Page 2 of 5
#16 Re: Definitive Build
Posted: Sat Jan 29, 2022 7:23 pm
by Ray P
I think the answer is 'it depends' but I'm not an expert EE and I'm thinking that Nick et al will have a fuller answer - why not try them and see if you like what you hear. You'll connect them in parallel with the exisitng caps.
One think to be aware of is that you can get quite a kick from the energy stored in the caps so make sure you bleed them down with a resistor before messing with them and maybe have a bleed resistor as part of the power supply?
#17 Re: Definitive Build
Posted: Sat Jan 29, 2022 7:25 pm
by pre65
Ray P wrote: ↑Sat Jan 29, 2022 7:23 pm
One think to be aware of is that you can get quite a kick from the energy stored in the caps so make sure you bleed them down with a resistor before messing with them and maybe have a bleed resistor as part of the power supply?
I'd agree with that Ray.
Many years ago I got a belt from a Quad 11 amp that had not been powered up for a while.
*
#18 Re: Definitive Build
Posted: Sat Jan 29, 2022 7:33 pm
by ed
there was a recent post on diyaudio where this was discussed at length.....I can't lay my hands on it at the moment...
from my suspect memory I think the gist of it was that you can go so far and then any further is just wasting your money...I will have to find the thread to find out how far that is.....more in the fullness of time..
on all my SS amps( firstwatt clones) I use 120,000uf, as do most folks who've built firstwatt amps....but that will really be dictated by how much power the amp wants
meanwhile, if you have the caps at hand then try it and see what you think
#19 Re: Definitive Build
Posted: Sat Jan 29, 2022 8:10 pm
by Nick
The answer I would give would also be it depends. In part on how much current you want the amp to provide. The real answer is yes if more capacitance is an advantage. And if you want more capacitance wire them in parallel. As I think you are making a solid state amp I would not worry about the voltage being dangerous but the current could do damage if shorted. Worry about metal rings on fingers and watch straps. Solid state amps will pull the voltage down on their own well enough. Unlike valve stages they don't need the heaters to pass current. So don't worry about bleed resistors.
#20 Re: Definitive Build
Posted: Sun Jan 30, 2022 7:34 am
by karatestu
It is said that there is a sweet spot regarding the amount of capacitance provided for different amplifiers. RD said he added more whilst listening and stopped when adding more stoped bringing any audible improvement.
I have seen it mentioned that too much psu capacitance can make a power amp sound slow and too little can make it sound anaemic. I have not really explored that idea too much and just work on a simple rule of the more current required the more capacitance I throw a it..
My psu's are probably not optimal but I am past caring now. When I changed from Avondale to NVA amplifier boards I just reused tha psu's's which have three 10,000uf caps in parallel fir the output stage and three 2,200uf for the driver, VAS and LTP. This may be overkill when the transformers I am using are 100VA 25-0-25 EI types. Sounds wonderful to me and I really can't be arsed doing any more experimentation there.
#21 Re: Definitive Build
Posted: Sun Jan 30, 2022 10:04 am
by Nick
RD said he added more whilst listening and stopped when adding more stoped bringing any audible improvement.
I am always suspicious with those sorts of stories with no actual evidence. Its odd how it always ends up with an solution that happens to be convenient with respect to cost and convenience of manufacturer, Even if it was true, would it still give the same answer with your source playing your music through your speakers in your room?
I have seen it mentioned that too much psu capacitance can make a power amp sound slow
The thing I cant think of any mechanism where that nice sounding idea can be true. None of the amps phase response or frequency response should be altered by the power supply capacitance unless it was fundamentally wrong in the first place and the extra capacitance extended the low frequency response to the point where the problem could be heard. Likewise the source or speakers or room could have a problem when the frequency response of the amp was extended down enough to show them. In all those cases the extra capacitance is not the cause of the problem, its just allowed the other problem to be heard. And so it creates a nice easy to believe story that becomes a thing.
and too little can make it sound anaemic
I can believe that as it would indicate that its reducing the bass response. What I have seen a lot is that if (for example) the HF is rolled off in a system then it will sound balanced if the LF is also rolled off (in the opposite direction), Fixing the problem at either end will make it sound like the problem is at the other end, So for example a amp that's flat from 20Hz to 20kHz will sound fine, as will one that is flat from 60Hz to 10kHz, But extend that 60Hz to 20Hz and the limited HF will become audible (true even more with speakers BTW).
I would suggest then that the correct amount is either as much as you can afford both in cost and space, or at least as much as you can add before the amp stoops changing its sound in your system, and if that sound is better with less than that capacitance, then its time to find and fix the cause of that problem.
Likewise the effect of cheap caps should be fixable by adding film caps in parallel. If the effect o this is not what you hoped, the cause it probably a fault somewhere in the HF response of the system. Its very easy to try a treat the symptom instead of the disease.
#22 Re: Definitive Build
Posted: Sun Jan 30, 2022 10:28 am
by Cressy Snr
Nick wrote: ↑Sun Jan 30, 2022 10:04 am…. Fixing the problem at either end will make it sound like the problem is at the other end, So for example a amp that's flat from 20Hz to 20kHz will sound fine, as will one that is flat from 60Hz to 10kHz, But extend that 60Hz to 20Hz and the limited HF will become audible (true even more with speakers BTW).
Yes, it’s certainly true for speakers. I’ve long resisted the calls for a supertweeter to help out my Fane wideband drivers. They have limited bass extension which cannot be overcome without getting thoroughly impractical. That’s just the way the drivers are. Extend the bass with a helper and the sound will become dull. Put a supertweeter on and it’ll become too bright.
A true full range system will become seriously expensive if high efficiency remains the goal. Plus it will be ridiculously stupid to try doing that in a 12x12 room. It’s far too small for that sort of tomfoolery.
Of course I could ditch the SE valve amplification, put the 60W NVA derivative on and immediately the vista of speaker possibilities would stretch from horizon to horizon.
#23 Re: Definitive Build
Posted: Sun Jan 30, 2022 10:44 am
by karatestu
Nick wrote: ↑Sun Jan 30, 2022 10:04 am
RD said he added more whilst listening and stopped when adding more stoped bringing any audible improvement.
I am always suspicious with those sorts of stories with no actual evidence. Its odd how it always ends up with an solution that happens to be convenient with respect to cost and convenience of manufacturer, Even if it was true, would it still give the same answer with your source playing your music through your speakers in your room?
I have seen it mentioned that too much psu capacitance can make a power amp sound slow
The thing I cant think of any mechanism where that nice sounding idea can be true. None of the amps phase response or frequency response should be altered by the power supply capacitance unless it was fundamentally wrong in the first place and the extra capacitance extended the low frequency response to the point where the problem could be heard. Likewise the source or speakers or room could have a problem when the frequency response of the amp was extended down enough to show them. In all those cases the extra capacitance is not the cause of the problem, its just allowed the other problem to be heard. And so it creates a nice easy to believe story that becomes a thing.
and too little can make it sound anaemic
I can believe that as it would indicate that its reducing the bass response. What I have seen a lot is that if (for example) the HF is rolled off in a system then it will sound balanced if the LF is also rolled off (in the opposite direction), Fixing the problem at either end will make it sound like the problem is at the other end, So for example a amp that's flat from 20Hz to 20kHz will sound fine, as will one that is flat from 60Hz to 10kHz, But extend that 60Hz to 20Hz and the limited HF will become audible (true even more with speakers BTW).
I would suggest then that the correct amount is either as much as you can afford both in cost and space, or at least as much as you can add before the amp stoops changing its sound in your system, and if that sound is better with less than that capacitance, then its time to find and fix the cause of that problem.
Likewise the effect of cheap caps should be fixable by adding film caps in parallel. If the effect o this is not what you hoped, the cause it probably a fault somewhere in the HF response of the system. Its very easy to try a treat the symptom instead of the disease.
Wow, that is a very comprehensive answer. Thanks. I can adopt that thinking going forward.
Yes it's funny how RD's more expensive designs had more psu capacitance but that might be related to transformer rating and required power output ?
The slow and anaemic "thing" is/was very prevalent in the PFM diy room where this hobby all started for me. It just goes to show the folly in believing everything you read.
My amps have remote psu's with three 10,000uf caps in parallel which are just over a metre away from the amp boards (load). I was in two minds about adding another 10,000uf cap right next to the amp boards but your comment about providing as much as you can afford has put any doubts I had to bed. Thanks.
Sorry DQ, didn't intend to derail your thread but hopefully it is relevant and interesting to you.
#24 Re: Definitive Build
Posted: Sun Jan 30, 2022 3:51 pm
by Daniel Quinn
No probs and I would hardly call it a derailment
#25 Re: Definitive Build
Posted: Sun Jan 30, 2022 8:41 pm
by Daniel Quinn
I can’t figure out the wiring for 4 capacitors instead of 2 anybody know of a picture showing it . All I can find is capacitor banks on a pcb
#26 Re: Definitive Build
Posted: Sun Jan 30, 2022 8:47 pm
by simon
Just parallel them - join plus to plus and negative to negative. Don't get it wrong or it will go pop!
#27 Re: Definitive Build
Posted: Sun Jan 30, 2022 8:51 pm
by pre65
Is the power supply separate to the amp ?
#28 Re: Definitive Build
Posted: Sun Jan 30, 2022 9:18 pm
by Daniel Quinn
Nope the power supply ain’t separate and using two capacitors the plus is joined to negative with a bar on one terminal . And the other connector gives the + and ~ for the power amp board
#29 Re: Definitive Build
Posted: Sun Jan 30, 2022 9:30 pm
by pre65
Daniel Quinn wrote: ↑Sun Jan 30, 2022 9:18 pm
using two capacitors the plus is joined to negative with a bar on one terminal . And the other connector gives the + and ~ for the power amp board
Perhaps it's me being a thicko, but I don't understand that.
That's why diagrams are so useful.
#30 Re: Definitive Build
Posted: Sun Jan 30, 2022 10:16 pm
by simon
Sounds like a bipolar PS - +0-. If so the two caps joining - to + will be ground or 0V most likely, the point in the middle. Think of the two caps stacked vertically - from top to bottom the top cap will be + above -, the bottom cap will be + at the top - at the bottom. The negative of the top cap is joined to the + of the bottom cap.