KRELL KSA50 power supply - kind request of info.

[AndrewT]

Hi Kamps,
was that 2 by 56mF in series per rail or 2 in // per rail?
The original PSU had about 2A to 3A peak current capability but you improved this by a factor of 4.2 (series connected) or 11.2 (// connected). Low esr and low ripple and high current capability all in one go. As you say a new amp.
Comment on value for money on this modification.

 

[beppe61]

Dear Sir,

thank you immensely for your extremely kind and valuable reply.
This is indeed what I am looking for here: first hand experiences in modding existing equipment.
Please let me answer between the lines.

Just to provide some personal experience... i got an old Forte Audio model 3 (200wpc AB) from ebay and while i liked the mids and highs, the bass was not that great.

It is the very same case here. Maybe the Forte is very much better than my Samson Servo 260 but I like the sound of this amp, bass apart.

> I open it up and find that both channels share 4 24000uF caps 2 each in series so basically each rail had 12000uF and this was shared by both channels, so each channel got 6000uF per rail....

I have now 2 10.000uF of low quality.
I have bought 2 Siemens Sikorel of 15.000 uF (not much bigger but presumibly much better quality.
Rated among the best actually
for audio purpose).

> I first installed a couple of 56000uF's per rail, then bypassed them with low esr smaller caps of 10,000uF then bypassed with 1000uF and 0.1uF.

Very interesting technique. I will study it deeply.
But first I would like to try the two new caps alone.

> The resultant sound was radically different, the bass is sooo tight and punches hard.
Hard to beleieve its the same amp.

THIS IS A POINT VERY VERY INTERESTING POINT.
I know that also Odyssey Audio offers a cap upgrade.
If only a could get a drop of more bass I will be done.

> Bottom line, instead of 1 big or many small cap, I went with a hybrid approach and it worked very nicely for me.

I am not an expert but I really think that your strategy has very good basis.
Before going for a cap upgrade I was rambling about the question if a bad transformer is more detrimental for bass than low capacitance.
I have non fully understood the problem yet, but I am always more and more convinced of the importance of capacitance in terms of size and quality.
Your experience is extremely helpful for me indeed.

Thank you sincerely.

Kind regards,

beppe61

 

[jacco vermeulen]

I open it up and find that both channels share 4 24000uF caps 2 each in series so basically each rail had 12000uF and this was shared by both channels, so each channel got 6000uF per rail....

Hard to believe from a Pass design.
The 40 watt class A poor mans Threshold version i owned had a powersupply as normal as can be.
Beppe, this one => http://www.diyaudio.com/forums/showthread.php?postid=598742#post598742


Arif,
you and a bunch of others posted positive reports on the 1:10 cap rule more than once. More pragmatic than stuffing endless boards with small caps. Using series of 10uF-1uF-100nF on the rails of tube and class A amplifiers has been common practice overhere for a very long time.
Actually, i was not a lantern but a ticked off clock.
You're my guiding light, i've adopted the 1/10 thing as common PS rule.

Beppe,
i've used old Siemens Sikorels a few times, not on the PS of a power amplifier though.
Sikorels are the very best, but were also the highest priced for a long time.
I haven't put anything less than 100.000uF per channel in an amplifier after Desert Storm, Sikorels are too rich for my blood.

Overhere Mallory's were exotics.
Sprague, Sic-Safco Felsics, Roederstein, Siemens and of course Philips capacitors were obtainable ( that, and sheep sheet)
All prime US brands used Mallory capacitors then.
Finding a cap bigger than 47000uF with a higher rating than 40 volts wasn't possible.
Yeah, i am sentimental about Mallory.
Cornel Dubilier makes fine capacitors as well. And the ones i mentioned before are fine too, even the Philips.
Canadian Classé used Philips caps, i noticed even PassLabs does for smaller capacitors than the big cans.

btw:
also our ecologist designer Mr Ball loves placing PS capacitors in parallel and plenty capacity for stiff rails for his interesting Vas first stage PP complementary Mosfet design, with decoupling on the boards.
If i recall, he also did on his chip entry Mosfet amplifier in the 90s for the Oceania market.
As did Gunther Mania in Europe with his BB OPA627/OPA637 Mosfet AVM amplifiers in those same years.

 

[beppe61]

Dear Mr. Vermeulen,

thank you so much fro your very kind and extremely helpful reply.
Please let me reply between the lines.

>

Hard to believe from a Pass design.
The 40 watt class A poor mans Threshold version i owned had a powersupply as normal as can be.
Beppe, this one => http://www.diyaudio.com/forums/showthread.php?postid=598742#post598742

This is more or less my dream amp.
A huge high regulation toroidal surrounded by big blue high grade caps.
If this is a poor man amp, dear Sir, I would never be rich.
Very very nice indeed. Congratulations !

> Beppe, i've used old Siemens Sikorels a few times, not on the PS of a power amplifier though.
Sikorels are the very best, but were also the highest priced for a long time.

I hope the caps I have bought on ebay are not fake.
I am afraid so because the price is 10 euro/each for 15.000uF/100V. I hope they are original.

> I haven't put anything less than 100.000uF per channel in an amplifier after Desert Storm, Sikorels are too rich for my blood.

Wow, this is quite a capacitance!

> Overhere Mallory's were exotics. Sprague, Sic-Safco Felsics, Roederstein, Siemens and of course Philips capacitors were obtainable ( that, and sheep sheet)
All prime US brands used Mallory capacitors then.
Finding a cap bigger than 47000uF with a higher rating than 40 volts wasn't possible.
Yeah, i am sentimental about Mallory.
Cornel Dubilier makes fine capacitors as well. And the ones i mentioned before are fine too, even the Philips.
Canadian Classé used Philips caps, i noticed even PassLabs does for smaller capacitors than the big cans.

Thank you sincerely for your extremely valuable advice on the best capacitors brands.
I think another good method to select components is to look at what famous manufacturers use for their top of the line offers.
For instance:
MarK Levinson Mallory on the N33 amp, Nippon ChemiCon now.
KRELL Sprague on old amps (do not know now)
Halcro Siemens Sikorel
BAT NIchicon KG series
YBA ROE
and so on.
Anyway is not easy to see inside an amp, but in some reviews they show you it opened.
I am very curios to see what a simple caps replacement can do to an amp. Very curios.
and then a big high regulation toroidal.

Thank you very much indeed for your extremely valuable advice.

Kind regards,

beppe61

 

[jacco vermeulen]

Beppe,

15.000uF/100V Sikorels i can get for 38 Euro's the piece.
Could also be that you made a good deal getting Sikorels for 10.

I just aquired a large number of 33.000uF Rifa capacitors.
They look brandnew, supposedly from a company that had a cancelled order, for far less than list price.

I've done the same for years, reading magazines with a magnafying glass and trying to arrange a hearing at an audio show or a store.
Heard the YBA2 power amplifier, even had the good fortune of catching Yves Bernard André at an audio show in France.

Long time ago i heard the Gassmann batteryfed Metaxas amplifier, quite a difference.
At one point i had the crazy idea to use submarine battery cells for an amplifier powersupply.
(sub battery cells weigh 1000lbs, do a little over 1 volt)

 

[beppe61]

Dear Mr. Vermeulen,
thank you so much for your always kind, welcome and valuable reply.

>

Beppe, 15.000uF/100V Sikorels i can get for 38 Euro's the piece. Could also be that you made a good deal getting Sikorels for 10.

I hope so. I am very curious, very. Some years ago I read an article about re-capping performed on a Roksan integrated.
With every caps swapping the sound changed !

> I just aquired a large number of 33.000uF Rifa capacitors.
They look brandnew, supposedly from a company that had a cancelled order, for far less than list price.

From what I read very good units indeed. Lucky you.
I will do the same, if the first experience will be positive.

> I've done the same for years, reading magazines with a magnafying glass and trying to arrange a hearing at an audio show or a store.
Heard the YBA2 power amplifier, even had the good fortune of catching Yves Bernard André at an audio show in France.

I remember a Stereophile review on a YBA preamp.
The capacitance was so high that the preamp could go on sounding for minutes after disconnetting it to the mains !
Great brand indeed and they use huge amount of capacitance and no voltage regulators at all !
And for the Signature series they even make their own capacitors !

> Long time ago i heard the Gassmann batteryfed Metaxas amplifier, quite a difference.

Metaxas. Another excellent brand. Their designer is one of the best in my opinion. and also the execution is wonderful.

> At one point i had the crazy idea to use submarine battery cells for an amplifier powersupply.
(sub battery cells weigh 1000lbs, do a little over 1 volt)

I know that batteries could very well be the ultimate solution to power and audio equipment.
But also with batteries, expecially lead acid ones that have a high internal impedance, the caps at the output and their quality is fundamental.
I start to be intrigued by those power reservoirs and their construction.
I hope to be able to report some nice findings soon.
As I said I am quite curious, quite.

Thank you very much again.
Kind regards,

beppe61
ITALY

 

[ANDYLASER]

With regard to bypass caps on large cans, where is the best place to physically place them? Is it better to try and attatch them to the cans themselves or place them nearer to the board they are supplying?

I wouldnt have thought trying to fix 3 or 4 caps including a 10,000uf one, on to a large can was that easy to accomplish.

 

[beppe61]

Dear Friend,

My humble opinion on by-passing big caps with small film caps.
The small by-pass high quality film caps is like a sweet cherry on a cake.
There is no cherry that can make a bad cake taste good.
The problem is always in the cake.
I am referring to high capcitance caps of course.
On small ones the story can be different.
Maybe I am missing something.
Nevertheless I would vey much appreciate some feedback on the topic.

Kind regards,

beppe61

 

[K-amps]

Hi Kamps,
was that 2 by 56mF in series per rail or 2 in // per rail?
The original PSU had about 2A to 3A peak current capability but you improved this by a factor of 4.2 (series connected) or 11.2 (// connected). Low esr and low ripple and high current capability all in one go. As you say a new amp.
Comment on value for money on this modification.

Andrew,

56kuF each cap (effective 56kuF) per rail replaced 48000uF (12000uF effective.)).Sorry for earlier post where I said a couple of 56kuf per rail, actually it is only single 56k per rail)

Total capacitance of amp after mod close to 140kuF effective

 

[K-amps]

With regard to bypass caps on large cans, where is the best place to physically place them? Is it better to try and attatch them to the cans themselves or place them nearer to the board they are supplying?

I wouldnt have thought trying to fix 3 or 4 caps including a 10,000uf one, on to a large can was that easy to accomplish.

Andy as close as possible to the big cans.. so forget cosmetics

in my case, I cable tied the 10kyF to the gnd buss bars. If this is not possible, at least the 1000uF and 0.1uF should be mounted on the Big can terminals to avoid inductance/ oscillations.

 

[AndrewT]

Hi,
sorry to disagree with Kamps but I think the bypass caps should be placed at the output of the PSU. The smallest cap right at the cable output connections and progressively back towards the main cans with each larger bypass. This is to minmise inductance at the output and so make the hi frequency caps more effective.
Long wires/traces will add series inductance to any cap and will completely destroy the benefit that small caps (1nF to 100nF) are intended to bring. The trick will be to make the whole assembly compact and with a minimum of track length from output back to first cap in the bank. In this bypass senario ceramic caps are perfectly acceptable indeed preferred.

 

[K-amps]

Andrew what you are referring to is what I call local reservoir caps which is another way of setting them up. In this application the idea is to reduce ripple caused by the cables and allowing the OP devices to draw on these local caps for short term needs.

Another way of setting them up no doubt.

 

[AndrewT]

Hi Kamps,
we are at odds.
The poster and my reply refer to PSU main cap bypassing.

I think you are now interpreting my reply as if it were on board decoupling. It is not, however I would repeat the bypassing philosophy for the decoupling as well. i.e. progressively smaller caps towards the PSU output and towards the main current consumers/glitch producers on the amp PCB.
Does that clarify my thoughts?