Power Supply Resevoir Size - Page 88 - 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 29th August 2012, 12:22 PM   #871
diyAudio Member
 
Join Date: Feb 2005
Frank,

Could you repeat the sim with an added 100mR in series as per a typical C-R-C supply - very informative picture with the current charging pulses so clearly shown - is there any provision to show the diode turn-off reverse spike so might be able to extend the model to show the behaviour of a typical Hagerman snubber, etc ...
__________________
... jh
  Reply With Quote
Old 29th August 2012, 01:17 PM   #872
DF96 is offline DF96  England
diyAudio Member
 
Join Date: May 2007
Quote:
Originally Posted by gootee
So hey, DF96 (or anyone): "Got equations?"
I have some approximate equations for the simplest case, in which the transformer series resistance R is sufficiently low that the CR product (C is reservoir cap) is significantly shorter than the available recharge time so the cap more or less fully recharges each half-cycle. This would roughly correspond to standard engineering design, I think. Still working on the case where R or C is higher. Very high C roughly corresponds to 'modern' or 'audiophile' design, where if C is good then 100C 'must' be better.
  Reply With Quote
Old 29th August 2012, 04:51 PM   #873
AndrewT is offline AndrewT  Scotland
diyAudio Member
 
Join Date: Jul 2004
Location: Scottish Borders
Quote:
Originally Posted by jameshillj View Post
...........very informative picture with the current charging pulses so clearly shown - is there any provision to show the diode turn-off reverse spike so might be able to extend the model to show the behaviour of a typical Hagerman snubber, etc ...
I would consider this type of sim is more appropriate to modeling the amplifier and examining the PSRR and such.
Not here where we are looking at a different effect.
__________________
regards Andrew T.
  Reply With Quote
Old 29th August 2012, 07:09 PM   #874
gootee is offline gootee  United States
diyAudio Member
 
Join Date: Nov 2006
Location: Indiana
Blog Entries: 1
Quote:
Originally Posted by jameshillj View Post
Frank,

Could you repeat the sim with an added 100mR in series as per a typical C-R-C supply - very informative picture with the current charging pulses so clearly shown -...
Jameshillj,

The "load-current pulses" very-often have a greater effect on the rail voltage than the charging-current pulses.

You might have been looking at the effects of the load-current pulses (unless you're looking at the only two wide ones, wich are centered at roughly 88.5 ms and 97 ms; those are the effects of the charging pulses). The load current pulses cause the rail to drop. So the "downward-pointing" voltage spikes/pulses are caused by the load-current pulses that the transistors are allowing to flow out of the positive rail.

Check the time scale at the bottom. The faster-occurring pulses are 1 ms apart, i.e. 1 kHz. And he said he applied a 1 kHz sine signal that gave the maximum output amplitude.

Quote:
...is there any provision to show the diode turn-off reverse spike so might be able to extend the model to show the behaviour of a typical Hagerman snubber, etc
Wow, man, you really need to download LT-Spice (free from linear.com). You don't need additional "provisions", to look at something. You can show almost ANYTHING.

The mouse pointer instantly turns into a PROBE(!) as soon as it moves over the schematic, which I posted just a few posts back, and which is also at http://www.diyaudio.com/forums/attac...tic_square.jpg .

And with a quick right click on a wire and then a left click on a pop-up menu, you can set the probe reference point anywhere you want, if you don't want it referenced automatically to ground.

Also, moving the probe over the top of a two-terminal device, or to a pin of a multi-pin device, changes the probe into a CURRENT probe.

Click whatever or wherever you want on the schematic and it's instantly plotted. Watch the plots unfold as the simulation time progresses, if you're simulating in the time domain. But it's just as easy to switch to the frequency domain.

Note that there IS a snubber in the circuit, already (but not aimed specifically at the turn-off spike): R2 is the snubber. The optional C in series with it was unnecessary in this case because the optimal dampng resistance was so high that its dissipation was low-enough that no frequencies needed to be restricted from it, making the optional series C unneeded.

The snubber was needed in an earlier version of the circuit in order to damp out a very strong and persistent ringing that was at a fairly-low frequency (below 1 MHz). I can't remember the details, right now. But the circuit changes whenever the transformer model is reconfigured or the reservoir capacitance is swept.

When simulating, I usually only bother with designing a snubber if it is needed in order to prevent the sim from slowing to a crawl. I haven't checked this one without the R2 installed, lately, but it's a high-enough R value that I just left it in.

By the way, that snubber was designed by using the steps I posted at:

paralleling film caps with electrolytic caps

(I realize that the snubber looks a little weird, with such a large R value. But it did work extremely well.)

I have seen an almost-identical snubber-design method in several different papers, and also in some other places, on the web. I read Jim Hagerman's snubber stuff "a few" years back, along with a lot of other snubber materials, but I can't remember if his method was similar. Anyway, did he invent a special case or type of snubber, since then? (Otherwise I would have to wonder why snubbers should be named after him.)

Regards,

Tom
  Reply With Quote
Old 29th August 2012, 07:51 PM   #875
diyAudio Member
 
Nico Ras's Avatar
 
Join Date: Nov 2005
Location: East Coast of South Africa
Tom & Frank, I am sorry I have run out of nice things to say. But be aware I have not missed a single post nor missed making notes that I feel could eventually be helpful in the conclusion or dying posts in this thread. Thanks again guys.

I have laid a little headphone amp board to the test the "sonic" effect of multiple smaller caps as I am currently using the same amp with a single cap per rail.

I feel headphones are the best test system because you have some of the worlds best speakers only a centimetre from your ear and no room acoustics or furnishings or any other obsticle to affect what you hear.
Attached Images
File Type: jpg amp.jpg (260.9 KB, 154 views)
__________________
Kindest regards
Nico
  Reply With Quote
Old 29th August 2012, 10:47 PM   #876
diyAudio Member
 
Join Date: Feb 2005
Thanks Tom, my old CRO has trouble producing such clearly defined signals with all the diode and supply line hash messing up the picture - I see the loading influence and it still surprises me to see how much the rail is modulated with the input/output signal of the F5 classA amp, for example, and the different effects of various capacitors types and sizes in the power supplies.

You're right, I'll have to have another go at sorting out the LT Spice tool - my computer skills are limited and interpreting the results require skill and knowledge that is sadly lacking here - very much appreciate your, and everyone else, detailed info.

Jim Hagerman was the paper "of my day" and just about everyone that is interested in snubbers in "hifi land" has at least heard of it, so it's a 'sort of' reference - as you say, many other studies and papers around - don't think there's been any further developments except better definition/recognition of the problems, particularly in conjunction with the falling quality in power distribution and transformer quality.

A bit off subject, did you complete your version of George's Lightspeed Vol Control a few years ago, and if so, how did it work out?
__________________
... jh
  Reply With Quote
Old 29th August 2012, 10:56 PM   #877
fas42 is offline fas42  Australia
diyAudio Member
 
fas42's Avatar
 
Join Date: Jun 2012
Location: NSW, Australia
Blog Entries: 11
Thanks for clarifying things, Tom; trying a C-R-C variation of the type suggested by james would show nothing, because its impact would be swamped by the presence of other behaviours.

This whole business is an intricate dance, participated in by the non-perfect behaviour of the mains, the parasitics of the power supply, the PSRR of the circuit driven, and finally the typically complex impedance of the speaker load. There are no straight answers here, which is why when people start fiddling with things in this area the sound quality changes ...

Frank
  Reply With Quote
Old 30th August 2012, 01:11 AM   #878
gootee is offline gootee  United States
diyAudio Member
 
Join Date: Nov 2006
Location: Indiana
Blog Entries: 1
Quote:
Originally Posted by fas42 View Post
Thanks for clarifying things, Tom; trying a C-R-C variation of the type suggested by james would show nothing, because its impact would be swamped by the presence of other behaviours.

This whole business is an intricate dance, participated in by the non-perfect behaviour of the mains, the parasitics of the power supply, the PSRR of the circuit driven, and finally the typically complex impedance of the speaker load. There are no straight answers here, which is why when people start fiddling with things in this area the sound quality changes ...

Frank
<rant>Ten minutes of typing a reply just %$#&%$# disappeared! (And yes, I know about Undo, and Ctrl-Z and Ctrl-Y, and Back and Fwd, and I usually periodically do a Select All and Copy.) Shouldn't there at LEAST be a &%$%#@ keyboard buffer, or SOMETHING, that allows one to get back what they have just typed?!</rant>

Anyway, yes, it's deceptively simple-looking, which has been mentioned in every one of the papers I've read where they derive the "real" equations for the "simple" transformer, rectifier, and capacitor circuit. It always ends up being many PAGES of complex integral and differential equations, with many transcendental functions in them, and also seemingly always some integrals that "have no closed-form solution". And none of the ones I have found has even attempted to include an active load.

If I ever get around to it, I will probably try to just substitute every transcendenal function with its Taylor Series Expansion, which turns everything into polynomial forms that can be made as accurate as needed by just adding on higher-order terms. That way, everything would be trivial to integrate and differentiate and we would end up with a system of polynomial equations that would be a true model (or "true-enough", at least), and could be crunched by any spreadsheet software. We could even get things like three-dimensional plots that could bound the usable values of several variables after others were already chosen, for example. The possibilities are almost endless, and very attractive.

Anyway, speaking of "...the sound quality changes", our model does have the capability, right now, of varying things (either automatically or manually) and then looking for resultant values or changes or conditions in the output signal.

And if we use WAV files for both input and output, we can COMPARE the input and output files to each other, or compare pairs of output files to each other, automatically, using software like diffmaker (a free download).

AND, we can also LISTEN TO the output files!

BUT, we don't know a lot about WHAT to measure, from within spice simulations, for example, in order to be able to predict "sound quality", unless it would just be something like looking for the minimum sum or average of the arithmetic differences between output and input WAV files or voltage plots. The calculated THD could be helpful but (I think) we think we know that low THD alone is not a good predictor of sound quality. Maybe it's "necessary but not sufficient". (But it is "easy". Right now, tonight, you could do things like generating plots of THD versus capacitor ESR or ESL, for example, or anything else in the model.)

So, anyway, I can envision a scenario where, initially, someone could be cranking-out WAV files, with LT-Spice simulations, and sending them to a team of volunteer "listeners", who could somehow rate them for "sound quality".

That might eventually enable someone to discover some measurements, or, maybe more-likely, even just simple configuration or parameter changes, that DO correlate with sound quality. That information could then be used to better-inform design-parameter choices/tradeoffs, or optimization schemes, etc etc.

A lot would probably depend on the sequences of what was tested and rated. There are people here, like SY, who seem to know a lot about the proper design of "human-based rating systems". But, as a first guess, think about when you go to get eyeglasses: They use a sequence of "binary" choices, where you are always rating which is the better of only TWO options at a time, and you eventually converge on the best one. (It's the old familiar "binary tree" concept that is/was used so much in things like database search algorithms.) But WHAT was to be compared and rated would still need to be determined. Yes it could be complex and difficult to design that part of the expriment. But it could also generate quick successes. And there is already a large framework of opinions and ideas available.

I could continue to make-it-up-and-elaborate-as-I-go, but that's all just a sleepless dream, at this point, anyway. However, it might be the beginning of finding a way to "leverage" some capabilites (i.e. 1. automation using simulation: cheaper, faster, and more repeatable and controllable than alternatives, and 2. cooperative efforts of many people, using the internet to quickly and easily send out the test files and receive the results) in order to greatly speed-up the process of developing some understanding and some methods that could be useful in improving everyone's ability to repeatably generate better audio system designs.

I'm not quite sure how I managed to diverge to that point. Oh well.

Later,

Tom

Last edited by gootee; 30th August 2012 at 01:20 AM.
  Reply With Quote
Old 30th August 2012, 01:29 AM   #879
tsiros is offline tsiros  Greece
diyAudio Member
 
tsiros's Avatar
 
Join Date: Jun 2012
Location: Patra, Greece
Send a message via MSN to tsiros Send a message via Skype™ to tsiros
Quote:
Originally Posted by Nico Ras View Post
I feel headphones are the best test system because you have some of the worlds best speakers only a centimetre from your ear and no room acoustics or furnishings or any other obsticle to affect what you hear.
if the recording is binaural, hells yeah! If it is a normal recording, not so much.
  Reply With Quote
Old 30th August 2012, 03:24 AM   #880
fas42 is offline fas42  Australia
diyAudio Member
 
fas42's Avatar
 
Join Date: Jun 2012
Location: NSW, Australia
Blog Entries: 11
Not a bad effort after being thoroughly p***ed off by the joys of normal IT engineering, Tom! I feel for you, having been there many times before, it's always the pieces where one puts a lot of energy and enthusiasm into it that go down the gurgler ...

I might invent a new phrase, something along the lines of ... "Tomorrow is another day!". How do you reckon that'll go down ...?

Frank
  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 07:37 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