Power Supply Resevoir Size - Page 3 - diyAudio
Go Back   Home > Forums > Amplifiers > Power Supplies

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 18th July 2012, 06:31 PM   #21
diyAudio Member
 
Join Date: Oct 2011
I understand you Nico. When I started building amps, I started with transistor amps and messed around with them a lot. At that time yes I can say that isolating the voltage section of the amplifier used to make the amplifier open up and ease up. Then I moved to Chip amps such as TDA and STK. STK used to isolate with 100 Ohm and 100uf.

Will continue the story....
  Reply With Quote
Old 18th July 2012, 06:42 PM   #22
Mooly is offline Mooly  United Kingdom
diyAudio Moderator
 
Mooly's Avatar
 
Join Date: Sep 2007
Quote:
Originally Posted by Nico Ras View Post
................In other words should we spend less on mega farads power supplies when 4700 uF (or less) will do and pay more attention to the front end driver stage power supply ?
That's very possible (and I'm not really a fan of huge values of C in PSU's) and yet I can't help but feel its how everything in a complete amp design comes together that really counts.

[If you saw the PSU I used (and still use) for initial amp development you would laugh. It's an 18VA (yes, no 0 on the end) 15-0-15 tranny feeding 4700uf caps. It's SC proof up to point and hasn't enough welly to damage outputs and yet still allows an amp to be driven to moderate levels on music. That's where the lateral amp started its life and where it first started to sound so good that I just knew I had to build it as a successor to the blameless I was then using]

If I were building the lateral amp again I might look at higher separate rails for the front end which would be easy at such low currents. A voltage doubler and zener would be more than adequate I think. Would it be as good though.

It's and interesting subject but I don't think there are hard and fast rules, certainly no magic formula.
__________________
-------------------------------------------------------
A simulation free zone. Design it, build it, test it.
  Reply With Quote
Old 18th July 2012, 06:53 PM   #23
diyAudio Member
 
lanchile's Avatar
 
Join Date: Sep 2008
Location: AREA 51
Very interesting subject! I hear all the time people saying monsters power supply (capacitance) will sound better... Will they? Well I try long time ago when I was building some of the NXV200 from Aussie amplifiers a big and strong power supply 80.000uf per rail, the sound was very good, but then I try with just 20.00uf per rail to see if there was so much difference. The result was that with 20.000uf the amp sounded faster (if that is a definition). from then till now when I build an amp (class A/B) let's say from 50 to 100 watts RMS per channel I use no more than 20.000uf per rail (using like 6.800uf caps values) with excellent results. I just concentrate in the quality of the capacitors in my case I use Mundorf audio grade 125c. I did try many brands before, but my first choice is Mundorf audio grade caps for their quality and sound.
  Reply With Quote
Old 18th July 2012, 07:38 PM   #24
diyAudio Member
 
Nico Ras's Avatar
 
Join Date: Nov 2005
Location: East Coast of South Africa
Hugh, AndrewT, Carlos, Andrej, Homemodder, Michael, Sakis, lets have views from all the old other experienced guys, because you are past speculation you have tried stuff that worked, tell us what worked and why you think it was a better solution.

I promise I will try out some stuff practically from what is disclosed here in the next few weeks and report my findings, maybe we can establish some audio rule of thumbs that can be made applicable. Maybe what our friend Lancile has to say is very important it is quality in stead of quantity, low impedance, high ripple current, high temp, etc.
__________________
Kindest regards
Nico

Last edited by Nico Ras; 18th July 2012 at 07:43 PM.
  Reply With Quote
Old 18th July 2012, 08:44 PM   #25
diyAudio Member
 
Join Date: Oct 2011
Default special caps under development

There are these special caps under development at the lab
Attached Images
File Type: png CAPS.png (6.4 KB, 1204 views)
  Reply With Quote
Old 18th July 2012, 10:11 PM   #26
diyAudio Member
 
Join Date: Mar 2008
Larger banks of capacitors takes longer time to reform.
The connection distance from the capacitor bank to the amplifier.
The cooling of the capacitors.
The quality of the capacitors. (the purity in all used materials)
The are some information of this here:
Capacitor Characteristics
Power supply considerations here:
http://www.profusionplc.com/images/d...ecf10-demo.pdf
  Reply With Quote
Old 19th July 2012, 03:05 AM   #27
gootee is offline gootee  United States
diyAudio Member
 
Join Date: Nov 2006
Location: Indiana
Blog Entries: 1
This has the makings of a Really Great Thread, Nico.

Only tommy1000 seems to have touched on the inductance of the power and ground connections between the reservoir caps and the active devices they are supposed to supply.

I believe that decoupling caps, across (and very close to) each and every active device's power/gnd connections, are even more important than the reservoir caps (or, at least they're at least as important).

Why? Transient Response accuracy, which equates to proper phase and time alignment of all frequency components, and is probably responsible for all of the good sound qualities and nuances that obviously cannot be estimated by measuring simple "distortion" (THD+N, IMD, etc) of the Steady-State Response. (Accurate transient response also prevents things like overshoot and ringing. If you purposely slightly-misalign the phase angles of the Fourier components of a square wave, even just in a monotonically-frequency-dependent way, you can get what looks EXACTLY like overshoot and ringing after the leading edge of the square wave. It makes sense, eh?)

Rail inductance makes it impossible for the reservoir caps to do the Transient Response accurately. And if they are forced to try, large rail voltage disturbances will result from V = L di/dt, from the inductance L of the supply and ground rails.

Are ripple and other rail-voltage disturbances "bad" for sound quality? It depends on the PSRR (which usually gets worse with higher frequency, et al). Obviously, ripple + L(di/dt) disturbances COULD be bad for sound quality. After all, all types of transistors are simply voltage-controlled or current-controlled RESISTORS, i.e. electronically-controllable "current valves". (And "the signal path" that we actually hear is the current, which goes from the power supply reservoir caps and the decoupling caps, through the power transistors, to the speakers, and back again.)

Anyway, we go to a lot of effort to accurately set the resistance of the transistors, according to the input signal, but the actual current that flows from the caps to the speakers then ALSO depends on the voltage across the transistor, since I = V/R.

So unless your circuit is tracking the power supply ripple and adjusting its control of the transistors to correct the resistance it sets them to, at each instant, based on the variations of Vsupply (as seen at the transistor), then the current (I) will not be the V/R that was intended, but would be (V+Vripple+L(di/dt))/R. Feedback of the output signal should also take the power supply voltage variations into account, but not "directly"; basically it's after the problem is already occurring. So it's probably much better to avoid the supply rail and ground rail voltage disturbances ahead of time.

It all boils down to being mindful of the power supply characteristics, AT the power transistors' connections (i.e. NOT back at the output of the PSU, which, for transient response, might as well be miles away if there are several inches or more of supply and ground conductors).

So we can look at that in both the frequency domain (is the impedance seen, when looking from the point of view of right AT each transistor's local power supply connections, staying low-enough, up to a high-enough frequency?) and the time domain (possibly mostly in terms of: a) is there enough local capacitance for worst-case transient current demands to be met accurately-enough and with low-enough induced voltage disturbances? b) what is the maximum inductance that can be tolerated in the local capacitance's connections to the device, to still achieve (a)?, and c) can the distant reservoir caps supply everything else accurately-enough?).

Anyway, there are relatively-easy ways to calculate everything I have mentioned. And the example calculations I have done so far do point to specific physical layout characteristics that should be considered, possibly much more than they usually seem to be, mainly in terms of getting enough capacitance close-enough to where it's needed, so that the total inductance is low-enough.

I haven't finished my work on that, yet, but so far it appears that it's surprisingly easy to not do it well-enough, which could be one of the reasons that some amps that measure extremely well in terms of THD and IMD just don't sound as wonderful as some might think that should imply.

If anyone is interested, here are most of the posts where I started trying to develop the equations for minimum capacitor size and maximum tolerable connection length:

paralleling film caps with electrolytic caps

paralleling film caps with electrolytic caps

paralleling film caps with electrolytic caps (PSU impedance, at the load, AS LOW AS YOU WANT: parallel copies of supply and ground rails, from each reservoir cap all the way to each decoupling cap and the point of load)

paralleling film caps with electrolytic caps

paralleling film caps with electrolytic caps (downloadable LT-Spice model files)

paralleling film caps with electrolytic caps (very interesting, about reservoir caps)

+/-30vDC @ 10A PSU

Cheers,

Tom
  Reply With Quote
Old 19th July 2012, 03:43 AM   #28
diyAudio Member
 
Nico Ras's Avatar
 
Join Date: Nov 2005
Location: East Coast of South Africa
Thanks for the contribution Tom, interesting material, I will follow the threads you posted.
__________________
Kindest regards
Nico
  Reply With Quote
Old 19th July 2012, 03:52 AM   #29
diyAudio Member
 
Nico Ras's Avatar
 
Join Date: Nov 2005
Location: East Coast of South Africa
Tom,

About two years ago MiiB (Michael) and I was exchanging ideas on an amp design and he came forward with a PCB layout that had a reservoir cap of 10000uF between each power device. i.e. there was a large cap sitting a few mm from the active device.

I thought this was quite a novel idea at the time. I have not tried it in reality but it seems to tally with what you are promoting here Tom. It may not only be what is in the power supply but how it is distributed that makes the difference.
__________________
Kindest regards
Nico
  Reply With Quote
Old 19th July 2012, 04:55 AM   #30
gootee is offline gootee  United States
diyAudio Member
 
Join Date: Nov 2006
Location: Indiana
Blog Entries: 1
Yes, I believe that the distribution is extremely important and your friend was on the right track.

One improvement that now seems obvious would have been to use several smaller caps in parallel, in place of each 10kuF cap.

For large capacitance values, paralleling N smaller caps is almost always much better, especially if their connections are also paralleled (i.e. kept separate) as much as possible. Then you get 1/N x inductance, 1/N x ESR, and N x capacaitance, all at the same time.

Also note that I was doing everything I mentioned above with the idea that most DIY people can not use six-layer PCBs with real power and ground and signal and guard planes. But if we use many smaller reservoir caps in parallel, AND have a pair of SEPARATE parallel power and ground conductors for EACH reservoir cap, that run all the way to the active load, and also have one of the (multiple parallel) decoupling caps across each of them, on the load end, and we replicate that as many times as we can with the space available, then we can get PSU and transient performance that is as good or better than what we could get with power and ground planes. (The reason it could be better is that the designer of the planed version might not be able to arrange his caps so that all their currents take separate paths, on the power and ground planes, which would cause mutual inductance and the maximum inductance reduction due to paralleling would not occur. Or, he could be unable to distribute multiple parallel power or ground connections to each of the planes, which would have a similar effect.)

For my next DIY power amp project, I am envisioning multiple ribbon cables for the power and ground rails
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Valve power supply - How to size transformer? SanderW Power Supplies 25 4th January 2013 04:12 PM
How do you calculate choke size in a power supply? Original Burnedfingers Tubes / Valves 25 5th January 2012 12:23 AM
power supply bypass cap size BigE Power Supplies 11 5th July 2011 02:59 PM
Power Supply Case Size diymixer Power Supplies 1 10th October 2010 05:47 AM
What size power supply should I get for repair work? spooney Car Audio 3 6th December 2007 11:50 PM


New To Site? Need Help?

All times are GMT. The time now is 12:31 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2