Going (semi) active....

[YNWaN]

I hope this is in the correct part of the forum and even if it is I admit that I was in two minds whether to post this as it's a bit 'specific' and, I suspect, outside the interest of many AoS readers - but what the hell, here goes ;);

I've just converted my speakers to run in a semi-active configuration. The speakers in question (Yamaha NS-1000Ms) are a three way and the mid and tweeter are driven through a passive crossover but the bass drivers have their own amp and an active crossover. To achieve this I purchased a Lab.gruppen IPD 1200 amp (a manufacturer of amps for the profesional audio market); this amp is a Class D design (300wpc into 8Ohms and 600 into 4) with a digital front end that allows extensive crossover mapping and room correction, plus a bunch of other stuff like adding delay and phase reversal. The IPD 1200 also has input (and output) level adjustment and was easily tailored to produce the same output level as my regular amps. All the amp adjustments are made via a very convenient bit of software you download to your computer, or via an iPad.

I've only had it all working a couple of days but I must say I'm very impressed - very impressed indeed!

http://i62.photobucket.com/albums/h90/markemark_2006/Hi-Fi/02C031B3-9200-4552-9D51-F114306E597F_zpsvixhgx4g.jpg

http://i62.photobucket.com/albums/h90/markemark_2006/Hi-Fi/318B310F-26B1-4903-80CC-9EC95EFB3405_zpshxued6lu.jpg

The squiggly cable is the CAT5 connecting the IPD 1200 to my Netbook whilst I fiddle with the settings :).

http://i62.photobucket.com/albums/h90/markemark_2006/Hi-Fi/45578DD5-8DCC-4883-A475-5583EDC1C33E_zpszehnoclh.jpg

http://i62.photobucket.com/albums/h90/markemark_2006/LabGruppensystemstructure_zps0c29865b.jpg

I need to measure the mid and bass outputs properly and no room correction has been applied, but the L.g really does grip those bass drivers!

A similar sort of thing could also be accomplished with a MiniDSP module and a separate amp.

[Macca]

Lots of AoS members will be interested in this, Mark. As am I. What are the subjective benefits, then?

[wee tee cee]

Marvellous...let us know how you get on!!!

[YNWaN]

Well, there are bass benefits, greater texture, a more visceral quality - dynamic range is increased. The stunt bass on the second side of The XXs first album is astonishing! There are other benefits too - complex or very dense passages of music sound less like instruments are vying for the same space - the system doesn't sound strained in any way (I think it's actually impossible to clip the amplification now as the speakers would give up first). There are advantages in the mid and top end too, everything is just a touch more delicate and nuanced. The whole sounds very low distortion - high levels can be achieved without discomfort but, at the same time, lower level listening is also rewarding.

On it's own, used full range, the L.g is a bit dry and Teutonic - makes the Naim amps sound like valve amps in comparison; but used like this the sound is much more layered and airy. Sonically it blends very well with the Naim amp.

[YNWaN]

My friend Dave has a proper measuring microphone and we can do frequency sweeps of the drivers and set the crossover accurately - just need him to be available...

The room correction is adjusted in a separate part of the software (although, in reality, adjusting one is effectively adjusting the other). One advantage of doing the crossover digitally is that the crossover curves are very true.

[awkwardbydesign]

I have just gone from 3 way active (all analog, 24dB octave slopes) to 3 way passive, because my wife found it too complicated. In the 80s when there was an organisation (ALSO http://www.whathifi.com/news/obituary-richard-hay-1944-2012) promoting active crossovers, someone described the sound of passives as "mud". I'm afraid I have to agree. I know clever designers get passives to sound fantastic, but having nothing between the amps and the drivers does make a difference. And a big SS amp for bass, 845 SET for mid and EL84 PP for treble means you can get the best from each amp, too. I wish I could go back, I even have the crossover, but it seems unlikely.

[Puffin]

I have thought about going this route before as passive bass is likely to be the weak point in any set up. Googling tells me that the amp is £900?

[synsei]

Mark (Reid Malenfant) formally of this parish ran his NS1000's fully active to very good effect, I've heard his system several times and it is mightily impressive.

[Ali Tait]

Interesting stuff. Lot to be said for going active IMHO. I use a miniDSP to run my OB's actively, sounds great to me.

[walpurgis]

Nice scheme hope it continues to impress. I reckon those Yams must really kick rigged up that way. Are those the original bass drivers? They look different somehow.

[YNWaN]

I have thought about going this route before as passive bass is likely to be the weak point in any set up. Googling tells me that the amp is £900?

Yes, I think it does retail for about £900. A friend has the Lab.gruppen account so I was lucky enough not to pay RRP. lab.gruppen are (apparently) a premium brand in the pro amp world and you can achieve a similar thing for quite a bit less if you choose Behringer etc (not quite such a 'joined up' solution though).

If I had three IPD 1200s I could drive the speakers fully active (900wpc 8Ohms!) - but experimentation so far suggests that this wouldn't actually sound as good.

There a number of high-end speakers that adopt a similar solution to this - active bass (often class D) and passive mid and tremble.

The system isn't more complicated to use than it was before though - in fact it's exactly the same. Having said that, my system is hardly a minimalist solution (you should see the amount of cable that goes to each speaker!).

[YNWaN]

Nice scheme hope it continues to impress. I reckon those Yams must really kick rigged up that way. Are those the original bass drivers? They look different somehow.

Yeah, they are the original bass drivers. The baskets have been sprayed black though and they don't have the grills over them.

They do indeed 'kick ass' :).

[Mika K]

My friend here in Finland had one biamped NS-1000M system with upgraded passive xo made from HQ parts and serious amps. That system was already quite remarkable performer and always though that to be my route with my NS-1000 domestics. Well now have been thinking whether to do that in future for the JBLs I have as that seems to be the ultimate solution for these also if you believe the more experienced members of the Altec-Lansing community.. :)

[Wakefield Turntables]

Mark,

Good look with your experimentation. I'm fully active and I dont think I could go back now unless I had to. I look forward to future posts.

Good luck! :)

[kininigin]

Mark (Reid Malenfant) formally of this parish ran his NS1000's fully active to very good effect, I've heard his system several times and it is mightily impressive.

I always wanted to hear Mark's system.He had a crazy amount of amplification if i remember correctly! Would be right up my street i suspect!

Don't think i could ever go back to passives.Having said that,i've not exactly had any top notch kit in that department! Or even heard any to be honest! I do like what actively driven speakers do though (when implemented well).If i had the money and time,getting some NS1000's and turning them active would be something i would love to try out!

[YNWaN]

My friend here in Finland had one biamped NS-1000M system with upgraded passive xo made from HQ parts and serious amps.

I've done a pretty extensive rebuild of mine too - crossovers included :):

http://i62.photobucket.com/albums/h90/markemark_2006/Yamaha%20NS-1000M%20Refurbishment/baseandframeandcrossover_zps2fb80144.jpg

[aquapiranha]

Great stuff!

[YNWaN]

Mark,

Good look with your experimentation. I'm fully active and I dont think I could go back now unless I had to. I look forward to future posts.

Good luck! :)

Thanks Andrew :). All I really have to do now is tweak the crossover slopes and rollover frequency - they aren't far off now as Troels Gravesen has already published many of the NS-1000Ms parameters. Then a bit of room correction to counter the major room modes :).

I tell you what though, I've heard a lot of high-end systems and this is one of the most resolved and most complete sounds I've ever heard!

[YNWaN]

Funnily enough I've just seen this:

http://3.bp.blogspot.com/-cxotYWgAr1w/U3M1nztvjBI/AAAAAAAAQ7k/ETLW1m42tyA/s1600/tess.jpg

It's the statement Tesseract speaker by Avalon - the woofers are powered by inbuilt amps and the mid and tweeter have a passive crossover :).

[Macca]

So you don't have any trepedation about running the bass signal through an AD/DA loop? Not that it would bother me but I thought you were a bit of a purist? ;)

The other thing that surprises me is that the Naim amps are a good match for the NS1000M. Well okay doesn't surprise me but it is interesting. I would have thought they would prefer more forgiving amps.

[Barry]

Funnily enough I've just seen this:

http://3.bp.blogspot.com/-cxotYWgAr1w/U3M1nztvjBI/AAAAAAAAQ7k/ETLW1m42tyA/s1600/tess.jpg

It's the statement Tesseract speaker by Avalon - the woofers are powered by inbuilt amps and the mid and tweeter have a passive crossover :).

Looks like it was designed for Wilson Audio by someone who is a closet member of the Ku Klux Klan.

[Macca]

LOL - yes I wonder if they do a version in white?

[Macca]

Burning Cross Loudspeakers present 'The Klansman' ;)

[YNWaN]

So you don't have any trepedation about running the bass signal through an AD/DA loop? Not that it would bother me but I thought you were a bit of a purist? ;)

Yes and no with regard to the trepidation. One thing I was concerned about was the latency caused by the conversion but an enquiry to Lab.gruppen confirmed that it was very low. I was (am) less concerned about applying this technology to the bass than I would be using for the rest of the frequency range. In addition, if one is to address room bass modes (and the major room coloration is likely to be in the bass) the only practical way is with DSP - passive room correction isn't very effective at lower frequencies


The other thing that surprises me is that the Naim amps are a good match for the NS1000M. Well okay doesn't surprise me but it is interesting. I would have thought they would prefer more forgiving amps.

The Naim amps are a really good match. I've heard NS1000Ms a few times powered by valve amps - far too much tail wagging the dog for my taste ;).

[walpurgis]

Those Avalons look like you could put wings on them and fly over Iraq without being seen.

Funny thing is, I've heard a few Avalons and Wilsons and never been desperately struck by the sound. But then, I could say the same about B&W and Focal J M Labs.

[Magna Audio]

Very interesting.
I have DSP X/O / Dac on 4 channels but I run the 4th on a passive, the signal for all 4 does go through the DSP however with the lower 4 channels using DSP X/O and the upper channel on pass through X/O wise. I time align all 4 in DSP.
The 5th channel signal (tweeter from 11KHz up) is taken off the 4th channel feed physically split from the DSP output. As the 4th channel was on pass through X/O wise it works just fine (has all the signal) and with such high frequencies meaning short physical distances, the tweeter and 4th channel are physically time aligned.

The benefits of room correction on the bass channels are great aren't they!

With multiway systems DSP really is the only way to got. I did have the system on full passive setup and the two bass horn inductors and caps are so big they cost nearly as much as the DSP solution! Passives have no flexibility or room mode correction feature either:)

As an experiment I frequency response corrected my mid and upper channels and it sounded dire!

Have you thought about trying all channels on DSP X/O and even time aligning?

[YNWaN]

Converting the whole frequency range to digital and then back again doesn't appeal to me - the compromise doesn't outweigh the benefits for me. The drivers are already reasonably well time aligned and the in room response over the mid and top is pretty smooth as is.

[Clive]

Yep, bass done this way is very controllable. I agree about leaving the mid and treble as pure analogue. Glad it's working so well for you Mark.

[YNWaN]

Thank you. I'm still trying to interpret the NS-1000M pages that Troels Gravesen has written. He talks about a fourth order slope but what he has built is clearly a third order - presumably he means that when combined with the bass drivers own roll of it forms a fourth order. I've emailed him about it but he isn't very forthcoming :(.

[Clive]

Thank you. I'm still trying to interpret the NS-1000M pages that Troels Gravesen has written. He talks about a fourth order slope but what he has built is clearly a third order - presumably he means that when combined with the bass drivers own roll of it forms a fourth order. I've emailed him about it but he isn't very forthcoming :(.
3rd order plus mechanical rolloff could be described as 4th order though it is misleading. If he's doing this actively I suppose you cant tell if he's inverting polarity vs the mid. 3rd order should be...

[YNWaN]

He's doing it passively. I actually know all the values of his crossover as someone has kindly told me them :). As I understand it a third order is 90 degrees rather than 180 so not necessarily inverted polarity - he certainly isn't doing so.

I've just spent the last two hours experimenting with the 3rd order slope against the standard 2nd order - it's the work of literally two seconds to change between them once the values are loaded into presets!

Looking at Troels' crossover (bass) it's definitely electrically third order, inductor in series, capacitor in parallel and another inductor in series. However, if I type his values into a crossover modeller and assume an 8Ohm driver I get a very low crossover frequency of 342Hz (the 2nd order Yamaha one is 500Hz). Even then, the capacitor he has chosen is some 5uF larger so I don't quite get it. I assume he is not using 8Ohm as the driver impedance...?

[Clive]

Oh yes, 90 deg makes more sense. Have you got variable phase (not just invert)? I struggle to hear phase variations of up 30 to 45 deg anyway.

[YNWaN]

No, can only invert phase - but Troels seems to imply that phase tracking is better at the lower end of the mid with a 3rd order low pass filter used - listening would seem to support that. I certainly think the steeper 3rd order slope works better in terms of reducing mid coloration.

[Reffc]

He's doing it passively. I actually know all the values of his crossover as someone has kindly told me them :). As I understand it a third order is 90 degrees rather than 180 so not necessarily inverted polarity - he certainly isn't doing so.

I've just spent the last two hours experimenting with the 3rd order slope against the standard 2nd order - it's the work of literally two seconds to change between them once the values are loaded into presets!

Looking at Troels' crossover (bass) it's definitely electrically third order, inductor in series, capacitor in parallel and another inductor in series. However, if I type his values into a crossover modeller and assume an 8Ohm driver I get a very low crossover frequency of 342Hz (the 2nd order Yamaha one is 500Hz). Even then, the capacitor he has chosen is some 5uF larger so I don't quite get it. I assume he is not using 8Ohm as the driver impedance...?

Wonder who that might have been :scratch::lol:

3rd on bass with 2nd mid will mean +90 degree electrical phase on bass. You can achieve absolute phase by moving mid/treble forward but that's out of the question with existing speakers. As Clive says, sometimes it's hard to pick out phase differences until they're actually quite significant. The issue about 3rd appearing as 4th order acoustically is true of reflex loaded cabs where the rolloff is steeper in the bass at the lower end due to the port but at bass/mid interface, it doesn't quite work like that. In that case, it is more dependant upon the natural roll off characteristics of the bass unit and the inductance of the bass unit voice coil combined with the crossover frequency. These things aside, go with what your ears tell you. Having calibrated mics helps a lot so "see" what's going on, but in their absence, you'll know what sounds right for you. Best of luck with the project Mark...very interesting it is too.

[YNWaN]

Thanks Paul. If it's OK with you I think I will email you the TG crossover details as I would like to extrapolate the electrical crossover corner frequency from it - assuming that is possible?

I was sent the crossover details some time ago by a chap who had read about my Yamaha modifications. I don't want to post them on an open forum as Troels sells the modified passive crossovers (well, the plans - the actual components come from Jantzen).

It is interesting to compare the crossover slopes - it's incredibly easy to input the crossover type and corner frequency - notch filters can also be easily added.

http://i62.photobucket.com/albums/h90/markemark_2006/5bd0ecbd6a90b5d2a9467a4418257d4a_zps5ec2f2ec.jpg

[Reffc]

When you refer to the "corner" frequencies Mark, I presume you mean the roll-off points? Actual electrical crossover frequencies cannot easily be calculated without measuring acoustic slope of the drivers in question, fitted into the cab baffle and run without a crossover (except tweeter....always wise to put a cap in for that to limit low point but allowing assessment of acoustic profile at upper crossover point), they can only be guessed at. The other point often missed when looking at the electrical and acoustic slopes for individual drivers is that the point at which the slope starts isn't the crossover point. That point depends upon the chosen crossover "Q" and the actual point of crossover will be that point at which the summed response of both drivers involved matches the summed response point for the chosen Q. eg Butterworth 3rd order summed response will peak at crossover by several dB, 2nd order L-R will be summed flat. Going from 3rd B'w to 2nd LR will result in something between the two. Also, dont forget that phase will only be 90 degrees at crossover, but will vary below and above that.

E:mail away. Will help if I can.

[YNWaN]

Hi Paul, yeah, I appreciate all of that :) - it's been a steep learning curve. For me that's the fun though, new projects focus the mind and encourage me to look properly into areas I had only previously glimpsed into. It's fascinating to hear how the voicing of the speaker changes when changes are made to the crossover rollover frequency and crossover slope.

@ I thought 'corner frequency' was a term used to describe the point that would exist if the crossover slope was a straight line and where it would meet the 0dB line - as you say, the roll-off point?

[Reffc]

Hi Paul, yeah, I appreciate all of that :) - it's been a steep learning curve. For me that's the fun though, new projects focus the mind and encourage me to look properly into areas I had only previously glimpsed into. It's fascinating to hear how the voicing of the speaker changes when changes are made to the crossover rollover frequency and crossover slope.

@ I thought 'corner frequency' was a term used to describe the point that would exist if the crossover slope was a straight line and where it would meet the 0dB line - as you say, the roll-off point?

It is a fascinating area of design Mark and glad you're having fun with it. Yes, the term does seem to be bandied about, but the roll off point itself is secondary to the summed response, actual crossover point and changes in off-axis response which a good design will aim for minimal disruption to stereo image. This presupposes that phase is considered and also designed to be as close to zero degrees as possible, but that's often not possible so the overall phase response above and below the crossover point over several octaves shows the bigger picture. Lots to juggle with!

[YNWaN]

Yesterday I experimented with listening to two low pass options - one was the Yamaha 2nd order 500Hz and the other a 3rd order at 405Hz. The 3rd order did seem to be better focused in the bass and have greater clarity in the lower mid - presumably because of the steeper slope.

[Reffc]

Yesterday I experimented with listening to two low pass options - one was the Yamaha 2nd order 500Hz and the other a 3rd order at 405Hz. The 3rd order did seem to be better focused in the bass and have greater clarity in the lower mid - presumably because of the steeper slope.

Yes, steeper slopes often clean up the area just above crossover as it avoids the area where the bass driver is producing most distortion (ie cone breakup) and as long as the mid can be cleanly run well above its's lower resonance point, which it is being, then you will benefit from a cleaner presentation. The trick as I'm sure you'll find, is in choosing the correct point as a crossover to ensure that summed response stays within a few dB of the desired frequency response (ie flat response). Before anyone jumps in saying "but a flat response sounds dead!" that's not what is meant. A flat response is simply that when you pass a test sweep that all frequencies are presented at a similar sensitivity otherwise a recording will not play back the way it was intended to. Using this as a starting point you can then offer compensation for the desired listening levels (remembering F-M curves) usually by either a gently downward slope or a slope which is slightly lifted in bass and treble (or recessed in the mids a la BBC monitor design for some of their broadcast monitors).

[YNWaN]

:) at least I do understand what you meant by a flat response and, as you say, choosing the crossover frequency appears to be key. With this crossover/amp you can save the settings as presets and hot swop between them which is very illuminating.

The third order option does seem to give better subjective focus than even order options. I must say, I'm very glad that it is only the bass driver response I am fiddling with!

P.S. If I use the wrong terminology please feel free to correct me - such things matter to me and I'm on a steep learning curve.

[YNWaN]

A friend visited yesterday and we did some measuring (and listening). We concluded that the 405Hz 3rd order did blend very well, but only if the bass drive was phase inverted. We also found that the room has a significant 50Hz mode (no great surprise there) and managed to reduce that significantly by applying EQ.

[YNWaN]

http://www.gryphon-audio.dk/media/9897/benefits-of-semiactive-rev.pdf

The Gryphon Pendragon - semi-active (active bass).

YG Acoustics also used to sell semi-active speakers that used Class D amps to drive the bass and adjustable EQ. Oddly, they seem to be moving back to fully passive systems though I suspect this has more to do with not wanting to be tied to third party amp manufacturers than anything else (YG have been moving more and more of their production 'in house' in recent times with manufacturing their own drivers and inductors - other crossover components are Mundorf).

[Magna Audio]

It can be made to work fantastically well.
Several systems I've heard prove that to me. Once you go truly multi way and have proper deep bass where the only way to do it correctly is time assigned you can solve many of the problems in DSP.
Analogue input still has that analogue sound / difference to digital sources too!

[YNWaN]

I'm very happy with how the system is sounding, the digital crossover and class D amp have blended in better with the passive crossover and A/B amps than I could have hoped.

[Magna Audio]

I'm very happy with how the system is sounding, the digital crossover and class D amp have blended in better with the passive crossover and A/B amps than I could have hoped.

That's good.

I convert all channels to digital but on the upper two - 1200 upwards in my system I have no DSP X/O, only time alignment delay and use passive 1st order on the S2's and Raal ribbons. It just seems to work very slightly better. The control of a couple of room modes is essential below 120Hz. All else is left un-messed with correction wise.

[John]

Its great to see that you gone down this route I too have been using this route for a few years now (controlling bass via DSP and keep passive in the mid and treble) I love what it has done for my system. I been tempted to go fully active but after a good chat with Steve feel like I got the best of both worlds

[Paul Lange]

I have to say I am a convert to DSP correction. After reading this thread I downloaded a small equalisation program from Sourceforge called EQ APO that uses a simple text file to affect the equalisation, and a sinewave sweep generator program. Just a simple sweep from 40hz to 300hz allowed me to note where I had any serious resonances and dropouts. I then entered these into the text file using a bit of estimation and after a few iterations I had a much better response and nice tight bass. If I still had my mic I could take some room measurements and do a more accurate job.

Problem that I have is what could I use to process the sound from my turntable without compromising what I like about my analog front end?

[YNWaN]

At present, the route of active crossover with DSP for the bass and passive crossover for the mid and treble gives me the best of all worlds :). Unfortunately, this is only practical when using a three way speaker (as I do).

DSP at source is an interesting option but it inevitably means that the whole signal is so filtered - though this is not such an issue if already in the digital domain.

Paul, I'm afraid there is no easy solution to your turntable question. Well, the easiest is to pass your phonostage (assuming you have a standalone phonostage) into you computer, or via an external analogue to digital converter and then into your computer. However, if the absolute quality of the turntables sound matters to you (as it does to me) I wouldn't do this.

[Paul Lange]

I realised that I have a M Audio 2496 card in my music PC so I thought WTH and plugged the phono stage into the audio inputs. After a bit of head scratching I worked out how to get the music out of the PC ..........

and it aint half bad to be honest. Not quite the same signiture as before but certainly a lot to be happy about. I do think that getting the bass humps and drop outs smoothed trumphs my purist analoue sensibilities.

I think I will have to get a calibrated mic..

[Ali Tait]

http://www.minidsp.com/products/acoustic-measurement

[Paul Lange]

http://www.minidsp.com/products/acoustic-measurement

Thanks for that, seems much less ago than a mic and phantom power supply.

[YNWaN]

It doesn't really need to be calibrated if you are making comparative measurements. The one in Ali's link should be a good choice.

[Ali Tait]

It comes with a free calibration file.