Audio Research PSU - Capacitor's config. - UK Vintage Radio Repair and Restoration Discussion Forum

Diabolical Artificer · 16th Jul 2017, 12:27 pm

Whilst doing research for the big amp I'm building I looked at folk had built theirs. One of which is the Audio Research D125 100w amplifier. the PSU is interesting and quite complex but it's the reservoir caps that interest me.

They use four 800u caps, a 2u, two 1u and a 0.1u in parallel - see att. From what I can glean the small value caps are to shunt higher frequencies to ground. I've heard of using caps 10 times smaller than our reservoir cap, say 100u, 10u and a 1u but tried this, with AFAICT no appreciable difference in ripple voltage etc.

What are the reasons for putting smaller caps in parallel with large value reservoir caps and how would one work out what value's to use?

Secondly the HT/B+ is 420v, the voltage rating's of their caps are 450v and 425v, this seems a bit close to the knuckle. From all I've read and picked up here ont forum I was under the impression that you should chose a voltage rating 60% higher than your wkg V. We're AR being tight or is 60% over rating your PSU caps over egging the pudding?

Your thought's welcome as always, Andy.

G8HQP Dave · 16th Jul 2017, 2:05 pm

Putting caps in parallel like this is often a bad idea. What can happen is that the capacitive reactance of a small cap forms a parallel tuned circuit with the stray inductance of a big cap so you get a rise in net impedance at some frequency a bit above the audio range. RF people sometimes do this in a bias network, but RF people generally know what they are doing. Audio people do this because they don't always understand the possible snags, and journalists and 'audiophiles' expect to see it.

Such 'bypass caps' make no difference to ripple, as you found. The intention is to maintain a low PSU impedance over a wide range of frequencies, but sometimes it can make things worse.

What should be done is put suitable decoupling capacitors on the amplifier board near the output stage. These may appear to be directly in parallel with the PSU reservoir caps, but they are actually separated by wiring resistance and inductance so they do more good and less harm than direct 'bypasses'.

ms660 · 16th Jul 2017, 2:40 pm

Might be worth watching these Andy, 1st one covers capacitor SRF and some measurements, 2nd one is about supply rail decoupling:

https://www.youtube.com/watch?v=vi24SpKYYoQ

https://www.youtube.com/watch?v=9EaTdc2mr34

Lawrence.

Radio Wrangler · 16th Jul 2017, 7:39 pm

Choose electrolytic capacitors from a reputable maker,. Buy them from a real distributor who won't be laundering fakes (knowingly or unknowingly). Run them between 20% and 80% of their rated voltage, and then you only have to worry about the ripple current rating.

Sticking capacitors in parallel is a bit of a lottery. All sorts of things can happen. Electrolytic capacitors contain wound-up lengths of foils, and have a limited number of tapes coming out to their connections. So they look inductive at higher frequencies. A smaller capacitor will turn this into a resonant circuit. Things now depend on the resistive losses in the capacitors to damp the resonance. If you do this sort of thing, you need to have got realistic values for component strays and done some thorough modelling. The better the capacitor, the bigger the risk.

David

CambridgeWorks · 16th Jul 2017, 8:15 pm

Hi Andy, in light of what David just said, I confirm those BHC 820uF 400V were I think direct from the makers. Certainly used perfectly well in large 24/7 smps, 28v 32A. 15A fuse, straight off the mains, into a bridge then 2 of those caps as smoothing. Only failure was caused by lack of cleaning the fan filter, allowing a build up of dust across the pcb tracks of those caps and BANG! A flashover. I once stopped all the production women talking with such a loud bang! Frightened them, and me somewhat!
Rob

Craig Sawyers · 17th Jul 2017, 10:02 am

I had one of these for a number of years, and very good it was. It is actually a hybrid with an FET front end driving a whole bunch of 6550 valves. Weighed a ton. The only thing I did with mine was replace the cooling fans with the same type (Papst) because the originals had gone a bit noisy. I did not mess with the circuit, because Audio Research's design is massively detailed.

There are lots of internal photos of this unit on the web. This one is of the PSU board https://www.***********/photos/sachcu...n/photostream/ but you get different views scrolling through the album. This is not mine, buy the way.

Craig

Diabolical Artificer · 17th Jul 2017, 3:57 pm

Thanks all. i understand the circuit better now.

I'd watched those video's before Lawrence but hadn't understood the second one fully. It illustrates the point's made here.

I'm not using those caps I got off you here Robin, the value is a coincidence.

there's a video of one here with fan issues Craig - https://www.youtube.com/watch?v=QP87Tu3nrUE&t=1654s he cured his with a spray lube.

A.

16th Jul 2017, 12:27 pm	#1
Diabolical Artificer Dekatron Join Date: Jul 2009 Location: Sleaford, Lincs. UK. Posts: 7,662	Audio Research PSU - Capacitor's config. Whilst doing research for the big amp I'm building I looked at folk had built theirs. One of which is the Audio Research D125 100w amplifier. the PSU is interesting and quite complex but it's the reservoir caps that interest me. They use four 800u caps, a 2u, two 1u and a 0.1u in parallel - see att. From what I can glean the small value caps are to shunt higher frequencies to ground. I've heard of using caps 10 times smaller than our reservoir cap, say 100u, 10u and a 1u but tried this, with AFAICT no appreciable difference in ripple voltage etc. What are the reasons for putting smaller caps in parallel with large value reservoir caps and how would one work out what value's to use? Secondly the HT/B+ is 420v, the voltage rating's of their caps are 450v and 425v, this seems a bit close to the knuckle. From all I've read and picked up here ont forum I was under the impression that you should chose a voltage rating 60% higher than your wkg V. We're AR being tight or is 60% over rating your PSU caps over egging the pudding? Your thought's welcome as always, Andy. Attached Thumbnails __________________ Curiosity hasn't killed this cat...so far.

16th Jul 2017, 2:05 pm	#2
G8HQP Dave Rest in Peace Join Date: Sep 2008 Location: Solihull, West Midlands, UK. Posts: 4,872	Re: Audio Research PSU - Capacitor's config. Putting caps in parallel like this is often a bad idea. What can happen is that the capacitive reactance of a small cap forms a parallel tuned circuit with the stray inductance of a big cap so you get a rise in net impedance at some frequency a bit above the audio range. RF people sometimes do this in a bias network, but RF people generally know what they are doing. Audio people do this because they don't always understand the possible snags, and journalists and 'audiophiles' expect to see it. Such 'bypass caps' make no difference to ripple, as you found. The intention is to maintain a low PSU impedance over a wide range of frequencies, but sometimes it can make things worse. What should be done is put suitable decoupling capacitors on the amplifier board near the output stage. These may appear to be directly in parallel with the PSU reservoir caps, but they are actually separated by wiring resistance and inductance so they do more good and less harm than direct 'bypasses'.

16th Jul 2017, 2:40 pm	#3
ms660 Dekatron Join Date: Apr 2011 Location: Cornwall, UK. Posts: 13,454	Re: Audio Research PSU - Capacitor's config. Might be worth watching these Andy, 1st one covers capacitor SRF and some measurements, 2nd one is about supply rail decoupling: https://www.youtube.com/watch?v=vi24SpKYYoQ https://www.youtube.com/watch?v=9EaTdc2mr34 Lawrence.

16th Jul 2017, 7:39 pm	#4
Radio Wrangler Moderator Join Date: Mar 2012 Location: Fife, Scotland, UK. Posts: 22,876	Re: Audio Research PSU - Capacitor's config. Choose electrolytic capacitors from a reputable maker,. Buy them from a real distributor who won't be laundering fakes (knowingly or unknowingly). Run them between 20% and 80% of their rated voltage, and then you only have to worry about the ripple current rating. Sticking capacitors in parallel is a bit of a lottery. All sorts of things can happen. Electrolytic capacitors contain wound-up lengths of foils, and have a limited number of tapes coming out to their connections. So they look inductive at higher frequencies. A smaller capacitor will turn this into a resonant circuit. Things now depend on the resistive losses in the capacitors to damp the resonance. If you do this sort of thing, you need to have got realistic values for component strays and done some thorough modelling. The better the capacitor, the bigger the risk. David __________________ Can't afford the volcanic island yet, but the plans for my monorail and the goons' uniforms are done

16th Jul 2017, 8:15 pm	#5
CambridgeWorks Nonode Join Date: Jun 2010 Location: Spalding, Peterborough, Cambridgeshire, UK. Posts: 2,859	Re: Audio Research PSU - Capacitor's config. Hi Andy, in light of what David just said, I confirm those BHC 820uF 400V were I think direct from the makers. Certainly used perfectly well in large 24/7 smps, 28v 32A. 15A fuse, straight off the mains, into a bridge then 2 of those caps as smoothing. Only failure was caused by lack of cleaning the fan filter, allowing a build up of dust across the pcb tracks of those caps and BANG! A flashover. I once stopped all the production women talking with such a loud bang! Frightened them, and me somewhat! Rob __________________ Apprehension creeping like a tube train up your spine - Cymbaline. Film More soundtrack - Pink Floyd

17th Jul 2017, 10:02 am	#6
Craig Sawyers Dekatron Join Date: Nov 2010 Location: Oxford, UK. Posts: 4,983	Re: Audio Research PSU - Capacitor's config. I had one of these for a number of years, and very good it was. It is actually a hybrid with an FET front end driving a whole bunch of 6550 valves. Weighed a ton. The only thing I did with mine was replace the cooling fans with the same type (Papst) because the originals had gone a bit noisy. I did not mess with the circuit, because Audio Research's design is massively detailed. There are lots of internal photos of this unit on the web. This one is of the PSU board https://www.***********/photos/sachcu...n/photostream/ but you get different views scrolling through the album. This is not mine, buy the way. Craig

17th Jul 2017, 3:57 pm	#7
Diabolical Artificer Dekatron Join Date: Jul 2009 Location: Sleaford, Lincs. UK. Posts: 7,662	Re: Audio Research PSU - Capacitor's config. Thanks all. i understand the circuit better now. I'd watched those video's before Lawrence but hadn't understood the second one fully. It illustrates the point's made here. I'm not using those caps I got off you here Robin, the value is a coincidence. there's a video of one here with fan issues Craig - https://www.youtube.com/watch?v=QP87Tu3nrUE&t=1654s he cured his with a spray lube. A. __________________ Curiosity hasn't killed this cat...so far.