PDA

View Full Version : Solid State Power Amp (Yes Solid State I Say)



Primalsea
10-01-2010, 19:41
This is another reincarnation of the first amp I ever made. It was once a glorious success until my wife decided to move the speakers one day while the amp was on. She pulled the cables out of the back and shorted them together!:(. When I got home I turned the hifi on and it didnt work. A closer look inside the amp showed some signs of things getting hot. I asked her about it and she said I thought your amp smelt funny:doh:.

I then rebuilt it into a small case and used it upstairs just for listening to music on the computer and for testing other DIY thingys. I've always liked the amp, it has a clean dynamic sound but never harsh and agressive. Also very fluid, even compared to valve amplifiers. A while ago I started to use it in place of the WD88 and actually preferred it! So I managed to get 2 250VA toroids on Ebay for a very reasonable amount, some PSU caps & diode bridges and with all the other bits and pieces I have laying around I started to make a heavy Dual Mono, Split High Current & Low Current Rail Supply.

The Rack Mount 2U case is a bit on the large side but very nice and had just been sitting around for quite some time so I decided to use it rather than buy another case just for this. As you can see the amp boards are quite small, especially when compared to the PSU. At least it give me loads of room to add more stuff or even more channels later on.:eyebrows:.

The filter circuit is from a photocopier and is designed for quite high currents. I have modified it a bit but its a bit of a winging it thing at the moment as passive filters have to be optimised for the load to get the result intended and it may need a bit of honing later on.

The amplifier circuit is a simple affair but I have always preferred amplifiers with the simple circuits over the complicated ones soundwise. To me many amplifiers are far too complex and usually the result of mental masturbation just to hammer down a particular distortion figure at the expensive of musicality. For me its simple circuits with good well thought out supporting circuits which can be as complicated as needed that work. The amp circuit might benefit from a more complex PSU or more current and accurate current sinks etc but this will be something for later on. I will get it going which what I have so far and see what happens and enjoy it for a while first, as long as it doesn't blow up!

http://i44.photobucket.com/albums/f23/primalsea/DSC01053.jpg
http://i44.photobucket.com/albums/f23/primalsea/DSC01052.jpg
http://i44.photobucket.com/albums/f23/primalsea/DSC01051.jpg
http://i44.photobucket.com/albums/f23/primalsea/DSC01050.jpg
http://i44.photobucket.com/albums/f23/primalsea/DSC01049.jpg

DSJR
10-01-2010, 20:20
Isn't that the sort of thing that ALL amps should be?

I'm sure that this is one reason that well designed and laid out valve gear works, as the circuits are generally simple (Glenn's products have becaome more this way as time's gone on and weren't ever complex at the beginning).

Where did you get the circuit or ideas for such?

Primalsea
11-01-2010, 19:07
The Amp circuit is based on that from ESP Audio but I spent quite a long time, about 2 years modifying it until the blowing up incident.

The PSU is just based on what I know and have experienced in the past.

Themis
11-01-2010, 19:49
Nice stuff, Paul ! :)

Is this a class A amp ? How much did it cost to build ?

Beechwoods
11-01-2010, 19:52
The wiring and construction looks extremely tidy... nice work, Paul!

anthonyTD
11-01-2010, 20:07
hi paul,
what output devices does it use?
A...

Primalsea
11-01-2010, 20:33
Cheers guys,

Its a class AB amp but bias is set optimally in order to reduce cross over distortion. I have belief that not all AB amps are biased properly so that they run cooler and have less issues with reliability. A number of them I have worked on were under biased. I think a lot of AB amps get a bad name as a result of this.

I will have to get back to you, Anthony, on the output devices after I've wondered upstairs and had a look. I know they are genuine Toshiba but cant remember which ones.

anthonyTD
11-01-2010, 20:52
Cheers guys,

Its a class AB amp but bias is set optimally in order to reduce cross over distortion. I have belief that not all AB amps are biased properly so that they run cooler and have less issues with reliability. A number of them I have worked on were under biased. I think a lot of AB amps get a bad name as a result of this.


I will have to get back to you, Anthony, on the output devices after I've wondered upstairs and had a look. I know they are genuine Toshiba but cant remember which ones.
:)

leo
11-01-2010, 21:35
At a guess I'd say 2SC5200 and 2SA1943:)

Nice work Paul, is this the P3A by any chance?

Alex Nikitin
11-01-2010, 21:39
Hi Paul,

IMHO, the heatsinking with just that thin case (even if it is aluminium) won't be sufficient. I would at the very least put a couple of 5-6 mm thick aluminium plates about 60x120 mm in size under each pair of devices, otherwise the life of the amplifier may be exciting but very short!

Cheers

Alex

leo
11-01-2010, 22:10
Hi Paul,

IMHO, the heatsinking with just that thin case (even if it is aluminium) won't be sufficient. I would at the very least put a couple of 5-6 mm thick aluminium plates about 60x120 mm in size under each pair of devices, otherwise the life of the amplifier may be exciting but very short!

Cheers

Alex

Well spotted, I thought it was one of those cases with the thick panel sides used as heatsinks but looking closer it seems to be a bog standard rack mount type

leo
11-01-2010, 22:19
BTW if it is P3A I'd try adding some higher uf caps on the amp board for local decoupling, the P3A could be a little touchy especially with the faster more popular Toshiba transistors

Primalsea
12-01-2010, 19:14
Hi guys,

Leo you are right about the output devices used. It started life as a P3A on the ESP OCB's that I ordered about 6 years ago. After many rebuilds and mods thye PCB's became too worse for wear so I changed i to the point for point wiring on the matrix board you see now.

I will look into the tip about the decoupling caps, usually you can never have enough. The amp boards are OK as I have been using them for a number of years in a smaller case with a smaller PSU.

Also Alex I think you are right about about the heatsinking, The old case was made from thick steel and had the heatsinks bolted on to the top of it. I was going to fire the amp up and after setting the bias run it into a dummy load and monitor the temp of the OP devices. However you got me reconsidering this and I should really just use the heatsinks as you can never have too much of that either!

I really think the original P3A amp is very good. I know its really just a simple circuit and many consider it nothing special but I think too many people get hung up on making more complicated amps in the delusion that they will always be better. I think all that happens is you trade some benefits for others.

It wont be long before its firing up time:eek:, always a worrying thing. The PSU have been tested and thats OK, the amp boards have been used before so they should be OK. Hopefully I have split the low and high current rails properly!

Primalsea
17-01-2010, 17:29
I'm nearly done! After Alex's advice on the heatsinks (something that I should have known better on) I made some up using Alu plates fanned out. Its not quite as much metal as Alex suggested but they are doing the job better than well and after a temperature test (finger placed on large transistor for 5 seconds) on the output devices after a warm up period all seems OK.

Bias adjustment was quite easy with none of the constant chasing your tail that can go on with some amps. I dont have the desktop with its decent sound card anymore so I haven't done any distortion tests but checking the power rails and the speaker outputs with a scope gives a fairly decent set of readings. Ripple at the speaker terminals is just a couple of mV:). The low current PSU rails are unregulated but still have a decently low ripple. The high PSU rails are fairly low too. It all seems good and all that's left is to put the top on and then fine some feet for the case. The filter stills needs checking out more but I plan to do that when I buy one of those plug in power meters and a DMM that does Inductance tests. I used to do this by using an old amp to put a Sinewave through the inductor and into a load and then measure the -3db point, a pain in the arse really!!


http://i44.photobucket.com/albums/f23/primalsea/DSC01058.jpg
http://i44.photobucket.com/albums/f23/primalsea/DSC01059.jpg
http://i44.photobucket.com/albums/f23/primalsea/DSC01060.jpg
http://i44.photobucket.com/albums/f23/primalsea/DSC01061.jpg

Primalsea
18-01-2010, 22:31
I checked out the sound card on my laptop and even though the frequency response is fairly dire and its useless at driving even light loads properly at low volumes at 1Khz the THD is less than 0.003%. This was verified by looping the output back into the laptops audio input.

Just as a quick yardstick (that's all that THD is good for IMO) I measured the amp at a little more that 1Khz sine wave into 8 ohms resistive load at 1 Watt. Distortion came out at less than 0.03%.

This seems fairly good as the ESP Audio P3A that the circuit is pretty much based on with just a few small changes is published as having 0.04%. I could see up to the 5th Harmonic but 2nd was the most prominent so it looks like cross over distortion is subdued quite well even for a class AB amp.

All I need are some feet for the case and I shall give it a listening test.

leo
19-01-2010, 08:38
Thanks for sharing the updates Paul, let us know how it sounds:)

Heres some of the diy class A/B amps I've tried, P3A, Avondale NCC200, Aksa 55+ , Gainclones

http://pic20.picturetrail.com/VOL1293/4634025/12749957/271121643.jpg

Primalsea
19-01-2010, 19:24
Hi Leo,

All of those kits you have are quite well known and I've heard are quite good. What do you think?? I read too many times that people berate them in favour of more complex circuits. I do wonder if many people just don't want to admit that the complex amps don't always sound better after they have spent a small fortune and many hours making them.

I have never liked any solid state amplifier with a complex topology. They may have 0.0001% distortion (what signal into what load who knows sometimes) but I find they tend to sound hard and aggressive.

Ali Tait
19-01-2010, 19:35
Paul,have you tried any of Nelson Pass amps? I've been toying with trying an F5.Looks a nice simple crcuit.

DSJR
19-01-2010, 19:47
Nowt wrong with Class AB or even B these days as the devices are so much more capable and linear apparently.

I know *some* of you liked AVI's amps in the past. Class B but VERY carefully researched device types, practically non-existent crossover distortion and VERY SIMPLE CIRCUITS as I recall..

My Crowns have what looks to be a complex circuit, but apart from the IC on the input stage, the signal path is quite straight-forward and simple, all the rest of it being carefully applied local feedback and various VI limiting stages, which do prevent 4 Ohm use by and large..

Primalsea
19-01-2010, 20:00
Hi Ali, I haven't made any Pass amps but I do have quite a bit of info printed out on them. There one amp that had an interesting feedback circuit. It used a small cap on the output of the VAS stage to the inverting input of the LTP. It gave the amp a very good high frequency performance. I've tried a variation on of it on a few amps and it seems to work quite well.

Ali Tait
19-01-2010, 20:42
Interesting stuff.I've just been looking recently,and the F5 seems to get very good ratings on the net.The B1 passive pre sounds really good too.I've heard one,and for a few quid's worth of parts,it's astonishing.

leo
19-01-2010, 21:23
Hi Leo,

All of those kits you have are quite well known and I've heard are quite good. What do you think?? I read too many times that people berate them in favour of more complex circuits. I do wonder if many people just don't want to admit that the complex amps don't always sound better after they have spent a small fortune and many hours making them.

I have never liked any solid state amplifier with a complex topology. They may have 0.0001% distortion (what signal into what load who knows sometimes) but I find they tend to sound hard and aggressive.

Hi Paul,

All those amps have their strength's and weakness, I'd probably say the NCC200 was the best out of those in the pic(IMO) all worked fine into normal loads

Depends a lot on the circuit but I also can't see why a complex design would offer more in sound over a simple circuit, certainly more to go wrong with something more complex

I'm still wanting to try a new amp project, tempted by the F5, dunno yet, tbh I'd like something to be a serious step up from what I've tried so far rather than a side step

Leo

Primalsea
20-01-2010, 12:13
Hi Ali, Had a look at the F5 its an elegant design. As its class A it will probably need PSU capacitors with the highest ripple current you can find just to cover yourself.

Hi Leo, I have some info on interesting mods for P3A (and similar topology amps). I will have to dig it out and send it to you. You could try building a Leach Amp, or you could try digging your heart out with a wooden spoon, which is probably less painful!

MartinT
21-01-2010, 12:11
I really liked the Linsley-Hood 75W circuit I built as a teenager. It used BDY56 NPN power transistors in a quasi-complimentary circuit (because high current PNP transistors were so much more expensive in those days) that was elegant and not too complex and worked well in class AB. It sounded excellent and drove a pair of Cambridge R40 transmission line speakers rather well.

I just found the circuit here:
http://www.tcaas.btinternet.co.uk/HFN11-72.gif
I removed Q7 & Q8 at some point (protection transistors) and it sounded even better.

Primalsea
21-01-2010, 22:19
There is a concern that the protection transistors can momentary "open" on musical peaks and short drive to the output devices. I dont know if its true but I have always felt that after removing them from amps they have sounded better.

snip
09-04-2010, 18:35
Any of you guys ever build the 'Death of Zen' on Elliot sound lab pages?

I am in the middle of rigging up one of these for headphone use for a friend.


//J

MartinT
09-04-2010, 20:19
Looks interesting, but I don't like using zener diodes in audio circuits because of their noise. I wonder if the current source could be made more temperature stable whether the need for a regulated supply could be done away with?

snip
10-04-2010, 15:57
Looks interesting, but I don't like using zener diodes in audio circuits because of their noise. I wonder if the current source could be made more temperature stable whether the need for a regulated supply could be done away with?

Otherwise, it's not exactly hard to control 12W through a regulator :)

With a regulated supply the zeners are not needed.
//J

MartinT
10-04-2010, 16:56
With a regulated supply the zeners are not needed.

Yes, indeed, but perhaps I should have said 'noise sources' rather than zeners. The problem with many regulators is that they are noisy too, especially the three-pin types.

snip
11-04-2010, 02:09
Yes, indeed, but perhaps I should have said 'noise sources' rather than zeners. The problem with many regulators is that they are noisy too, especially the three-pin types.

A good noise shunt or a discreet regulating circuit quickly sorts it out, I guess for power amplifier use it is a bit different, but building it for headphones where the total power consumed is not much, it's pretty easy getting it under control.

Even the cheap and cheerful LM317T can easily be made pretty quiet, I recently stuck one in my DAC and it was pretty silent when I tested it.

I appreciate that it can easily end up with a power supply circuit that is more complex than the amplifier it self though.

Once I get the thing rigged up on some stripboard for some testing I will try it without the zeners and a regular 'brute smoothed' supply and report back.

//J

DSJR
11-04-2010, 13:30
Possibly off-topic (apologies), but in addition to the NCC200 boards, has anyone ever worked with the Quad 405-2 circuit, especially in its latest Dada implentation - later layout, decent bits and limiting removed? I always liked the musical qualities of this amp once it had several hours to lose a slight treble "sting" and the principle's ok, certainly now with higher power transistors available....