Musings on Class D .... is it actually 'there' now? Should we all be using it?

[montesquieu]

The big Radford STA100 is off with Valvebloke (/Ampregen) for a spot of attention after it (embarrasingly) started making noises while I was trying to show it off to some visitors from overseas who were staying with us for a couple of weeks.

In its place I've put a Tripath TA2022 jobbie I picked up from Duncan which he sold with some single driver speakers and a RPi with Dac ... 90w so more than enough in theory at least to drive the big Tannoys.

It hasn't got all the sortedness of the big Radford, in terms of detail or dimensionality - you'd hardly expect it to. But it certainly has something ... solid in the bass, really very pleasant in terms of timbre, it's actually quite sweet, very composed at the top end. A touch more detail and transparency, a touch better with space, and it would really take a bit of beating. (I've certainly owned 40w valve amps that would struggle to do what it does).

In a way this should't be surprising, my front end is really wonderfully well sorted now and with a tube preamp to hold it all together, so long as there's a match with the power amp, and the power amp has enough welly to light up the HPD315a's (which not all power amps do, by any means) all should be well at 90s, and indeed it is. I had great results with a 275w Lyngdorf Class D previously, before my front end was as sorted as it is now.

So I guess my question is whether there's a Class D amp that would really take on the likes of my Radford, and other high end tube amps, into real world speakers, and BEAT them? (you know what I mean, aside from high-efff). In my mind they've certainly made big Class A sold state muscle amps a bit redundant - why bother with all that brute force and heat, Krells and stuff, for difficult speakers, when Class D does it already? But it seems to me that in the 40w-100w sort of area, valve amps have still ruled the roost.

I've been looking at the NuPrime stuff (that came out of the NuForce technology before the brand name was sold off), Nuforce (current) of course, current Lyngdorf gear. Has anyone (I mean anyone impartial - of course Class D has its cheerleaders) really made a study of Class D and really made the comparison with tube amps? I understand there were quite a few about at the Munich show.

'Scuse me if I'm late to this discussion but I've been dismissive up till now based on my own ears of just how 'right' a good valve amp can sound - unbeatable to date, in fact, in my setup. But my sense is that things are actually changing. As switching rates get higher and higher, power supplies get cleaner, surely we are bound to reach a point where Class D doesn't just cut it in terms of measurement, but also cuts it musically as well?

And then what!?

Anyway I'm quite keen to get hold of some NuPrime STA-9s (bridgeable to 290w monoblocks!) to hear just what state of the art might sound like, unfortunately no UK distributor it seems. $650 USD / €699 Euros in Holland. But I think very interesting to try. Any UK-available alternatives? (At sensible dosh of course).

[Nihil Sleighride]

I was interested in the NuPrime amps too but didn't want to commit to ordering from abroad. Audiologica have the new Nuforce gear and in the end I just went for STA120 to give me a flavour.

Got to say really impressed with the little thing and would be happy to give their and NuPrime's bigger monos, like what you are interested in, a go in the future.

[YNWaN]

I've listened to NCore amps and directly compared them to another amp (not valve). The NCores had great bass control but sounded a bit lacking in life and texture in the upper reaches which made them a bit 'shut in'. I also own a very powerful Class D amp that I drive my speakers bass drivers with and have used that as a full range amp in the past. Again that had amazing bass control and an extraordinary stop-start sort of immediate sound - very strongly propulsive. The mid was a bit hard sounding though and the upper frequencies lacked differentiation and sweetness.

[montesquieu]

I've listened to NCore amps and directly compared them to another amp (not valve). The NCores had great bass control but sounded a bit lacking in life and texture in the upper reaches which made them a bit 'shut in'. I also own a very powerful Class D amp that I drive my speakers bass drivers with and have used that as a full range amp in the past. Again that had amazing bass control and an extraordinary stop-start sort of immediate sound - very strongly propulsive. The mid was a bit hard sounding though and the upper frequencies lacked differentiation and sweetness.

Yes this was my experience too of an early Hypex 180w amplifier and to some extend the Tripath 2022 ... but not so much with the Lyngdorf.

My question is really about the very latest implementations of the Class D concept, my sense is that things have been streadily improving ... the stuff in the past for the most part fell always short in comparison to good valve amps.

[Eisener Bart]

I've been looking at the NuPrime stuff (that came out of the NuForce technology before the brand name was sold off), Nuforce (current) of course, current Lyngdorf gear. Has anyone (I mean anyone impartial - of course Class D has its cheerleaders) really made a study of Class D and really made the comparison with tube amps? I understand there were quite a few about at the Munich show.

I wanna show you only 2 photos.

It was before.

http://f20.ifotki.info/org/0a7a24d2491e4e7a0c971b0fdee181772eafa3251407688.jp g (http://i-fotki.info/)

Now.

http://f20.ifotki.info/org/9fe42a5e30acd30a3003e23e7bc84a722eafa3251407756.jp g (http://i-fotki.info/)

And full post of onwer.

"Well its been quite a while since I posted my first comments about the R20's back in March. (#12 in this thread) There's a reason and its a long story. I'll spare most of the details but first I want to give a huge shout out to Jason and Bob for the incredible service and response they have provided. These guys are the BEST and are true professionals in every way!!! It is very comforting to know that these guys are there, passionately standing behind the terrific Nuprime products that they represent.

Next.......and please pay close attention to this; if you are using subs with a Neutrik connector like RELs, do not hook up the sub cables to both terminals. Next, do not only hook up the hot connector and let the ground float. Finally, DO attach the positive connectors to the positive terminal and the negative to the tiny "ground screw" to the left of the negative amp terminal. Otherwise you will blow up the amps like I did; repeatedly. :x Then you will have to send them back to Bob like I did; repeatedly.

Currently I'm running the amps with Maggie 3.7is (Mye stands), REL G2's, CAT Legend preamp, Shunyata loom, Triton v2, Typhon, VPI Signature with 3D arm, Koetsu Rosewood Signature, and a Modwright (tube mods) Oppo 105 in a highly treated room. I listen near field about 7' away from the center of the speakers and have spent hundreds of hours tweaking every component. This is a very revealing system that exposes almost everything upstream. Other amps I own and have used for comparison are Carver Black Beauty 305's (KT120's/ 300w each) and a pair of Conrad Johnson Premier 8's with all factory teflon and KT120 upgrades (also 300w each). Both of these all tube mono blocks are outstanding and grab the Maggies with a vice-like grip that won't quit. I've also got a GAT 2 that has temporarily been sidelined while the CAT Legend is commanding center stage. For about the past couple of weeks I have been mixing and matching these electronics to get a good idea about the R20's and as you can see, they are in pretty good company. Plus, I'm a hard core tube, planar, and vinyl guy and in the past have avoided solid state, Class D, and digital like the proverbial plague.

So, what's the verdict on the R20's? Unfrickingbelievable! First, the tonality is unlike anything that resembles classic solid state. Instead, the sound is warm, lucious, and full, with layer upon layer of detail. The attack is sharp and crisp but the decay leaves notes sort of hanging in space with each one wrapped in a bubble of three dimensional reverberant air. All of the textures seem just right too. On cello, violin or stand up bass, you can hear the bow or the plucking of a string, but then the wood of the instrument reverberates in space until that note naturally fades or is overcome by the next. Drum skins are unreal. Horns have a natural, glorious, power to them that captures all of the subtle breathiness of quiet passages, but explodes without a hint of shrillness at a crescendo. And look out for tympani drums; a loud thwack is like getting slugged in the jaw! Dynamics are off the charts. Plus these amps have presence in spades. At times, with the lights turned low, its like the music just showed up on the soundstage and I've been transported to the venue, without the slightest recollection that there are electronic devices actually converting electrons into notes. Its tough to decide what to play next, although I've found myself migrating repeatedly to live recordings because it all sounds so real.

Meanwhile, my tube amps are scattered around the floor like audio jewelry. Beautiful to look at, but..... Anyway, I'm really not one to get carried away and gush about the latest and greatest newcomer to the audio scene. But these amps are something unique, especially at the price. I almost wish they were bigger and heavier and maybe had some KT150's glued to the top of the cases. Or I needed a hand truck to move them around. That way I could sort of rationalize their world class status and come to terms a little easier with how good they sound.

One more quick snippet. Right now I'm listening to "A Case of You" by Dianna Krall. We've all probably heard that cut way too many times. But as the music is playing beautifully, I'm fascinated by her pedal work on the piano. Right beneath her voice and breath, the keys striking the piano wires, the ambient sound of the hall, and the occasional sniffle in the audience, is a strong visual impression of her feet decisively working the pedals. I haven't quite experienced that before. And with almost every album, the R20's are revealing more, much more, than I have ever heard before. These amps get my strongest possible recommendation, and if they work this well with my Maggies, they should have a pretty easy time with just about any other speaker.

Thanks again to Jason and Bob for their great products and for being such terrific folks to work with. You guys are the best!"
http://www.audiocircle.com/index.php?topic=140718.40


You may find plenty information from happy owners of NuPrime products here:

http://www.audiocircle.com/index.php?board=199.0

and direct link on STA-9:
http://www.audiocircle.com/index.php?topic=138614.0

[Eisener Bart]

sorry, repeated post.

[struth]

Think my monoblocks, which can be used as stereos as well are class D.... Ive got 4. Cheap I guess but to me keep beating other things i try. They are fairly heavy and were designed to be used in Juke boxes.

[YNWaN]

My question is really about the very latest implementations of the Class D concept, my sense is that things have been streadily improving ... the stuff in the past for the most part fell always short in comparison to good valve amps.

Both of the amps I tried are very recent Class D designs.

[RichB]

I still use an Amptastic mini-t with a Croft valve pre/phono up front. There's definitely something to be said about a valve pre with them. Where it fails for me is in the bass. Not for the lack of it, but in the quality and 'realness' and believability of it. Once you've heard it you listen for it.

I've recently been mucking around with an Arcam FMJ amp and compared to the mini t the bass sounds more right and authentic than with the mini-t.

[jandl100]

I've been a fan of class D amps for years.
Some are pretty tame, and would suit the P&S brigade more than me - the TA2022 chip falls about half way in my experience - a bit bland.
There's something distinctive about the sound of some class D that works for me - I especially like the (sadly) low powered TA2020 chip - there's an openness and transparency and sheer transient speed that sounds very close to real to me.

But of those I have tried (quite a lot!) I have never heard any that could compete with large valve or large non-class D solid state amps.
I guess the Crown XLS1500 was the most recent. Rated at 300wpc (8 ohms) it sure as hell did not sound like a 300wpc amp should. It didn't even come close to the awesome majesty of my old (220wpc) Parasound A21 amp.
Power supplies are to blame here, I guess - or the lack of adequate PS in the Crown.

I've not heard a class D fitted out with a pukka large PS - Rowland, ARC come to mind as possibles, maybe?
Perhaps that would do the job, I would certainly like to hear a heavy class D amp!

[wee tee cee]

I have been using class D amps for a few years now and rate them very highly. Temple audio mono blocks when partnered with a decent pre can really punch above they're weight.

They tend to be very kind to awkward speaker loads and work particularly well with statics IMHO.

They respond to cable and PSU upgrades-the latter I settled on a big car battery that really lowered the noise floor.

I was looking at the NAD 7050...a few people I trust rate it highly.

I haven't tried them against high end valves but do own a nice KEL84 power amp and a pair of KEL84 mono blocks-the Temples sound more right to my ears (the valvey majik does encourage prolonged bourbon consumption though)

One of the troops ran a mini 1 with a modest psu whilst awaiting the build of his LDA 330B....It didn't disgrace itself in his well resolved system by any means.

Well worth some experimentation.

[Arkless Electronics]

Solid state class A is still the ultimate IMO and probably always will be. It needs to be done in a fairly money no object way for best results of course, which in my own case and for my own use means money is saved by having NO styling, aesthetics... or even casework!

Load sensitivity is a major issue for class D, becoming worse as frequency goes up and due to the output filter network (similar to a crossover filter in a passive speaker) at the output of class D amps. Engineering is all about balancing compromises in some areas and this is a prime example. Switching frequency V RFI V output filter network V switching losses V distortion V crossconduction protection etc...

ps yes this is a hugely simplistic sweeping statement but further technical analysis is beyond the scope of such threads....

[Gazjam]

Class D....Devialet's output section is the "Gold Standard" as far as I've ever heard. (And I've heard a fair few)

Heard a Deavielet 400 (think it was) compared to some "done right" valve power amplification and phono stage.
Some French wizardy afoot...Devialet hates Scotty and actually DOES "break the laws of physics", but valves done right is better.

Blows my mind that the current (French) digital guru's are catching up with what was done 40 years before.
Valves rool. (Its a rule)

[montesquieu]

I have been using class D amps for a few years now and rate them very highly. Temple audio mono blocks when partnered with a decent pre can really punch above they're weight.

They tend to be very kind to awkward speaker loads and work particularly well with statics IMHO.

They respond to cable and PSU upgrades-the latter I settled on a big car battery that really lowered the noise floor.

I was looking at the NAD 7050...a few people I trust rate it highly.

I haven't tried them against high end valves but do own a nice KEL84 power amp and a pair of KEL84 mono blocks-the Temples sound more right to my ears (the valvey majik does encourage prolonged bourbon consumption though)

One of the troops ran a mini 1 with a modest psu whilst awaiting the build of his LDA 330B....It didn't disgrace itself in his well resolved system by any means.

Well worth some experimentation.


I've heard/owned most of of these in some shape or form, and no agrument about sound per pound, but I'm more talking about high-end, state of the art/art of the possible, up there with the highest-end valve amps.




Class D....Devialet's output section is the "Gold Standard" as far as I've ever heard. (And I've heard a fair few)

Heard a Deavielet 400 (think it was) compared to some "done right" valve power amplification and phono stage.
Some French wizardy afoot...Devialet hates Scotty and actually DOES "breaks the laws of physics", but valves done right is better.

Blows my mind that the current (French) digital guru's are catching up with what was done 40 years before.
Valves rool. (Its a rule)

Totally agree this has always been the case up till now ... but if development continues, at which point might that cease to be true? This is what I've been musing about really.

@Jez ... at a high level I'm aware of the technical challenges of doing Class D right but I'd be keen to hear your thoughts ... in particular on what would need to be addressed in order to truly equal Class A? Like I said in my view the Lyndorf SDA2175 came pretty close to getting there.

[User211]

Heard a Hypex NC1200 design in my system with SMPS and an Audio Note M8. Absolutely awful. And very recent and very expensive.

An NC400 design with JFET input stage and no SMPS - best I have heard in class D. A prototype made by Jon Oakey of Reality Audio. I reviewed it on the Wam.

Avoid SMPS designs at all costs, is Jon's advice.

Jerry's Crown was quite nice sounding but a bit lacking and didn't "sound" powerful. Actually, nor did NC1200.

I had that Lyngdorf you had Tom in my system at a bake off. It was OK, not brilliant though.

The class G/H (http://www.audioholics.com/audio-amplifier/amplifier-classes) amp Jerry and I have is very good if slightly veiled. It has a lot of very positive qualities. I really like it and I think Jerry really likes his too.

[User211]

Class G/H is not similar to class D. So this is a little irrelevant I guess, but some blub from the Apart Champ manual:

"In the following pages we will explain how we have reached our goal. Champ-One features an
exceptionally dynamic output stage, a unique combination of class G and H amplifier topology,
all packed in a two unit fanless enclosure. While our competitors are struggling with often poorly
designed so called energy efficient class D digital amplifiers…well, we have taken a close look
and listened carefully to some of these designs, and we were not impressed. We got inspired to
make something APart from all the rest, and started from scratch.

Take a look at our new champion with his striking design and powerful set of features that will
exceed the demands of passionate music lovers as well as system integrators. Designed with a
continuous 4 ohm, 2 ohm dynamically stable high current output stage, Champ-One will surprise
you with its sonic excellence and dynamic capacities. Remember, this unique amplifier has been
developed with the impact of live music dynamics in mind. A true beauty, inside as well as outside,
but also a beast if necessary."

This amp easily beats the very expensive NC1200 and SMPS in my book, with a MUCH cheaper preamp than an M8. Massively so. It just sounds waaayyy better than that combo.

[YNWaN]

Jez's description of how class D behaves into low loads is how I understood it. Another contributor to this thread said that class D was good with punishing loads but that isn't really the case and power output rarely doubles as the impedance halves - in fact they are rarely good into low loads like 2 Ohms. The specs of my own amp do claim a doubling of wattage into 4 Ohms over 8 (600wpc instead of 300), but any lower impedance than that and output falls off steeply. I think this is likely the reason that class D amps often fail to sound as powerful as their specs would suggest (as Jerry mentioned earlier).

[montesquieu]

I had that Lyngdorf you had Tom in my system at a bake off. It was OK, not brilliant though.



I think it depends what it's driving, its 275w was about right for Tannoys that normally come alive with 40w+ of valve power, I wouldn't expect it to light up the Apogees very well.

Just checked the power specs on the Champ-One

Power consumption : 18VA idle, 1KVA full program, 1.5KVA peak

Yikes!

[User211]

Look at the efficiency rating Tom. It is better than class A/B designs at around 50-70%. It stays cool driving the Apogees lest you really push it. It never used 1KW that is a max continuous power thing.

[User211]

That said the link I provided to class G/H states that it is more efficient than class A/B but it still quotes 50-70% for class A/B.:scratch::mental:

[YNWaN]

As I understand it the Devialet approach is rather like the Musical Fidelity Superchargers (if you remember them) - they weren't class D but joined with ones system in a similar way. A small low powered Class A amp (remember that most active pre-amps operate in class A) of a couple of watts (probably less) drives a class D stage that does the real amplification. The idea is that the characteristics of class A are reflected in the output but rather assumes the class D stage is doing no harm of its own.

[Barry]

As I understand it the Devialet approach is rather like the Musical Fidelity Superchargers (if you remember them) - they weren't class D but joined with ones system in a similar way. A small low powered Class A amp (remember that most active pre-amps operate in class A) of a couple of watts (probably less) drives a class D stage that does the real amplification. The idea is that the characteristics of class A but rather assumes the class D is doing no harm of its own.

How does that differ from Quad's 'Current Dumping' technique?

[struth]

Jez's description of how class D behaves into low loads is how I understood it. Another contributor to this thread said that class D was good with punishing loads but that isn't really the case and power output rarely doubles as the impedance halves - in fact they are rarely good into low loads like 2 Ohms. The specs of my own amp do claim a doubling of wattage into 4 Ohms over 8 (600wpc instead of 300), but any lower impedance than that and output falls off steeply. I think this is likely the reason that class D amps often fail to sound as powerful as their specs would suggest (as Jerry mentioned earlier).

used mine into my telefunkens that are 2.75 ohms without issue. they are sensitive mind which may make a difference

[montesquieu]

How does that differ from Quad's 'Current Dumping' technique?

Presumably in using a Class D amp after the initial stage, rather than one in AB?

[Arkless Electronics]

Heard a Hypex NC1200 design in my system with SMPS and an Audio Note M8. Absolutely awful. And very recent and very expensive.

An NC400 design with JFET input stage and no SMPS - best I have heard in class D. A prototype made by Jon Oakey of Reality Audio. I reviewed it on the Wam.

Avoid SMPS designs at all costs, is Jon's advice.

Jerry's Crown was quite nice sounding but a bit lacking and didn't "sound" powerful. Actually, nor did NC1200.

I had that Lyngdorf you had Tom in my system at a bake off. It was OK, not brilliant though.

The class G/H (http://www.audioholics.com/audio-amplifier/amplifier-classes) amp Jerry and I have is very good if slightly veiled. It has a lot of very positive qualities. I really like it and I think Jerry really likes his too.

Eee so many things to answer but I'll start here....

A SMPS in effect IS a class D amplifier!! A linear regulated supply is basically a class A amplifier being used to give a DC voltage out, regulation being achieved by comparison with a fixed voltage reference. Same with a class D amp being used against a reference voltage to give a DC voltage out. The high efficiency of a SMPS is achieved in precisely the same way as the high efficiency of a class D amp... as they are the same..

[Arkless Electronics]

Class G/H is not similar to class D. So this is a little irrelevant I guess, but some blub from the Apart Champ manual:

"In the following pages we will explain how we have reached our goal. Champ-One features an
exceptionally dynamic output stage, a unique combination of class G and H amplifier topology,
all packed in a two unit fanless enclosure. While our competitors are struggling with often poorly
designed so called energy efficient class D digital amplifiers…well, we have taken a close look
and listened carefully to some of these designs, and we were not impressed. We got inspired to
make something APart from all the rest, and started from scratch.

Take a look at our new champion with his striking design and powerful set of features that will
exceed the demands of passionate music lovers as well as system integrators. Designed with a
continuous 4 ohm, 2 ohm dynamically stable high current output stage, Champ-One will surprise
you with its sonic excellence and dynamic capacities. Remember, this unique amplifier has been
developed with the impact of live music dynamics in mind. A true beauty, inside as well as outside,
but also a beast if necessary."

This amp easily beats the very expensive NC1200 and SMPS in my book, with a MUCH cheaper preamp than an M8. Massively so. It just sounds waaayyy better than that combo.

...and moving on.....

That quote is 100% meaningless marketing drivel! Class G and H are basically adjuncts to other amplifier classes and can be used with class A, class B, class AB or even class D. It can even combine any of the previous classes! Class G and H (I won't describe the difference between G and H as it's not that relevant to the discussion and means the opposite anyway depending on in Europe or USA) are rail switching topologies. To put it simply a low powered amp is used, say 5W, which can be taken up to being say a 200W amp on peaks only, or when the extra power is required, by switching the voltage rails of the amp to a much higher voltage in synchronisation with the music. This is called commutation. The idea is that say 70% of the time only the low power setting is switched in and so it only consumes the power and gives the heat of the low power amp for most of the time. The first commercial use in hi fi I can recall was the Hitachi "Dynaharmony" range....

[Arkless Electronics]

Jez's description of how class D behaves into low loads is how I understood it. Another contributor to this thread said that class D was good with punishing loads but that isn't really the case and power output rarely doubles as the impedance halves - in fact they are rarely good into low loads like 2 Ohms. The specs of my own amp do claim a doubling of wattage into 4 Ohms over 8 (600wpc instead of 300), but any lower impedance than that and output falls off steeply. I think this is likely the reason that class D amps often fail to sound as powerful as their specs would suggest (as Jerry mentioned earlier).

....and next...

When I mentioned load sensitivity it should be borne in mind that there can be many types of load sensitivity ;) What I was referring to is the effect of the output filter network on a class D amp. This filter has dual purposes, it is a reconstruction filter to give us a linear audio output and it is also a LPF (Low Pass Filter) to try and remove the severe RF interference output from class D amps.

The audio output of a class D amplifier is nothing more than supersonic square waves at high level! How's that for the ultimate interference source :eek: Much effort has to be taken to prevent it acting as a high power radio transmitter with the speaker cable as the aerial.

The filter at the output is there to filter out this high power RF and, by acting as a low pass filter, it allows the audio pulse width modulation of the train of square waves to be "demodulated" into audio. The filter is electrically similar to the section of a passive speaker crossover feeding the bass driver, in this case though it's all happening at a much higher frequency with the "crossover frequency" being more like 30KHz. Just as a speakers crossover network must be designed to work into the correct impedance bass driver for a flat response, so must the filter at the output of a class D amp... which is where load sensitivity comes in. If the filter is designed for a flat response with an 8R load then there will be treble drop into a 4R load and treble boost into a 16R load etc.

Oh and for any engineers cringing out there I am of course trying to keep this simple and I'm referring to generic class D and leaving aside the "ah but's... " etc to prevent this getting too beyond the scope of the matter...

[Arkless Electronics]

That said the link I provided to class G/H states that it is more efficient than class A/B but it still quotes 50-70% for class A/B.:scratch::mental:

...once more into the breach!

The theoretical maximum efficiency for a pure class B amplifier is 78.5%. This never achieved in practice and in the case of a class A/B amp 50-70% is about right... BUT only at maximum output! At lower levels they are vastly more inefficient than this and with a 100W amp working at say 5W efficiency would typically be only 15% or so....

Class G and H are used to gain greater efficiency and are very rarely used in hi fi amps due to commutation issues. They are most commonly found in big PA amplifiers where they can deliver practical efficiency in the 50 - 80% area and so result in cheaper, lighter and cooler running amps. Class D can deliver efficiency usually in the 85 - 95% area.

Class G and H, and specifically whichever version one wants to call the switching of different tiers of output devices, (as I hinted earlier, ask engineers from Europe, USA and Japan and you will get differing answers as to which is which between class G and H), has only one particularly interesting application to hi fi and that is to do most of the amplification with a pure class A small amp of say 10WPC and then switch in a tier of output devices in class A/B for the peaks. I'm not sure if any commercial amp has been made along these lines but I'd guess there will be one somewhere....

[Arkless Electronics]

...on to Devialet.

The basic operating principle of the Devialet amps is that it is a class A amp of a few Watts output being "swung between the rails" sort of class G or H stylee, of a couple of class D amps basically being used as SMPS's (see earlier for similarity of class D and SMPS). This gets a bit difficult to explain here... but a class A amp of say 3W can supply the entire output of say a 200W amplifier providing it can handle the full current delivered to the load. The purpose of the class D amps (which could be class A/B etc) is to keep the voltage seen by the small 3W amp down to the point where it only dissipates the power of a 3W class A amp even at much higher levels. It's not quite a first in that similar things have been tried before and even built, by Technics amongst others, but is a relatively untapped technology... I have my own technology along these lines which I have done a fair bit of theoretical development work on as well but I've not taken it beyond that.... hey it's virtually impossible to get people to buy conventional technology when you're a largely unheard of one man band, never mind stuff that's "out there" :)

The beauty of the Devialet system is the use of advanced digital signal processing to control the nitty gritty details of the technique that have proved a bit awkward before they came along ;)

[anthonyTD]

Some good posts there Jez! :)
A...

[Arkless Electronics]

I've heard/owned most of of these in some shape or form, and no agrument about sound per pound, but I'm more talking about high-end, state of the art/art of the possible, up there with the highest-end valve amps.





Totally agree this has always been the case up till now ... but if development continues, at which point might that cease to be true? This is what I've been musing about really.

@Jez ... at a high level I'm aware of the technical challenges of doing Class D right but I'd be keen to hear your thoughts ... in particular on what would need to be addressed in order to truly equal Class A? Like I said in my view the Lyndorf SDA2175 came pretty close to getting there.

Now that's the $64K question! Push pull class A is the "gold standard" for audio amplification (not single ended, which should not be used in audio power amps), especially if using solid state to avoid the output transformer and lack of PNP valves... It has the lowest possible distortion, highest damping factor, best control of load etc etc of any class and is also the most "purist" in theory and practice. I doubt that class D (or anything else) will ever beat it.

An interesting corollary though is that class D works well with no feedback, in fact it is difficult to apply negative feedback to a class D amp, and class A push pull is the best topology for zero feedback designs in non switching amps. It may well be that the lack of feedback in some class D amps gives them the subjectively good results many report and that although class A is superior, class D could help to give better sound per pound in the future.... After all, the advantages of class A are so huge that the only reason it is not used for all hi fi power amps is the huge cost, weight and heat produced!

[Arkless Electronics]

Some good posts there Jez! :)
A...

Thanks Anthony :) It's just difficult to keep to the main task and not fly off at all the ifs and buts and exceptions that prove the rule and the "what about the Acme model...." that folk will no doubt now come back at me with after I've kept it so generic and simplified... Hope it's of use to some though;)

[anthonyTD]

Your Right Jez, it is difficult, but the fact that you havent Flown off the handle, and instead, taken the time to explain [all be it in a simplified way] the adavantages, and disadvantages of the diffrent operating classes, will' I am sure be of great benefit to those genuinely interested in, and contemplating getting their head around the diffrent topologies.:)
Thanks Anthony :) It's just difficult to keep to the main task and not fly off at all the ifs and buts and exceptions that prove the rule and the "what about the Acme model...." that folk will no doubt now come back at me with after I've kept it so generic and simplified... Hope it's of use to some though;)

[BilliumB]

An interesting corollary though is that class D works well with no feedback, in fact it is difficult to apply negative feedback to a class D amp, and class A push pull is the best topology for zero feedback designs in non switching amps. It may well be that the lack of feedback in some class D amps gives them the subjectively good results many report and that although class A is superior, class D could help to give better sound per pound in the future.... After all, the advantages of class A are so huge that the only reason it is not used for all hi fi power amps is the huge cost, weight and heat produced!

Hi Jez - could you say a bit more about the benefits of zero feedback in amplifier design.

Cheers, Bill

[walpurgis]

Now that's the $64K question! Push pull class A is the "gold standard" for audio amplification (not single ended, which should not be used in audio power amps), especially if using solid state to avoid the output transformer and lack of PNP valves... It has the lowest possible distortion, highest damping factor, best control of load etc etc of any class and is also the most "purist" in theory and practice. I doubt that class D (or anything else) will ever beat it.

An interesting corollary though is that class D works well with no feedback, in fact it is difficult to apply negative feedback to a class D amp, and class A push pull is the best topology for zero feedback designs in non switching amps. It may well be that the lack of feedback in some class D amps gives them the subjectively good results many report and that although class A is superior, class D could help to give better sound per pound in the future.... After all, the advantages of class A are so huge that the only reason it is not used for all hi fi power amps is the huge cost, weight and heat produced!

I agree with everything you've said Jez.

[Arkless Electronics]

Hi Jez - could you say a bit more about the benefits of zero feedback in amplifier design.

Cheers, Bill

The flippant answer would be "there aren't any".... other than a total lack of any possibility of instability at ultrasonic frequencies.

Virtually every aspect of the performance of an amplifier is improved in direct proportion to the amount of negative feedback used, it's that much of a miracle, and when invented by Harold S Black whilst on the ferry on his way to work in 1928 and jotted down literally on the back of a fag packet, the world of electronics was turned on its head... eventually! It was considered so "out there" even by its inventor that it wasn't utilised or patented for a few years afterwards. It's first big application (and the one Black was looking for an answer for) was the linearisation of SET valve amplifiers (using some valves still beloved by the SET aficionados) used as repeater amplifiers for the telephone system (he was employed, as were many of the worlds greatest engineers, by Bell Labs). Having 5-10% distortion, when these amps were used in chains for long distance communications the distortion after going through several of these amps made the output garbage by the time it reached the other end.... Enough of the history lesson though:D

Without negative feedback hi fi would not exist, not as we know it anyway. It is the thing which allows good measured performance and which makes even acceptable performance affordable and reproducible. It is the magic which gives us low hum and hiss, distortion below a couple of % as a matter of course, wide flat frequency response and good damping factor, again as a matter of course. Without NFB (Negative FeedBack) the matching between channels of a stereo amp would require careful hand selection and measuring of individual components... so no budget gear for a start! and stuff at the ghetto blaster end of the market would be so dire as to be not fit for purpose or so expensive no one would want it!

So what's not to like about NFB? It all comes down to subjectivism and to matters which even us engineers can only speculate about at the end of the day!

Technically NFB is the nearest thing to a free lunch and win, win situation in analogue electronics. It's only technical drawback is that there is a limit to how much NFB can be applied before, at some ultrasonic frequency, it becomes positive feedback, the amp goes unstable, and all hell breaks loose usually blowing output transistors etc.
The "accepted wisdom", for most of the last 100 years or so anyway, is to always use as much NFB as possible as it can only do good and the more you use the better it gets. It is precisely this quest for more NFB that lead to Arthur Radford designing better output transformers and Arthur Bailey designing a better phase splitter for the Radford amps. This allowed them to use more NFB without it getting to the "Oh fuck" point I outlined above.... It's also what "TL" stands for in the Leak TL12 Point One... TL is "Triple Loop" meaning 3 NFB loops and yes that's how they got it down to 0.1% THD...

Excessive use of NFB can, sometimes, lead to an excessively "grey", "leaden", "2d", "flat", "grainy" sound.... according to some... and yes I have found this to be true myself, but not always!! It must be used carefully and not to much as a "band-aid" to cure everything. Even amongst electronic engineers you will find considerable disagreement as to how much is too much, whether there is even such a thing as too much, when and how to use it etc.
As an interesting, if technical aside, and bearing in mind how this thread started, Bruno Putzeys ex of Philips and now one of the worlds top (if not THE top) designers of class D amps (he designed the Hypex and other technologies) wrote an interesting piece about NFB a few years back... IIRC he's of the "no such thing as too much" camp on NFB, as is Doug Self, who apparently never does any subjective testing on his designs....

Others believe it should be minimised or not used at all!! Myself? I use it in moderation and depending on the required results and methods being used to achieve it. Now and then, where appropriate, I use no NFB but it is rare that one can get away with this without it causing much bigger problems than any solved by not using NFB... like ending up with THD > 0.5% and which no Arkless Electronics product would leave the drawing board with anywhere near that amount of THD!

As a generalisation, the less NFB used then the more expensive and difficult it's going to be to get some minimum standard of measured performance

Valve power amps generally use far less NFB than solid state power amps and this is something which some have said is largely responsible for the valve sound.... it's partly responsible I would say. Why do they use much less NFB? They have to! The output transformer messes things up so much that only fairly small amounts can be used without the thing going haywire! Because they work usually in class A or largely in class A they can get away with less NFB.

To put things in perspective though, the budget or mid price solid state shoe box size amp would not be possible without lots of NFB.. It's the "cure all band aid" which makes such a thing capable of MUCH better than the 20% THD, loads of hum and hiss, frequency response of maybe 100 Hz - 8KHz (and different on each channel and each amp) plus unpredictable channel balance heap of shit nightmare which a £500 integrated amp would otherwise be!!

[Arkless Electronics]

I agree with everything you've said Jez.

Thanks Geoff. Pity though that properly digging into the meat of hi fi like this will usually produce little interest.... whereas if I'd started something about "do oak or spruce wooden cones under the CD player sound best" or some such bollocks, it would run for pages!
:sofa:

[walpurgis]

Come on Jez. Flippin Spruce? You know perfectly well it has to be Patagonian Rosewood for the best results. ;) :)

[BilliumB]

Hi Jez

Thanks for such a detailed response - very interesting.

The reason I started thinking about the issue is that a couple of years ago I purchased a second hand Conrad Johnson solid state amp (CA200 - they call it a Control Amplifier - it's a 180W integrated containing a passive pre) that claims to use zero global feedback (I assume this might mean it uses some local NFB?) - I'm very happy with it, but have not really compared it with other amps to know how good it is. Clearly it was an expensive beast when new, and uses very carefully selected components as you alluded to in your thoughts.

Since getting the CJ, I've looked for but not found much supporting the 'zero negative feedback' approach, with practically everyone extolling the benefits of NFB. However, I came across the following paper from Nelson Pass suggesting NFB increases higher order harmonic distortion products - any thoughts?

https://passlabs.com/articles/audio-distortion-and-feedback

I've separately been wondering whether widespread use of high levels of NFB is why many people seem to claim that all amplifiers sound the same (please be clear that I've listened to so few different amps that I don't have a view on this claim).

More thoughts would be very interesting.

Cheers, Bill

[Arkless Electronics]

Come on Jez. Flippin Spruce? You know perfectly well it has to be Patagonian Rosewood for the best results. ;) :)

:D I wondered how long it wood take for anyone to twig that I'd mentioned that unlistenable soft wood :eyebrows:

[jollyfix]

Come on Jez. Flippin Spruce? You know perfectly well it has to be Patagonian Rosewood for the best results. ;) :)

Sorry Geoff ,you are wrong, its Rosewood, but from Brazil that works best.....:lol:
Cheers Jez, enjoying your posts on this thread too...

[walpurgis]

Both my SS Class A power amps are described as 'zero feedback' jobs. They certainly sound damn good.

[Macca]

Jez - fascinating stuff - What about the Radford STA 25 - has to be the most fundamentally 'right' sounding amp I ever heard. Why is that?

[Arkless Electronics]

Hi Jez

Thanks for such a detailed response - very interesting.

The reason I started thinking about the issue is that a couple of years ago I purchased a second hand Conrad Johnson solid state amp (CA200 - they call it a Control Amplifier - it's a 180W integrated containing a passive pre) that claims to use zero global feedback (I assume this might mean it uses some local NFB?) - I'm very happy with it, but have not really compared it with other amps to know how good it is. Clearly it was an expensive beast when new, and uses very carefully selected components as you alluded to in your thoughts.

Since getting the CJ, I've looked for but not found much supporting the 'zero negative feedback' approach, with practically everyone extolling the benefits of NFB. However, I came across the following paper from Nelson Pass suggesting NFB increases higher order harmonic distortion products - any thoughts?

https://passlabs.com/articles/audio-distortion-and-feedback

I've separately been wondering whether widespread use of high levels of NFB is why many people seem to claim that all amplifiers sound the same (please be clear that I've listened to so few different amps that I don't have a view on this claim).

More thoughts would be very interesting.

Cheers, Bill


It's an old chestnut (back to wood :D) first, to the best of my knowledge, mentioned by M G Scroggie in the early 50's and then a few time here and there since... (Attwood, Baxendall...). The crux of it goes that if you use NFB too sparingly it can makes worse by attenuating the second harmonic and leaving greater total energy in the third and higher harmonics, which are musically dissonant, especially if odd harmonics. The cure is to make sure that if use use NFB use at least around 10dB as an absolute minimum;)

[Arkless Electronics]

Both my SS Class A power amps are described as 'zero feedback' jobs. They certainly sound damn good.

Yep. The more output devices you use and the heavier into class A they are biassed the better everything gets. With push pull SS class A, take it far enough and you will get <0.1% THD, plenty of damping factor etc etc all without any NFB. And wonderful sound. Or you can use a little NFB, in valve amp proportions, and get <0.001% THD as in my main home amp at the moment.

[Arkless Electronics]

Jez - fascinating stuff - What about the Radford STA 25 - has to be the most fundamentally 'right' sounding amp I ever heard. Why is that?

It's one of the most technically correct valve power amps around, designed above all else to measure good and made so by using loads of NFB... by valve standards, as I explained earlier. Just good electronic engineering without the slightest trace of foo! They have almost SS like measured performance in many ways ;)
There's probably a bit of luck in as much as "it just turned out that way" as well...

[BilliumB]

From the paper, I think Nelson is suggesting that an amp with few stages and no NFB might sound better than one using even moderate NFB. If you've got the time I'd be interested in what you think of his paper.

Cheers. Bill

[Macca]

It's one of the most technically correct valve power amps around, designed above all else to measure good and made so by using loads of NFB... by valve standards, as I explained earlier. Just good electronic engineering without the slightest trace of foo! They have almost SS like measured performance in many ways ;)
There's probably a bit of luck in as much as "it just turned out that way" as well...

Interesting that it is technically good too...I'm also more of a solid-state man but there is something about the Radford that is right on some sort of primitive part of the brain. An instinctive rightness, if you like. I guess it is about getting the balance of ingredients just right and there is always going to be some chance in that.

[Arkless Electronics]

From the paper, I think Nelson is suggesting that an amp with few stages and no NFB might sound better than one using even moderate NFB. If you've got the time I'd be interested in what you think of his paper.

Cheers. Bill

I'm familiar with the article yes. it's basically just a rewriting of earlier work by other authors I named earlier.. nothing new there but a perfectly good article. As i said even engineers don't agree on this! I like his Stasis circuitry from years ago (Threshold) but it has been made less relevant by better output devices. F5 and similar pretty good too but I would use more NFB than that myself in that application (and yes I've done it). I'm not keen on his Zen and son of zen amps as they are single ended... yes he's done his utmost to get the best from such simple circuitry and done a damn good job but they're still single ended and you can only polish a turd to a certain lustre.... I can't think of any plus points re single ended power amps of any type... but loads of negatives! For audio use, single ended amps must work in class A to even work at all in any way that wouldn't be so distorted that you could not make out words etc... they make even normal push pull class A look efficient and cool running... and it goes downhill from there really. An Audionote SET zero NFB amp I had in recently could only just meet the minimum spec for PA type amps for use as "Tannoy" systems in shopping centres etc of 10% THD, beyond which it is considered that intelligibility of speech could be effected and render fire evacuation announcements etc too garbled. European directive EN54/16 for anyone really that bored.... I once worked in the industry designing such gear.... boy was it exiting... not.

[User211]

...and moving on.....

That quote is 100% meaningless marketing drivel! Class G and H are basically adjuncts to other amplifier classes and can be used with class A, class B, class AB or even class D. It can even combine any of the previous classes! Class G and H (I won't describe the difference between G and H as it's not that relevant to the discussion and means the opposite anyway depending on in Europe or USA) are rail switching topologies. To put it simply a low powered amp is used, say 5W, which can be taken up to being say a 200W amp on peaks only, or when the extra power is required, by switching the voltage rails of the amp to a much higher voltage in synchronisation with the music. This is called commutation. The idea is that say 70% of the time only the low power setting is switched in and so it only consumes the power and gives the heat of the low power amp for most of the time. The first commercial use in hi fi I can recall was the Hitachi "Dynaharmony" range....

You've got to expect complete BS when it comes to hi-fi marketing. Almost all marketing, really.

I liked it purely because the amp sounds better to me than Bruno's dreadful sonic achievement with SMPS and NC1200. Any BS that gets people away from that sonic "way below averageness" has to be a plus.

[anthonyTD]

Again good post Jez,
Although, I will have to diasagree with you on the Single ended stuff, as far as ALL of it being quite crap,:eyebrows: However, I will agree that the usual basic conventional way of doing it is not great, but throw some time effort, and a load of modern semiconductor circuitry at it, and it can be quite a diffrent story, as in the Soul Series, Although they cannot be classed as a true valve amp' as they are hybrids, only using valves in the output, and as stated, using lots of semiconductor circuitry to control all of the valves parameters, including the heater circuits, And yes, they use more than the usual permited NFB, all these benefits become obvious when you test them and listen to them, for eg; one main aspect of their performance; flat frequency response at a few watts from around 10hz, to 250khz!
Compare that to conventional single ended, which would be around 50hz, to 18khz! :)
I'm familiar with the article yes. it's basically just a rewriting of earlier work by other authors I named earlier.. nothing new there but a perfectly good article. As i said even engineers don't agree on this! I like his Stasis circuitry from years ago (Threshold) but it has been made less relevant by better output devices. F5 and similar pretty good too but I would use more NFB than that myself in that application (and yes I've done it). I'm not keen on his Zen and son of zen amps as they are single ended... yes he's done his utmost to get the best from such simple circuitry and done a damn good job but they're still single ended and you can only polish a turd to a certain lustre.... I can't think of any plus points re single ended power amps of any type... but loads of negatives! For audio use, single ended amps must work in class A to even work at all in any way that wouldn't be so distorted that you could not make out words etc... they make even normal push pull class A look efficient and cool running... and it goes downhill from there really. An Audionote SET zero NFB amp I had in recently could only just meet the minimum spec for PA type amps for use as "Tannoy" systems in shopping centres etc of 10% THD, beyond which it is considered that intelligibility of speech could be effected and render fire evacuation announcements etc too garbled. European directive EN54/16 for anyone really that bored.... I once worked in the industry designing such gear.... boy was it exiting... not.

[George47]

Tom,

I am using a Class D over the summer here. It is the Nord NC500 and I have put together some views on the Wam.

In summary I have really disliked Class D so far including some expensive NC1200 amps from the US (the one User 211 had). All of them had this real flat undynamic sound, it was destailed and very neutral but otherwise 'snoresville'. But the Nords are different. Instead of using the standard £1.69 op amp input (LM4562) they offer a plug-in socket and you can use the Burson, Sparkos and Sonic Imagery 994. I preferred the 994s ($80) but the Sparkos do have a softer and warmer sound. I suspect they won't 'convert' people but those that are interested in Class D may find them better than other Class D apart from ultra expensives. It may be alternative and for a coffee or two and some biccies (I know I am tart) I could bring them round. :)

[Arkless Electronics]

Again good post Jez,
Although, I will have to diasagree with you on the Single ended stuff, as far as ALL of it being quite crap,:eyebrows: However, I will agree that the usual basic conventional way of doing it is not great, but throw some time effort, and a load of modern semiconductor circuitry at it, and it can be quite a diffrent story, as in the Soul Series, Although they cannot be classed as a true valve amp' as they are hybrids, only using valves in the output, and as stated, using lots of semiconductor circuitry to control all of the valves parameters, including the heater circuits, And yes, they use more than the usual permited NFB, all these benefits become obvious when you test them and listen to them, for eg; one main aspect of their performance; flat frequency response at a few watts from around 10hz, to 250khz!
Compare that to conventional single ended, which would be around 50hz, to 18khz! :)

With sufficient "gilding of the lily" sure, single ended can be made to work well. It is still always a compromise in some areas though IMHO. Most of the adjuncts to the basic single ended topologies do in fact take them more towards being push pull.... Constant current loading will get the distortion down and one can even make the current source driven to further improve linearity and efficiency. Taken far enough one then gets to the point of "is this single ended or push pull?", a good example being the Linsley Hood 10W class A amp.

A major issue with single ended is the inability to move into class A/B as load impedance falls and this is a big problem. Lets say we have a 20W class A push pull amp and a 20W single ended amp (has to be class A as it's single ended). Both are rated at 20W into 8R. Now lets say we want to use this amp with 4R speakers.... or that the 8R ones have dips down to 4R or less.... What happens then?

In the case of the push pull amp it is now capable of 40W into the 4R speaker, only the first 10W will be in class A but it will move seamlessly into class A/B to accommodate the greater current needed to drive 40W into 4R. 10W of class A is still enough so that only loud peaks will likely drive it into class A/B so you are still getting class A for most of the time.

In the case of the single ended though things are very different. Rated at 20W into 8R it can only deliver 10W into 4R at which point it hits a brick wall and clips.... asymmetrically usually.

Then add into the mix the fact that single ended is twice as inefficient as even normal push pull class A... and that's with constant current or choke loading! Simple resistively loaded single ended manages a heady 12.5% maximum efficiency:eek: Complete with gross distortion...

I really can't think of a single plus point for single ended... At its simplest it could be said simplicity itself is a plus point... but there would only be a slight saving number of parts used and it needs at least twice the power supply!! Add on all the extra electronics to give it a chance of decent performance and it's more complicated than push pull and still needs much bigger power supplies and heatsinks.

[Arkless Electronics]

Tom,

I am using a Class D over the summer here. It is the Nord NC500 and I have put together some views on the Wam.

In summary I have really disliked Class D so far including some expensive NC1200 amps from the US (the one User 211 had). All of them had this real flat undynamic sound, it was destailed and very neutral but otherwise 'snoresville'. But the Nords are different. Instead of using the standard £1.69 op amp input (LM4562) they offer a plug-in socket and you can use the Burson, Sparkos and Sonic Imagery 994. I preferred the 994s ($80) but the Sparkos do have a softer and warmer sound. I suspect they won't 'convert' people but those that are interested in Class D may find them better than other Class D apart from ultra expensives. It may be alternative and for a coffee or two and some biccies (I know I am tart) I could bring them round. :)

As a general note on these after market discrete op amps I would say caution is advised. IC op amps such as the LM4562 and similar have become so good that they take a lot of beating and are unlikely to be a major weak link, cheap as they are. There is no way that replacing one of the better modern audio op amps with a discrete op amp is going to make any sort of night and day, or "wow" difference... Subtle at best. Caveat emptor ;)

[anthonyTD]

:)
With sufficient "gilding of the lily" sure, single ended can be made to work well. It is still always a compromise in some areas though IMHO. Most of the adjuncts to the basic single ended topologies do in fact take them more towards being push pull.... Constant current loading will get the distortion down and one can even make the current source driven to further improve linearity and efficiency. Taken far enough one then gets to the point of "is this single ended or push pull?", a good example being the Linsley Hood 10W class A amp.

A major issue with single ended is the inability to move into class A/B as load impedance falls and this is a big problem. Lets say we have a 20W class A push pull amp and a 20W single ended amp (has to be class A as it's single ended). Both are rated at 20W into 8R. Now lets say we want to use this amp with 4R speakers.... or that the 8R ones have dips down to 4R or less.... What happens then?

In the case of the push pull amp it is now capable of 40W into the 4R speaker, only the first 10W will be in class A but it will move seamlessly into class A/B to accommodate the greater current needed to drive 40W into 4R. 10W of class A is still enough so that only loud peaks will likely drive it into class A/B so you are still getting class A for most of the time.

In the case of the single ended though things are very different. Rated at 20W into 8R it can only deliver 10W into 4R at which point it hits a brick wall and clips.... asymmetrically usually.

Then add into the mix the fact that single ended is twice as inefficient as even normal push pull class A... and that's with constant current or choke loading! Simple resistively loaded single ended manages a heady 12.5% maximum efficiency:eek: Complete with gross distortion...

I really can't think of a single plus point for single ended... At its simplest it could be said simplicity itself is a plus point... but there would only be a slight saving number of parts used and it needs at least twice the power supply!! Add on all the extra electronics to give it a chance of decent performance and it's more complicated than push pull and still needs much bigger power supplies and heatsinks.

[George47]

As a general note on these after market discrete op amps I would say caution is advised. IC op amps such as the LM4562 and similar have become so good that they take a lot of beating and are unlikely to be a major weak link, cheap as they are. There is no way that replacing one of the better modern audio op amps with a discrete op amp is going to make any sort of night and day, or "wow" difference... Subtle at best. Caveat emptor ;)

Hmmm.... I tried the Sonic Imagery 994 and the Sparkos op amps and I would say they are quite different. Not night and day but pretty easy to hear with the Sparkos being warmer, with a more powerful bass but having a softer overall sound. The SI 994 was a more dynamic device.

These amps also have discrete component regulators which may add more to the whole package.

To some extent it is not relevant what matters is what comes out of the speakers and I preferred this amp compared to the Merrill versions. Even if they have enormous negative feedback which Bruno believes/found had a much better measured performance whilst accepting that not applying enough leads to problems.

[Arkless Electronics]

Hmmm.... I tried the Sonic Imagery 994 and the Sparkos op amps and I would say they are quite different. Not night and day but pretty easy to hear with the Sparkos being warmer, with a more powerful bass but having a softer overall sound. The SI 994 was a more dynamic device.

These amps also have discrete component regulators which may add more to the whole package.

To some extent it is not relevant what matters is what comes out of the speakers and I preferred this amp compared to the Merrill versions. Even if they have enormous negative feedback which Bruno believes/found had a much better measured performance whilst accepting that not applying enough leads to problems.

It all started as far as any commercial thang goes with the JE990 from Jensen Transformers and designed by Deane Jensen, this was back in 1979...
It was mainly intended for pro audio use and compared to any IC op amp available in 1979 it was a step up... It's still superb today in fact!
However, there has been so many improvements in IC op amps since then that I can't see them being particularly relevant any more, not as a simple plug in replacement for ordinary IC op amps anyway. Discrete op amps do have their uses, such as high voltage and or current output capabilities but this is not relevant to a replacement for an IC op amp being used from +&- 15V.... I actually designed and built a discrete op amp several years ago for a special purpose, it had to deliver 90V RMS to the load, which it did and still does superbly...

There are slight differences between the sound of op amps yes but they are pretty small IME and although there will also be changes between the sounds from different discrete op amps, and between them and IC op amps, IMO IC op amps have reached such a pinnacle that I would look long and hard at replacing £1.50 IC op amps with £80 discrete ones to get minuscule improvement....or in many cases minuscule changes rather than improvements, and especially when one will usually need several op amps as there will often be more than one in a piece of equipment. £80 a go soon adds up!

Discrete regulators still easily outperform IC ones but in the case of op amps it needs to be borne in mind that op amps are intrinsically fairly insensitive to their power supplies ;)

YMMV as usual and have fun tinkering :)

[Mr. C]

As with everything implementation is the key, they are many ways to skin a cat.

There are few exceptions in the class 'd' arena, however a true class 'd' is an amplifier that takes purely a digital input signal and nothing else for it to be a true class 'd'

A PWM switching amplifier with take an analogue signal to.


They key is controlling the amount of RFI generated, NOT using a SMPS to power them, linear psu ultra tight regulation on the control module, being able to take the switching frequency beyond 400/500Khz imho it needs to be well over 1Mhz high end and wide enough not cause other consideration.

They are many other items that need addressing however it can be done and they can generate full bodied texture, layering and three dimensionality (not just a pimped up 2D ala Devialet)

Everyone bangs on about the bass, about having wonderful control which they do have to a point but they very rarely can a PWM genuinely recreate truly deep, articulate and textured renderings of bass. NCores certainly can't they would make great PA modules, maybe PMC could have a tie up there :eek:

Then there is the bandwidth with the exception of two I know off they all top out around 50Khz, (the Devialet manages the forth harmonic, however it still sounds soulless and cardboardy) imho you cannot truly attempt to reproduce a decent stab at music until you have at least the firth harmonic +

A correctly design and built class a/b with produce a better sound in 95% of cases imho

[Arkless Electronics]

As with everything implementation is the key, they are many ways to skin a cat.

There are few exceptions in the class 'd' arena, however a true class 'd' is an amplifier that takes purely a digital input signal and nothing else for it to be a true class 'd'

A PWM switching amplifier with take an analogue signal to.


They key is controlling the amount of RFI generated, NOT using a SMPS to power them, linear psu ultra tight regulation on the control module, being able to take the switching frequency beyond 400/500Khz imho it needs to be well over 1Mhz high end and wide enough not cause other consideration.

They are many other items that need addressing however it can be done and they can generate full bodied texture, layering and three dimensionality (not just a pimped up 2D ala Devialet)

Everyone bangs on about the bass, about having wonderful control which they do have to a point but they very rarely can a PWM genuinely recreate truly deep, articulate and textured renderings of bass. NCores certainly can't they would make great PA modules, maybe PMC could have a tie up there :eek:

Then there is the bandwidth with the exception of two I know off they all top out around 50Khz, (the Devialet manages the forth harmonic, however it still sounds soulless and cardboardy) imho you cannot truly attempt to reproduce a decent stab at music until you have at least the firth harmonic +

A correctly design and built class a/b with produce a better sound in 95% of cases imho

I'm afraid your first statement is simply 100% wrong. The D in class D has sweet FA to do with digital and no it is not correct that only a digital input class D is true class D. Class D is analogue.

There is also no reason to avoid SMPS above and beyond any reasons for using/not using them with any other type of amplifier. As I explained much earlier a SMPS in effect IS a class D amplifier.

Bandwidth..... check out what I said about the output filter network being a Low Pass Filter... The top end of a class D will indeed drop like a brick beyond the cut off frequency of the filter. That's the idea. It has to in order to prevent huge amounts of radio interference. Take a portable radio somewhere near a class D amp, even 20 foot away... and tune through the bands and you will almost certainly find all sorts of strange whistles and squawks and usually even the music you are playing on the hi fi coming through the radio!
In effect you have a "passive crossover" at the output of class D amps and the properties of the inductors and capacitors will come into play...

As for the Devialet well it could be claimed it is a class A.... It depends how you want to classify your classes ("class distinction" as Doug Self once called it) and how much poetic license you give it ;)

[struth]

some info on the other types of amplifiers bar the usual A,AB etc...



Class D Amplifier – A Class D audio amplifier is basically a non-linear switching amplifier or PWM amplifier. Class-D amplifiers theoretically can reach 100% efficiency, as there is no period during a cycle were the voltage and current waveforms overlap as current is drawn only through the transistor that is on.

Class F Amplifier – Class-F amplifiers boost both efficiency and output by using harmonic resonators in the output network to shape the output waveform into a square wave. Class-F amplifiers are capable of high efficiencies of more than 90% if infinite harmonic tuning is used.

Class G Amplifier – Class G offers enhancements to the basic class AB amplifier design. Class G uses multiple power supply rails of various voltages and automatically switches between these supply rails as the input signal changes. This constant switching reduces the average power consumption, and therefore power loss caused by wasted heat.

Class I Amplifier – The class I amplifier has two sets of complementary output switching devices arranged in a parallel push-pull configuration with both sets of switching devices sampling the same input waveform. One device switches the positive half of the waveform, while the other switches the negative half similar to a class B amplifier. With no input signal applied, or when a signal reaches the zero crossing point, the switching devices are both turned ON and OFF simultaneously with a 50% PWM duty cycle cancelling out any high frequency signals.
To produce the positive half of the output signal, the output of the positive switching device is increased in duty cycle while the negative switching device is decreased by the same and vice versa. The two switching signal currents are said to be interleaved at the output, giving the class I amplifier the named of: “interleaved PWM amplifier” operating at switching frequencies in excess of 250kHz.

Class S Amplifier – A class S power amplifier is a non-linear switching mode amplifier similar in operation to the class D amplifier. The class S amplifier converts analogue input signals into digital square wave pulses by a delta-sigma modulator, and amplifies them to increases the output power before finally being demodulated by a band pass filter. As the digital signal of this switching amplifier is always either fully “ON” or “OFF” (theoretically zero power dissipation), efficiencies reaching 100% are possible.

Class T Amplifier – The class T amplifier is another type of digital switching amplifier design. Class T amplifiers are starting to become more popular these days as an audio amplifier design due to the existence of digital signal processing (DSP) chips and multi-channel surround sound amplifiers as it converts analogue signals into digital pulse width modulated (PWM) signals for amplification increasing the amplifiers efficiency. Class T amplifier designs combine both the low distortion signal levels of class AB amplifier and the power efficiency of a class D amplifier.

[Arkless Electronics]

some info on the other types of amplifiers bar the usual A,AB etc...



Class D Amplifier – A Class D audio amplifier is basically a non-linear switching amplifier or PWM amplifier. Class-D amplifiers theoretically can reach 100% efficiency, as there is no period during a cycle were the voltage and current waveforms overlap as current is drawn only through the transistor that is on.

Class F Amplifier – Class-F amplifiers boost both efficiency and output by using harmonic resonators in the output network to shape the output waveform into a square wave. Class-F amplifiers are capable of high efficiencies of more than 90% if infinite harmonic tuning is used.

Class G Amplifier – Class G offers enhancements to the basic class AB amplifier design. Class G uses multiple power supply rails of various voltages and automatically switches between these supply rails as the input signal changes. This constant switching reduces the average power consumption, and therefore power loss caused by wasted heat.

Class I Amplifier – The class I amplifier has two sets of complementary output switching devices arranged in a parallel push-pull configuration with both sets of switching devices sampling the same input waveform. One device switches the positive half of the waveform, while the other switches the negative half similar to a class B amplifier. With no input signal applied, or when a signal reaches the zero crossing point, the switching devices are both turned ON and OFF simultaneously with a 50% PWM duty cycle cancelling out any high frequency signals.
To produce the positive half of the output signal, the output of the positive switching device is increased in duty cycle while the negative switching device is decreased by the same and vice versa. The two switching signal currents are said to be interleaved at the output, giving the class I amplifier the named of: “interleaved PWM amplifier” operating at switching frequencies in excess of 250kHz.

Class S Amplifier – A class S power amplifier is a non-linear switching mode amplifier similar in operation to the class D amplifier. The class S amplifier converts analogue input signals into digital square wave pulses by a delta-sigma modulator, and amplifies them to increases the output power before finally being demodulated by a band pass filter. As the digital signal of this switching amplifier is always either fully “ON” or “OFF” (theoretically zero power dissipation), efficiencies reaching 100% are possible.

Class T Amplifier – The class T amplifier is another type of digital switching amplifier design. Class T amplifiers are starting to become more popular these days as an audio amplifier design due to the existence of digital signal processing (DSP) chips and multi-channel surround sound amplifiers as it converts analogue signals into digital pulse width modulated (PWM) signals for amplification increasing the amplifiers efficiency. Class T amplifier designs combine both the low distortion signal levels of class AB amplifier and the power efficiency of a class D amplifier.

careful when copying and pasting from the internet.....

Class S is not really a recognised class, some authors mention it but most don't acknowledge it. It is basically the use of class D for RF amplifiers and not used in audio. The delta-sigma modulation bit is also used in class D audio amps, it's one "type" of class D or method of generating class D.
Class S was sometimes used at one time in relation to the work of Dr Aubrey Sandman, which soon after was kind of nicked by Technics and renamed "class AA". He was I believe unsuccessful in suing them unfortunately. Sandman amps are similar-ish to Quad current dumping BTW...

Class I and class T are also not really recognised classes but variations on class D and are proprietary names from Crown (class I) and Tripath (class T).

Classes C, E and F are only appropriate to RF ie to drive the aerial of a radio transmitter, not for audio.

The main methods of class D are

1/ A PWM generator fed from a fixed frequency oscillator.
2/ A free running oscillator circuit using positive feedback and in which both frequency and pulse width will change with signal. This is the first type used, as in the Sinclair design from the 60's and others before that. Much updated by Bruno Putzeys it is the technique behind the Hypex etc amps.
3/ Delta Sigma Modulation (or Sigma Delta. Same thing)
4/ Digital generation of the PWM by DSP of the digital input. Need A/D converter to use analogue input. Stilll ends up producing an analogue PWM signal at the output.

The difference between class G and class H is that in one, secondary tiers of output devices (more than 2 tiers in some big amps), connected to higher voltage rails, are switched in to encompass higher output power levels. In the other, only one tier of output devices is used, as in a normal amplifier, but the power supply itself is switched to higher voltage levels to encompass higher powers. The effect is much the same. Which is which? Good question. Ask a European or American or Japanese engineer and you will get different answers and even this may not have remained constant over the years... As the first use of this I'm aware of (although I've not researched it) was by Hitachi in their "Dynaharmony" series I'm tempted to say that class H may be for Hitachi... BUT the classes were normally just a case of the next available letter as new technology came about and that would put class G first:scratch:

[anthonyTD]

All good stuff Jez, :)
But when you get down to what matters, ie; the sound, folk will always have diffrences of opinion, and it will matter not a jot to many as to which topology is more accurate on paper, or on test, it will always come down to what sounds the best, or should I say, which has the ability to connect with you on an emotional level!

[Marco]

Exactamundo! ;)

For me, it takes a lot to beat a reasonably powerful Class A P/P valve (power) amp, used with say, a high-quality series passive, into big and efficient speakers :)

Marco.

[walpurgis]

Surely all audio amplifiers are 'Push Pull'? They'd be DC devices otherwise (regardless of A, AB or B class, etc)

The DC offset in pure Class A, does not stop it pushing and pulling. :)

[Marco]

Yeah, sure. I meant as opposed to a single-ended valve amp, as good as they can be when implemented well. I'd always choose a good P/P, as I prefer having access to a bit of power :)

Marco.

[walpurgis]

I'd not opt for single ended either. Too many drawbacks.

[Macca]

Is it not that the advantage of a SECA is the simplicity? Running the signal through fewer components must count for something?

[walpurgis]

It does, but to get any kind of useful output you start looking at a lot of weight, size, power needs & heat amongst other things.

[montesquieu]

Tom,

I am using a Class D over the summer here. It is the Nord NC500 and I have put together some views on the Wam.

In summary I have really disliked Class D so far including some expensive NC1200 amps from the US (the one User 211 had). All of them had this real flat undynamic sound, it was destailed and very neutral but otherwise 'snoresville'. But the Nords are different. Instead of using the standard £1.69 op amp input (LM4562) they offer a plug-in socket and you can use the Burson, Sparkos and Sonic Imagery 994. I preferred the 994s ($80) but the Sparkos do have a softer and warmer sound. I suspect they won't 'convert' people but those that are interested in Class D may find them better than other Class D apart from ultra expensives. It may be alternative and for a coffee or two and some biccies (I know I am tart) I could bring them round. :)


Hi George ... I think you are on! I'm sitting sweltering here and the last thing I want to do is switch on a 100w Class A/B KT90 valve amp!

Tons of food for thought in this thread and really appreciate the technical commentary Jez (and Anthony).

Be interesting to expand on what sonic differences might be expected from the various Class D topologies ....

Interesting about the Soul amps Anthony I wasn't aware of what you've written here. Does what you've done improve the damping factor (always an issue with your typical single ended valve amp)?

[walpurgis]

Tons of food for thought in this thread and really appreciate the technical commentary Jez (and Anthony).

I agree. I've really enjoyed reading the well informed input.

[danilo]

Dunno.. after having tried a fair number of Amps over the decades, up to and including a bespoke Se 45 Meshplate gizmo.
My Current F6 is simply unequalled.. in my ears.. period.
Which is Why I'll likely keep it until I'm dead.
Only other alternate I'd seriously consider would be a Pass Xs 150.. if I could afford it.
Or as a lark adventure, maybe this thing
http://www.neurochrome.com/modulus-86-rev-2-0/

[YNWaN]

Running the signal through fewer components must count for something?

That theory seems logical but I'm not sure it's really true in real life - or, at least, that the compromises accrued from using overly simple solutions are worse than a more superficially complex solution. In a lot of complex circuits the majority of the components are not in the signal path and even when it comes to speaker crossovers I'm not convinced that minimum componentry is best.



100% Analogue

[Marco]

'Minimal' must always be combined with 'best' - otherwise there's no point! ;)

The BEST and most well-implemented circuit, populated with the BEST components, minimised in use where possible and advantageous, in order to create the shortest and simplest signal path, should result in the BEST possible sound :)

Of course what constitutes as "BEST", sound-wise in that context, is almost entirely subjective.....

Marco.

[Macca]

the compromises accrued from using overly simple solutions are worse than a more superficially complex solution.



Yes, that can happen. Although I get the impression from this thread that, like speakers, all amplifier topologies seem to be compromises of one sort or another.

[anthonyTD]

Hi Tom,
Sorry for the late reply, First of all, glad to hear you find my ramblings here of interest, now,,, as for your question on the Soul amps concerning the damping factor, over conventional Single ended, the answer is yes, due to the type of circuit design, and to a large degree, the amount of feedback one can apply, the damping factor compared to conventional single ended is indeed improved, However, All valve amplifiers as far as damping factor figures are concerned, will only be able to acheive a fraction of what is posible with Solid State, this is due mainly to valves being high impedance devices, therefore; they need some form of matching device [output transformer] to be able to transfer their power into a relatively low impedance, ie; the speaker, Again however, when you look at the amount of damping that is actualy needed to control most well designed loudspeakers, [after the first inductor in series with the main speaker driver, Damping factors become less important] you will understand why Valve amplifiers can still be capable of extraordinary performance as far as sonics are concerned.:)

Hi George ... I think you are on! I'm sitting sweltering here and the last thing I want to do is switch on a 100w Class A/B KT90 valve amp!

Tons of food for thought in this thread and really appreciate the technical commentary Jez (and Anthony).

Be interesting to expand on what sonic differences might be expected from the various Class D topologies ....

Interesting about the Soul amps Anthony I wasn't aware of what you've written here. Does what you've done improve the damping factor (always an issue with your typical single ended valve amp)?

[anthonyTD]

As for the question on Class D, I will leave that to Jez, or anyone else here who has more experience than I with its diffrent design characteristics, I have built a few over the years, including tripath designs etc, but never felt the urge to want to investigate them further.:)

[sq225917]

I've always found class d lacking, even ncores. Ok for bass, where they excel. The best amp I've heard, not pushed hard was a turbo'd f5. Horses for courses.

[montesquieu]

I've always found class d lacking, even ncores. Ok for bass, where they excel. The best amp I've heard, not pushed hard was a turbo'd f5. Horses for courses.

Tried a factory F5 against the Radford - with Tannoys, an easy win for the 100w tube amp - but I can see its potential for less demanding speakers.

The Hypex stuff has always left me cold but some of the others do far better.

[Arkless Electronics]

If there is enough interest I will look into offering a class A push pull SS power amp of about 10WPC and possibly an integrated amp version... Let me know folks. If no interest appears to be out there for something like this I'll leave it at that...
I have no interest in producing anything class D BTW!

[RichB]

If there is enough interest I will look into offering a class A push pull SS power amp of about 10WPC and possibly an integrated amp version... Let me know folks. If no interest appears to be out there for something like this I'll leave it at that...
I have no interest in producing anything class D BTW!

What sort of price point are you looking around for the power amp Jez.

Happy to discuss via PM than take this thread off topic.

[Arkless Electronics]

What sort of price point are you looking around for the power amp Jez.

Happy to discuss via PM than take this thread off topic.

It's just me kind of thinking out loud and seeing if there is enough interest to take it further really Rich... I'll pm you.

[Mr. C]

I'm afraid your first statement is simply 100% wrong. The D in class D has sweet FA to do with digital and no it is not correct that only a digital input class D is true class D. Class D is analogue.

There is also no reason to avoid SMPS above and beyond any reasons for using/not using them with any other type of amplifier. As I explained much earlier a SMPS in effect IS a class D amplifier.

Bandwidth..... check out what I said about the output filter network being a Low Pass Filter... The top end of a class D will indeed drop like a brick beyond the cut off frequency of the filter. That's the idea. It has to in order to prevent huge amounts of radio interference. Take a portable radio somewhere near a class D amp, even 20 foot away... and tune through the bands and you will almost certainly find all sorts of strange whistles and squawks and usually even the music you are playing on the hi fi coming through the radio!
In effect you have a "passive crossover" at the output of class D amps and the properties of the inductors and capacitors will come into play...

As for the Devialet well it could be claimed it is a class A.... It depends how you want to classify your classes ("class distinction" as Doug Self once called it) and how much poetic license you give it ;)

Good Afternoon Jez

Now I have some time available I will respond.

Ok I was being overly simplistic in my generalization regarding 'D' class word semantics Lets stick with PWM switching amplifiers that take an analogue input signal, rather than those that merely using an pure PCM or I2s digital input for now.

Two of areas of limitation of of the PWM design you have identified with, switching frequency noise interference and the reconstructed analogue signal bandwidth is mere a low pass filter.

Some of the earlier pwm switching amplifiers (sub 500kHz bandwidth) which used a smps to power them attempted to reduce the switching interference by using smps which switch at or very close to the switching point of the amplifier thus theory would suggest they two would have a cancelling effect on each.

One well know brand's early integrated design was very quiet when in operation, yet if the device was placed in standby a very distinct buzzing noise (all be it quiet) could be heard emanating from the speakers in one channel if you placed your ear near the speaker!


Also the reconstructed analogue if viewed under a scope would show a very distinctive pattern as you could 'see' all of the artifacts sitting on top of the wave form see image below.


http://www.facstaff.bucknell.edu/esantane/movies/10kHz7v4-8ohms.JPG


Now can you actually hear the artifacts which are present in the above waveform screen shot?, No the are high enough outside the pass band not to have an audible affect imho.

In the last sixteen years I have been involved with PWM switching amplifiers I have found that a quality linear power supply produces far more real tangible sound that an equivalent specification SMPS of coarse your mileage my vary, however I am resolute in this approach, however I fully understand mass production and volume sales.

For a quality pwm amplifier then ultra fast, ultra quiet FET's are required along with suitable boot strapping and anti-ringing circuitry.

I also feel that 500kHz switching frequencies do not produce the most favorable results but a calculated trade off between power/size/sound quality/the use of smps and importantly for these days price.

Then this brings us back to as Jez's mention's the final stage in a pwm amplifiers construction a low pass filter (LC network) and the figures that work out to deliver the most stable (electrically) and desirable sonically

These seem to figure between 0.22uf capacitance and around 18uH for the inductors. With some designs a damping circuit (RC) to was required to prevent the FET's (field effect transistors) having a hard time due to peaking in the LC circuit so damping filter usually consisted of resistance between 15 and 33Ohms and capacitance of around 01.uf The above give a bandwidth of around 80kHz (fourth harmonic)

Again the quality of the components, board lay out, proximity of components next to other critical circuit parts etc all have quite a significant hand in how the final product sounds.

Coupled with a quality linear ultra stiff psu with high quality fast transient and low esr caps, quiet rectification (interestingly snubber diodes never really improved the sound perceptually imho) the correct switching frequency, very short signal pathways, correct FET selection, switching module supply voltages need to be ultra clean and very tightly regulated again a quite significant difference can be made in this area, which I feel is very critical to obtaining high quality sounds from a pwm amplifier.

The perimeters and considerations maybe different for that of a class 'a' or 'a/b' but the result can deliver in spades.

You just have to understand a different set of rules.

The NCore bandwidth is 50kHz -2/3dB most of the older larger powered ICE power modules are 36kHz -2/3dB now your are getting into naim's lack of bandwidth :stalks:

Anyway not all things are created equally just don't tar all PWM amplifiers with the same brush.

Oh the phone close proximity interference I would suggest that any amplifier that has a crazy enough bandwidth will produce rfi pick up from a phone try going close to a D'Agastino amp bandwidth 2.4gHz :lol:

[Arkless Electronics]

Good Afternoon Jez

Now I have some time available I will respond.

Ok I was being overly simplistic in my generalization regarding 'D' class word semantics Lets stick with PWM switching amplifiers that take an analogue input signal, rather than those that merely using an pure PCM or I2s digital input for now.

Two of areas of limitation of of the PWM design you have identified with, switching frequency noise interference and the reconstructed analogue signal bandwidth is mere a low pass filter.

Some of the earlier pwm switching amplifiers (sub 500kHz bandwidth) which used a smps to power them attempted to reduce the switching interference by using smps which switch at or very close to the switching point of the amplifier thus theory would suggest they two would have a cancelling effect on each.

One well know brand's early integrated design was very quiet when in operation, yet if the device was placed in standby a very distinct buzzing noise (all be it quiet) could be heard emanating from the speakers in one channel if you placed your ear near the speaker!


Also the reconstructed analogue if viewed under a scope would show a very distinctive pattern as you could 'see' all of the artifacts sitting on top of the wave form see image below.


http://www.facstaff.bucknell.edu/esantane/movies/10kHz7v4-8ohms.JPG


Now can you actually hear the artifacts which are present in the above waveform screen shot?, No the are high enough outside the pass band not to have an audible affect imho.

In the last sixteen years I have been involved with PWM switching amplifiers I have found that a quality linear power supply produces far more real tangible sound that an equivalent specification SMPS of coarse your mileage my vary, however I am resolute in this approach, however I fully understand mass production and volume sales.

For a quality pwm amplifier then ultra fast, ultra quiet FET's are required along with suitable boot strapping and anti-ringing circuitry.

I also feel that 500kHz switching frequencies do not produce the most favorable results but a calculated trade off between power/size/sound quality/the use of smps and importantly for these days price.

Then this brings us back to as Jez's mention's the final stage in a pwm amplifiers construction a low pass filter (LC network) and the figures that work out to deliver the most stable (electrically) and desirable sonically

These seem to figure between 0.22uf capacitance and around 18uH for the inductors. With some designs a damping circuit (RC) to was required to prevent the FET's (field effect transistors) having a hard time due to peaking in the LC circuit so damping filter usually consisted of resistance between 15 and 33Ohms and capacitance of around 01.uf The above give a bandwidth of around 80kHz (fourth harmonic)

Again the quality of the components, board lay out, proximity of components next to other critical circuit parts etc all have quite a significant hand in how the final product sounds.

Coupled with a quality linear ultra stiff psu with high quality fast transient and low esr caps, quiet rectification (interestingly snubber diodes never really improved the sound perceptually imho) the correct switching frequency, very short signal pathways, correct FET selection, switching module supply voltages need to be ultra clean and very tightly regulated again a quite significant difference can be made in this area, which I feel is very critical to obtaining high quality sounds from a pwm amplifier.

The perimeters and considerations maybe different for that of a class 'a' or 'a/b' but the result can deliver in spades.

You just have to understand a different set of rules.

The NCore bandwidth is 50kHz -2/3dB most of the older larger powered ICE power modules are 36kHz -2/3dB now your are getting into naim's lack of bandwidth :stalks:

Anyway not all things are created equally just don't tar all PWM amplifiers with the same brush.

Oh the phone close proximity interference I would suggest that any amplifier that has a crazy enough bandwidth will produce rfi pick up from a phone try going close to a D'Agastino amp bandwidth 2.4gHz :lol:

I'm in agreement with most of that! Thanks for posting the oscilloscope photo as that is a fine example to the layman of just how much crap comes out of a typical class D amp when producing a 1KHz sine wave...

The use of SWPS in the manner you describe can indeed be fraught with problems and I came across this myself when once involved with the design of industrial class D amps and SMPS which were used in quantity in 6' racks. We tried all sorts of experiments in synchronous switching and also "spread spectrum" but in the end I designed small class A?B modules with integral fan cooling and this gave better all round results when RFI etc was taken into account.
However, suitable filtering and reservoir caps on a good SMPS should render it on a par to a linear supply for a class D power amp.

[Mr. C]

:
I'm in agreement with most of that!.


:stalks::stalks::stalks::comatose:

[walpurgis]

Thanks for posting the oscilloscope photo as that is a fine example to the layman of just how much crap comes out of a typical class D amp when producing a 1KHz sine wave....

Is it crap Jez? If that signal is 1kHz, I'd guess by the look of it, the carrier wave is around 35kHz. Way above the audio spectrum. Would it intrude?

[Clive]

Is it crap Jez? If that signal is 1kHz, I'd guess by the look of it, the carrier wave is around 35kHz. Way above the audio spectrum. Would it intrude?
That's not even an octave away from being audible. If there are harmonics working their way downwards then it's bad news.

[walpurgis]

From what I can see (not much) the carrier looks clean. If that's the case, harmonics if any, would be of very small amplitude and not be harmonics of the 1kHz signal in any case.

[Arkless Electronics]

Is it crap Jez? If that signal is 1kHz, I'd guess by the look of it, the carrier wave is around 35kHz. Way above the audio spectrum. Would it intrude?

Maybe... maybe not.... I don't like it myself though, it's intellectually dissatisfying at very least! Depending on the frequency and on the bandwidth of the source etc it can easily be seen that yes it could be a problem by intermodulating with the source signal....
It's more my amusement at people going to lengths to try and avoid RF etc on mains supplies, that is of less importance, and by using foo cables which won't even have any effect... and will then tolerate this!! Bloody laymen :D

[walpurgis]

What about cables that do have an effect? :D

Anyway. I'm a layman, I don't build amps. But I do like to question things. I stick with Class A though. 'Cos I know how it works and like the principle and the sound.

[Mr. C]

I have been very fortunate in the last decade or so to be able to own some of the most desirable power amplifiers every made, strange how this layman :lol: still has not changed from a noise inducing / fri spewing pwm switching amplifier all these years.

Must have cloth ears and suck 100watts of class 'a' valves for breakfast ;)

Anyway good thread chaps

[Arkless Electronics]

What about cables that do have an effect? :D

Anyway. I'm a layman, I don't build amps. But I do like to question things. I stick with Class A though. 'Cos I know how it works and like the principle and the sound.

Mains cables? There aren't any!

The point I'm making is that here is a major interference source, highly visible on a 'scope, measurable etc and yet audiophiles haven't been told to obsess about this by magazines, forums etc and so most probably had no idea it was even present.... There yer go foomeisters, a whole new range of special cables for class D amps to be exploited... you heard it first here:eyebrows:
And yet there will daily be postings on this and other forums from people thinking they can improve their sound with fancy mains cables etc... attacking a non existent problem with a solution that has no effect!
And on that bombshell.... That ends my participation in that particular area as the circular arguments will be whirring into life already and people who haven't understood a single word of this thread will be honing their 6 Moons vocabulary in praise of Russ Abbot... er.. Andrews :ner:

[walpurgis]

Just a thought. If Class D switches at frequencies above 200kHz why is that carrier wave around 35kHz? Or is this not from a Class D or even maybe Class T amp, which I believe switch slightly lower?

Am I being Dense? I thought switching frequencies were fixed and it was just the frequency of the 'carried' signal and its amplitude that varied.

[Macca]

Speaking of power cables saw this elsewhere:

All power cords, no matter which brand or type require about the same amount of time to burn in. We burn in power cords using a fan as a load and run it for 5-7 days of continuous current 24/7. During this period of time the cable will go through dramatic changes in sonic character. So judging the sound quality of a cord during the burn in period will not give you an accurate assessment of the sound quality of the cord.

The cable will continue to improve after the initial 5-7 days burn in and will completely stabilize around the third to fourth week on normal use.

Some people have reported that they did not notice any burn in effects with their power cord. This either means that the the person is not a careful listener or that their system or room acoustics need some work or that the quality of the power cord is not sufficient to hear burn in effects."

and this:

We've seen a power cord and nothing else be responsible for a loss of power amounting to about 40 watts in one of our amps. All I did there was measure the voltage drop across the power cord and I found the culprit. I did that with a 3 1/2 digit DVM. That kind of loss being pretty measurable I would expect it to be audible too, and it is. '

Wondered what you would make of it, Jez. The first quote sounds like bollocks to me but is the scenario on the second quote actually possible?

[walpurgis]

Bet those fans sound great after the 5-7 days! :lol:

P.S. Don't forget to demagnetise and polish the mains plug pins. ;)

[Arkless Electronics]

Just a thought. If Class D switches at frequencies above 200kHz why is that carrier wave around 35kHz? Or is this not from a Class D or even maybe Class T amp, which I believe switch slightly lower?

Am I being Dense? I thought switching frequencies were fixed and it was just the frequency of the 'carried' signal and its amplitude that varied.

Switching frequency is usually fixed but in some designs, especially "free running" amps, it can vary. Choice of switching frequency effects everything else and is a very important first step when designing a class D amp. Usually frequencies far above 35KHz will be chosen and in fact a class D amp working at 35KHz could only go up to 17.5KHz absolute maximum in terms of its HF response. We don't know what the actual frequency being displayed on that oscillograph is and it could be 20KHz not 1KHz! frequencies between about 50KHz (very low) and 1-2MHz are usually used in practise and as is normal with engineering, everything is a compromise... certain things get better with increasing frequency and vice versa.

[walpurgis]

is the scenario on the second quote actually possible?

No. Not unless there's a short or you're using 1000 metres of coiled thin bell wire as a mains lead (which would heat up and blow anyway).

[walpurgis]

Switching frequency is usually fixed but in some designs, especially "free running" amps, it can vary. Choice of switching frequency effects everything else and is a very important first step when designing a class D amp. Usually frequencies far above 35KHz will be chosen and in fact a class D amp working at 35KHz could only go up to 17.5KHz absolute maximum in terms of its HF response. We don't know what the actual frequency being displayed on that oscillograph is and it could be 20KHz not 1KHz! frequencies between about 50KHz (very low) and 1-2MHz are usually used in practise and as is normal with engineering, everything is a compromise... certain things get better with increasing frequency and vice versa.

Thanks Jez. I'd not considered that the lower the switching frequency, the lower the upper limit of audio frequencies would be. But thinking about that it makes perfect sense. The carrier ceasing to be able to carry a complete audio wave. I get it.

[Arkless Electronics]

Speaking of power cables saw this elsewhere:

All power cords, no matter which brand or type require about the same amount of time to burn in. We burn in power cords using a fan as a load and run it for 5-7 days of continuous current 24/7. During this period of time the cable will go through dramatic changes in sonic character. So judging the sound quality of a cord during the burn in period will not give you an accurate assessment of the sound quality of the cord.

The cable will continue to improve after the initial 5-7 days burn in and will completely stabilize around the third to fourth week on normal use.

Some people have reported that they did not notice any burn in effects with their power cord. This either means that the the person is not a careful listener or that their system or room acoustics need some work or that the quality of the power cord is not sufficient to hear burn in effects."

and this:

We've seen a power cord and nothing else be responsible for a loss of power amounting to about 40 watts in one of our amps. All I did there was measure the voltage drop across the power cord and I found the culprit. I did that with a 3 1/2 digit DVM. That kind of loss being pretty measurable I would expect it to be audible too, and it is. '

Wondered what you would make of it, Jez. The first quote sounds like bollocks to me but is the scenario on the second quote actually possible?

The first quote is bollocks of the highest order!
The second... well under certain conditions it could be true but it is totally irrelevant to hi fi. IF one of the 2KWPC and above PA amps available were to be used with a long mains cable, say 20 yards, and if it was on the thin side, then it would probably be possible to loose 40W off the 4KW total due purely to resistive losses in the cable. It would be much more likely with USA 115V supplies. It would be unlikely to be audible and the type/make/brand of cable would make no difference if they were all of the same rating.
Edit: In the case of an amp powered from a SMPS with "International" voltage compliance the SMPS would compensate for the resistive loss of the cable and there would be no power loss from the amp.

[Arkless Electronics]

Thanks Jez. I'd not considered that the lower the switching frequency, the lower the upper limit of audio frequencies would be. But thinking about that it makes perfect sense. The carrier ceasing to be able to carry a complete audio wave. I get it.

Look up "Nyquist Rate" ;)

[anthonyTD]

:)


The point I'm making is that here is a major interference source, highly visible on a 'scope, measurable etc and yet audiophiles haven't been told to obsess about this by magazines, forums etc and so most probably had no idea it was even present.... There yer go foomeisters, a whole new range of special cables for class D amps to be exploited... you heard it first here:eyebrows:
And yet there will daily be postings on this and other forums from people thinking they can improve their sound with fancy mains cables etc... attacking a non existent problem with a solution that has no effect!
And on that bombshell.... That ends my participation in that particular area as the circular arguments will be whirring into life already and people who haven't understood a single word of this thread will be honing their 6 Moons vocabulary in praise of Russ Abbot... er.. Andrews :ner:

[anthonyTD]

Ahh but desireable for what reason,:eyebrows: is it posible that none were paticularly good at playing music,and that's why you stuck with your class D.:D
Just kidding,mate.:)
I have been very fortunate in the last decade or so to be able to own some of the most desirable power amplifiers every made, strange how this layman :lol: still has not changed from a noise inducing / fri spewing pwm switching amplifier all these years.

Must have cloth ears and suck 100watts of class 'a' valves for breakfast ;)

Anyway good thread chaps

[Mr. C]

In my experience and in practice (dependent on power outputs) I find that between 1.38Mhz and 1.72Mhz produce a more pleasing sound to my ears.

Later I will take some scope images of 1Khz wave form on three different PWM amps for you and one from a decent class a/b

[Mr. C]

Hello Tony

This is one of those threads that polarize opinion to a great degree, like analogue / digital etc.

For me the biggest bugbear of PWM switching amplifiers is their total lack of soul or a engaging quality that makes you wish to listen to more music.

A good class 'd' exhibits usually exhibits quality bass control, very quiet background, detail, dynamics and a real solid sound which a great many mistake for a three dimensional sound.
?
The lack real textural layering, delicacy, tonal quality and any innate musicality imho. Devialet? come on, a thinking man's B & O in a very beautiful case sorry a Swiss army knife electronic device for Trendy Wendy's, Apple people and ex Krell owners who kid themselves the a pair of 800's is far superior to any other amplifier around. Of coarse YMMV

I would suggest that each amplifier class has its own pro's and con's and each individual identifies with some of the traits that each class shows, no more or less.

If you find the right PWM switcher it really can captivate you for the right reasons, not the WOW reasons, in the same way a pair of Gaku on's, Tron's, Pass labs, Dartzeel, TAD and Gryphon can (other amplifier brands are available!)

You pay your money and make your choice. (or like a few of us make your own)

[Macca]

Hello Tony

This is one of those threads that polarize opinion to a great degree, like analogue / digital etc.

For me the biggest bugbear of PWM switching amplifiers is their total lack of soul or a engaging quality that makes you wish to listen to more music.

A)

I don't understand this. The purpose of an amplifier is to boost the signal to a level that the loudspeakers can use it, whilst adding as little as possible in the way of distortion or noise, Surely it is for the music to have soul, not the kit?

[walpurgis]

Look up "Nyquist Rate" ;)

Got it now. Fits the picture of the circumstances I had in my head.

[Mr. C]

Martin

I'll rephrase it for you.

The ability to produce the music being played in a manner which the original artist, composer, group or orchestra layed down the original recording with feeling, passion and emotion.

This is not the same as a great sounding recording which can demonstrate a matter of fact-ness, though sometimes you are lucky and can have both.

A great many Class 'd' amplifiers demonstrate a clinical, sterile or monochromatic presentation some even call this have a 'cold' sound.

I would further that a most respectable class 'd' are not hard or grainy, however if you are used to certain euphoric valve tendencies you may use this word possibly.

[Macca]

Martin

I'll rephrase it for you.

The ability to produce the music being played in a manner which the original artist, composer, group or orchestra layed down the original recording with feeling, passion and emotion.

This is not the same as a great sounding recording which can demonstrate a matter of fact-ness, though sometimes you are lucky and can have both.

A great many Class 'd' amplifiers demonstrate a clinical, sterile or monochromatic presentation some even call this have a 'cold' sound.

I would further that a most respectable class 'd' are not hard or grainy, however if you are used to certain euphoric valve tendencies you may use this word possibly.

A 'cold' sound I am not sure about. But is not hard and grainy just distortion making itself known? I'm sort of of the opinion that any amp of any type that is in working order and measures within acceptable parameters will sound fine as long as there is no mis-match with the speakers or it is being clipped. And the last two happen a lot more than is commonly thought. When I say 'fine' I mean it will not sound grainy or hard. Not 'perfect' or even 'good', but acceptable.

[anthonyTD]

Only just read this, is this relating to my last post Tony ? :scratch:
Hello Tony

This is one of those threads that polarize opinion to a great degree, like analogue / digital etc.

For me the biggest bugbear of PWM switching amplifiers is their total lack of soul or a engaging quality that makes you wish to listen to more music.

A good class 'd' exhibits usually exhibits quality bass control, very quiet background, detail, dynamics and a real solid sound which a great many mistake for a three dimensional sound.
?
The lack real textural layering, delicacy, tonal quality and any innate musicality imho. Devialet? come on, a thinking man's B & O in a very beautiful case sorry a Swiss army knife electronic device for Trendy Wendy's, Apple people and ex Krell owners who kid themselves the a pair of 800's is far superior to any other amplifier around. Of coarse YMMV

I would suggest that each amplifier class has its own pro's and con's and each individual identifies with some of the traits that each class shows, no more or less.

If you find the right PWM switcher it really can captivate you for the right reasons, not the WOW reasons, in the same way a pair of Gaku on's, Tron's, Pass labs, Dartzeel, TAD and Gryphon can (other amplifier brands are available!)

You pay your money and make your choice. (or like a few of us make your own)

[Clive]

A 'cold' sound I am not sure about. But is not hard and grainy just distortion making itself known? I'm sort of of the opinion that any amp of any type that is in working order and measures within acceptable parameters will sound fine as long as there is no mis-match with the speakers or it is being clipped. And the last two happen a lot more than is commonly thought. When I say 'fine' I mean it will not sound grainy or hard. Not 'perfect' or even 'good', but acceptable.
I get the "cold" sound comment. I alternate between 4 amps - valve SE, SS SE, current dumping and class D. Each has it's own signature which is apparent with my rather revealing speakers. With other speakers of more typical sensitivity and power requirements I find there's less to choose between the amps and indeed the SE ones tend to struggle (loose dynamics). With the revealing speakers the class D is very similar to the currently dumper but there's something that draws me into the music more than with the class D. Class D has tremendously fast 'start/stop' which with some speakers can be too much of good thing (when splitting hairs for comparison's sake). With more a typical speaker the class D is a no brainer due to cost and the speakers not being so revealing that it then sounds as good as other far more expensive amps.

[YNWaN]

Only just read this, is this relating to my last post Tony ? :scratch:

Looks rather like Tony is responding to himself.


100% Analogue

[Macca]

Looks rather like Tony is responding to himself.


100% Analogue

Like some sort of Platonic dialogue.

[Mr. C]

Mark

I was replying to Anthony, to many Tony's around here :eek:

[anthonyTD]

Ahh, right, thought so, no worries! :D
Mark

I was replying to Anthony, to many Tony's around here :eek:

[George47]

Not sure what Class D amp was shown.

For nCore modules measurements by JA in Stereophile, whilst using a low pass filter, showed signal to noise ratios of greater than 100db. The low pass filter only filtered signals out that were greater than 200kHz. The frequency response was essentially flat for 2,4 and 8 ohms with little roll off up to 20 kHz at least. The frequency of the residual was at 453kHz. Heard anything at 453 kHz?

Class D has issues that people have been working on and some like Bruno Putzeys have had great success. Getting the right balance in all amps ain't easy and I am sure there were some real shockers in Class A, valve and SS in the past. Same with Class D. So can we not highlight today's issues rather than what has happened in the past even (in the case of Class D) the reasonably recent past?

[r100]

Does any of you gentlemen own one of theses amps from Nordacoustics ? It's a UK based company and they use the OEM ncore 500 modules from hypex ... http://www.nordacoustics.co.uk

[Yomanze]

I would have liked to be able to have a listen to these, but nothing in the UK. Seems like an audio engineer with an RF background is the right sort to work with class D. http://ar-t.co/VELLUTO_CIRCUIT_DESCRIPTION.html

[montesquieu]

I would have liked to be able to have a listen to these, but nothing in the UK. Seems like an audio engineer with an RF background is the right sort to work with class D. http://ar-t.co/VELLUTO_CIRCUIT_DESCRIPTION.html

Kind of bollocks to quote power into 2.7 ohm load though. Oh hold on I have some 2.7 ohm speakers in the garage .... (not). Useful description of the implementation though.

I'm currently pondering the Nuforce sta-200 I have a suspicion it could be quite sympatico with Tannoys though i could do with something with lower sensitivity.

[struth]

Ive got a set close to that lol

[Yomanze]

Kind of bollocks to quote power into 2.7 ohm load though. Oh hold on I have some 2.7 ohm speakers in the garage .... (not). Useful description of the implementation though.

I'm currently pondering the Nuforce sta-200 I have a suspicion it could be quite sympatico with Tannoys though i could do with something with lower sensitivity.

Maybe it's to show that the thing works well into such a low load unlike other Class D amps?

[montesquieu]

Maybe it's to show that the thing works well into such a low load unlike other Class D amps?

Fair point.

[Arkless Electronics]

It's as relevant as any other load as many speakers can dip that low...
Class D does not have ant specific issue with low load impedances....

[r100]

I would have liked to be able to have a listen to these, but nothing in the UK. Seems like an audio engineer with an RF background is the right sort to work with class D. http://ar-t.co/VELLUTO_CIRCUIT_DESCRIPTION.html

This IS a Gloucestershire, UK based company. They assemble the OEM version of the nCore 500modules (as opposed to the DIY nCore 400) with some refinements (http://www.nordacoustics.co.uk/nordoneuppoweramplifiers). These are not your average class D amp and I would love to know from someone who actually has one if it is as good as they say. Maybe you could have a listen at their premises. And BTW, they can drive 2 ohm speakers http://www.hypex.nl/technology/ncore.html @ 1200W :eek:

[struth]

my class d(at least I think they are lol) can handle my under 3 ohm speakers fine. Never had any problems when they were being run by it.

[spendorman]

my class d(at least I think they are lol) can handle my under 3 ohm speakers fine. Never had any problems when they were being run by it.

Size of heatsinks may give a clue to whether they are Class D or not. Perhaps some photos?

When I got my Class D TPA3116, it was hard to take in that the tiny heatsink was suitable for a 50W / channel amp, but it is. These amps are not that far short of 100% efficient.

[struth]

Size of heatsinks may give a clue to whether they are Class D or not. Perhaps some photos?

When I got my Class D TPA3116, it was hard to take in that the tiny heatsink was suitable for a 50W / channel amp, but it is. These amps are not that far short of 100% efficient.

Not opened one yet. Not sure how actually lol. Ive got a faulty one so must get round to it.

[The Alchemist]

In reply to "r100" I do not yet own the Nordacoustics nc500 amplifiers but I would like to! Colin North is a fellow IPL speaker builder and when I discovered that he was starting to build Hypex modules I had to go and have a listen for myself since I have been interested in the concept of the Hypex DIY kit for many years.
I went for an audition last year and have to say that I was mightily impressed with what he was doing and how good the amps sounded. I thought that the bass "grip" was exceptionably tight and as soon as I sell enough speakers I will be getting a Dual Mono SE version for myself.

Don't confuse these amps with the DIY modules that can be bought from Hypex. The OEM versions that Colin uses are different inasmuch as they utilise a buffer board with an OP amp which can easily be changed to allow a degree of choice as to how the amp sounds. At the moment he is using Sparkos or Sonic Imagery which make the amps much more "musical". I believe that Colin's amps offer fantastic value for money and if you go ahead and buy one I doubt that you would be disappointed!

[r100]

Very interesting ! Actually, I was quite surprised and heartened to learn that IPL was involved in this venture... their speakers are renowned and liked, so I deduct that the amps they put together at Nord are of equal quality. <thumbs-up icon here>

[The Alchemist]

Please don't misunderstand me. IPL has nothing to do with Nordacoustics other than to say that Colin buys the kits from Ivan and sells a finished speaker, the same as me. I know that Ivan is impressed with the Nord amps (he mentions them on his website) but has no other connection. The common thread between IPL and Nordacoustics is that both produce products which are towards the upper end of the spectrum for a very reasonable cost!

[r100]

Please don't misunderstand me. IPL has nothing to do with Nordacoustics other than to say that Colin buys the kits from Ivan and sells a finished speaker, the same as me. I know that Ivan is impressed with the Nord amps (he mentions them on his website) but has no other connection. The common thread between IPL and Nordacoustics is that both produce products which are towards the upper end of the spectrum for a very reasonable cost!

not familiar with the ins and outs but there does seem to be some love between the two companies ... see http://www.iqspeakers.co.uk/speakers which is positive, me thinks

[Yomanze]

Fair point.
I did a bit of research, looks like a cut and paste from B&O Icepower 250 specs.

[George47]

If anyone is interested I did a review of the Nord stereo power amps.

They are very good....http://www.hifiwigwam.com/nord-one-up-500-dm-st-dual-mono-stereo-amplifier-review/

The amps use the NC500 modules with individual SMPS. Colin uses discrete regulators and has put a socket on the board to allow the cheap op-amps to be swapped out. I found the Sonic Imagery op amps suited me best, others prefer the Burson. And as it is made on the UK the price is more palatable than an import. Acoustic Imagery make a similar mono version but without the discrete regulators or the ability to swap the op amps.

[SLS]

Having installed a Devialet Expert Pro 250 yesterday I might be biased. It employed an A/D hybrid system. It just does everything well in a very understated way. What was particularly impressive was how much better it is compared to the previous Le 250 version.

It does some sounds, in particular kettle drum, with spectacular realism, that is astounding. I use Harbeth SHL5+ that use an 8" mid-bass driver, proving to me at least that massive speakers are not required for a massive sound.

[Sherwood]

I can strongly recommend the Nuprime STA-9 power amps! I am currently using a pair to drive Magneplanar 1.7 speakers. I am speaking as an owner of a good quality valve power amp (Art Audio Quintet) and several other good solid state units. I have come to the STA-9s via an earlier Nuforce amp which unfortunately expired on me whilst I was working in Northern Vietnam and which was not cost effective to repair (I still have it in a cupboard somewhere). The Nuforce was extremely good and able to drive my earlier Magneplanar MG12s very well. However, the Nuprime is significantly better than the Nuforce on every dimension: detail, dynamics, imaging, transient response. Bass is both deep and tuneful with timbres resolved superbly and vocals sound very natural. On first encounter it might seem to some to be a little rolled off in the treble, but extended listening has convinced me that this is not the case: it is just very smooth, rather valve like. I did try using a single STA-9 with my Maggies and it acquitted itself extremely well. Impressive at this price point.

I suspect that many of those who decry digital amps, have not heard the best designs. Yes, some of the budget Tripath designs are excellent for the price (i have a few myself) but they are not in the same league as the new generation of digital and PWM amps.

Geoff

[agk]

I thoroughly enjoyed the couple I've owned.

It seems that op amps are the new valves.

[montesquieu]

I can strongly recommend the Nuprime STA-9 power amps! I am currently using a pair to drive Magneplanar 1.7 speakers. I am speaking as an owner of a good quality valve power amp (Art Audio Quintet) and several other good solid state units. I have come to the STA-9s via an earlier Nuforce amp which unfortunately expired on me whilst I was working in Northern Vietnam and which was not cost effective to repair (I still have it in a cupboard somewhere). The Nuforce was extremely good and able to drive my earlier Magneplanar MG12s very well. However, the Nuprime is significantly better than the Nuforce on every dimension: detail, dynamics, imaging, transient response. Bass is both deep and tuneful with timbres resolved superbly and vocals sound very natural. On first encounter it might seem to some to be a little rolled off in the treble, but extended listening has convinced me that this is not the case: it is just very smooth, rather valve like. I did try using a single STA-9 with my Maggies and it acquitted itself extremely well. Impressive at this price point.

I suspect that many of those who decry digital amps, have not heard the best designs. Yes, some of the budget Tripath designs are excellent for the price (i have a few myself) but they are not in the same league as the new generation of digital and PWM amps.

Geoff

Funny you should say that - look what just turned up today - Nuprime STA-9. Just one as I wanted to get a feel for it though if I like it I might well buy another and monoblock (290w per Tannoy should be quite entertaining). The plan is to tide me over till I get the big Radford back from restoration.

This is Class A (JFET) into Class D (though there seem to be some particularly analogue methods at work in the implementation). I've only had 30mins with it so far and I'm rather impressed, bass is very solid, highs clean enough, not sure about midrange warmth, inner detail or presentation of space vs the Radford but I'll let the thing warm up and play in first - it's fresh new out of the box.

http://i1181.photobucket.com/albums/x440/montesquieu61/IMG_0609_zpsd5bgmsz9.jpg

[Firebottle]

Quite a contrast in size there Tom, looking forward to your further impressions :)

[Sherwood]

Funny you should say that - look what just turned up today - Nuprime STA-9. Just one as I wanted to get a feel for it though if I like it I might well buy another and monoblock (290w per Tannoy should be quite entertaining). The plan is to tide me over till I get the big Radford back from restoration.

This is Class A (JFET) into Class D (though there seem to be some particularly analogue methods at work in the implementation). I've only had 30mins with it so far and I'm rather impressed, bass is very solid, highs clean enough, not sure about midrange warmth, inner detail or presentation of space vs the Radford but I'll let the thing warm up and play in first - it's fresh new out of the box.

http://i1181.photobucket.com/albums/x440/montesquieu61/IMG_0609_zpsd5bgmsz9.jpg

Tom,

I am not entirely convinced about burn in with electronics (it certainly does apply to speakers) but it took a couple of days for my STA-9s to "bed in".

I am guessing that the single STA-9 in stereo mode will be a fantastic match to your Tannoys. I tried one of mine in that configuration with my single driver Alnico speakers and the bass extension and slam was impressive (try Rickie Lee Jones "Under the Boardwalk").

However, the sound in mono mode is somewhat different. Obviously, the extra power comes into play, but the quality of the sound changes too. Musical timbres are easy to discern and the unit just sounds more valve like. I am not qualified to speculate on the technical reasons for the change but switching back and forth a couple of times confirmed the difference. What I have also discovered is the difference that balanced XLR inputs make. The immediate difference is the increase in loudness for any given volume setting. Other than that everything just seems a little more etched, more dynamic, and the already incredible (low) noise floor drops even further. I now plan to buy longer XLRs and switch to short speaker cables (<0.5m) with the STA-9s underneath the Maggies. For that I need to design and manufacture an alternative stand for the Maggies that will raise them by about 6inches and angle them a few degrees backward. Any ideas from the engineeringly gifted out there ........

Geoff