Go Back   Home > Forums > >

Solid State Talk all about solid state amplification.

CFA Topology Audio Amplifiers
CFA Topology Audio Amplifiers
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
Old 29th November 2013, 06:57 AM   #2271
PMA is offline PMA  Europe
diyAudio Member
 
PMA's Avatar
 
Join Date: Apr 2002
Location: Prague
OK, this was the section of the power amplifier board, the caps are about 5 cm from VAS and OPS
  Reply With Quote
Old 29th November 2013, 07:36 AM   #2272
astx is offline astx  Austria
diyAudio Member
 
Join Date: Jan 2011
Location: Tyrol / Austria
CFA Topology Audio Amplifiers
Quote:
Originally Posted by kgrlee View Post
The weak link was the ribbon cable. Though Toni maintained that it was of no consequence .. for the sims, simulating the ribbon cable was essential to show some of the effects he found.

To get SPICE and 'real world' closer would have require drastic revisions to his PCB(s) and physical layout. ie a single PCB with VAS & OPS close together. Too many possible trace inductances & capacitances on his present complicated layout for me to pontificate.

FWIW, it was the ground track on the ribbon which was important. The others, including the leads from VAS collector to OPS were relatively problem free, with or without distributed capactiance to go with the inductance.
...
Thought I have shown several times that the length of the ribbon cable has only a capacitive influence on the amp. A longer ribbon cable bettered the stability because the higher capacitance to ground has added some sort of VAS compensation (today I know that this slightly bettered the gain margin - see post #701 - The effect of too less gain margin). All ground wires over the ribbon cable where dirty ground. Clean ground never has been routed to output stage via ribbon cable!
BTW the real world amplifier uses TMC compensation and is pretty stable playing fine every day.
Think if you would recalculate the cherry compensation on the latest of my simulations (adding 2x28pF from VAS output to ground) and design it for a GM > 10dB and a PM > 60 dB maybe cherry compensation will work too. PM me if you want pcb samples.

Quote:
Originally Posted by kgrlee View Post
...
Electrolytics cos damping is vital and their ESR is almost ideal for this. ie bigger caps need smaller ESR.

These need to decouple to a Dirty Earth that goes back to the Star Point via the same route as the Power Rails. Toni re-routed one of his decoupling caps on my recommendation and got better results.
...
Yes I have re-routed 2 pieces of 0.1µF caps from clean ground to dirty ground (via ribbon cable to output stage).
But the most important re-routing may have been the collector of one of the VAS transistors from clean to dirty ground.

IMHO: to put the sensitive input/VAS stage away from all those big currents is better for those big amplifiers to get best THD and Noise values!

BR, Toni
  Reply With Quote
Old 29th November 2013, 02:36 PM   #2273
kgrlee is offline kgrlee
diyAudio Member
 
Join Date: Nov 2011
Location: Cooktown, Oz
Quote:
Originally Posted by astx View Post
Thought I have shown several times that the length of the ribbon cable has only a capacitive influence on the amp. A longer ribbon cable bettered the stability because the higher capacitance to ground has added some sort of VAS compensation (today I know that this slightly bettered the gain margin - see post #701 - The effect of too less gain margin). All ground wires over the ribbon cable where dirty ground. Clean ground never has been routed to output stage via ribbon cable!
BTW the real world amplifier uses TMC compensation and is pretty stable playing fine every day.

Think if you would recalculate the cherry compensation on the latest of my simulations (adding 2x28pF from VAS output to ground) and design it for a GM > 10dB and a PM > 60 dB maybe cherry compensation will work too. PM me if you want pcb samples.
Toni, I wasn't criticising your design. As I have said many times, it is excellent.

What I was referring to is that once we got to a certain stage, the sims no longer reflected your 'real world' results.

eg simulating the higher capacitance on the ribbon didn't make this closer to 'real world'

Once the sim no longer reflects 'real world' you have to be very careful to make predictions based on it. In the early days of the thread I suggested some Cherry mods which proved good. Your 'real life' results were in line with the sims.

But as the performance level rose, this was no longer the case.

After simulating the ribbon with some success, I couldn't go any further without asking for even greater physical topology revisions. So I stopped.

There's no point my suggesting more Cherry compensation ideas until I get my sim to reflect your latest results.

My main reason for participating in your thread was to get some 'real life' confirmation of SPICE world Cherry sims. My thanks for being so patient with me and indulging some of my crazy ideas.

I learnt a lot from the exercise including what NOT to do. Also, this was my really big learning experience with TMC.

Quote:
Yes I have re-routed 2 pieces of 0.1µF caps from clean ground to dirty ground (via ribbon cable to output stage).
But the most important re-routing may have been the collector of one of the VAS transistors from clean to dirty ground.
Yes. That was the important one.
  Reply With Quote
Old 29th November 2013, 03:11 PM   #2274
Bob Cordell is offline Bob Cordell  United States
diyAudio Member
 
Join Date: Sep 2006
Quote:
Originally Posted by Dave Zan View Post
Coupled via the track resistance or inductance?



I should have stated that I plan 10,000 uF main power supply electrolytics, paralleled one per OPS transistor.
Then on-board SMD electrolytics probably around 100 uF, also one per OPS transistor.
Plus any extra bypass if needed, this I don't know about yet.



Yes, had already worked this out for inductance/ stray field reasons.
Seems obvious but nice to have it checked.



Hi temp. here for sure. 105 C presumably.



Have seen at least one study of capacitors in parallel where they didn't behave as one would naively expect.
But 1 uF next to the 100 uF electrolytic seems plausible. Any advice on that?

Best wishes
David
Hi David,

I always like to put about 1000uF per rail on the amp board close to the output transistors. If one does this, the wiring length to the big PSU reservoir caps does not matter much. These should of course be bypassed with iuF or so. These caps pretty much take care of the HF issues, making the arrangement far less dependent on the wiring to the PSU for distortion, etc.

An interesting half-crazy idea is to go far non-conventional when one does this, not using big wires to the PSU, but rather small-tightly coupled ones, including ground. The little bit of resistance may actually help matters and even give a little additional filtering from the PS at HF. Consider shielded microphone cable, wiring the rails on the inside wires and using the shield as the ground return to the PSU reservoir caps. When doing this, I return the speaker output to the amp board at the center of the on-board caps rather than a star ground at the PSU when I do this.

I realize that Self does not recommend twisting together the two rail leads and the ground, instead preferring to twist together only the two rail wires.

I always prototype with relatively long wires to the PSU.

Cheers,
Bob
  Reply With Quote
Old 29th November 2013, 04:18 PM   #2275
31697B is offline 31697B  Thailand
Account disabled at member's request
 
Join Date: May 2012
Local or on-board cap/regulator is always the best approach.

The main filter/energy storage caps can be mounted with the terminals facing downward so they are close to the chassis and the PS and return wiring run against the chassis. Ground the caps' metal case also, if it has one, to the chassis via the hold-down mounting hardware. if mounted in the typical way with terminals facing upward, run the PS wires to it against the chassis and close up against the sides of the grounded cap case.

-Thx-RNMarsh

Last edited by 31697B; 29th November 2013 at 04:24 PM.
  Reply With Quote
Old 29th November 2013, 07:23 PM   #2276
kgrlee is offline kgrlee
diyAudio Member
 
Join Date: Nov 2011
Location: Cooktown, Oz
Quote:
Originally Posted by Bob Cordell View Post
An interesting half-crazy idea is to go far non-conventional when one does this, not using big wires to the PSU, but rather small-tightly coupled ones, including ground. The little bit of resistance may actually help matters and even give a little additional filtering from the PS at HF.
This isn't crazy at all. The limit is Cherry's A New Distortion Mechanism in Class B Amplifiers where he puts series inductance in the PSU leads with large local electrolytics.

I don't think its necessary to go to his lengths (the inductors) but this was the paper that refined my thinking on the subject.

He does this on his ETI NDFL amp that is on jan.didden's website. Loadsa useful practical details therein.

In da old days, it was difficult to find good caps small enough to put near the OPS & VAS. 220u was the biggest I could squeeze in. I remember marvelling at some really supa dupa full power THD20k measurements on a 50W prototype .. until the local caps blew up.

Last edited by kgrlee; 29th November 2013 at 07:35 PM.
  Reply With Quote
Old 30th November 2013, 12:45 PM   #2277
ontoaba is offline ontoaba  Indonesia
diyAudio Member
 
ontoaba's Avatar
 
Join Date: Nov 2009
Location: Kudus, Malang, Dieng
Quote:
Originally Posted by kgrlee View Post
Which ones have you tried and what did you hear that's not neutral & natural?

I don't see how mentioning the brand can change its neutrality & naturalness
Just buy some opamps then use your headphones. You could see their characteristics through their sheets. If you don't accept 0.01% thd is unhearable, then do the test, may be you could hear additional 0.01% signal added to 100% signal, then just show me the way to do the test.

Quote:
Originally Posted by Esperado View Post
You have a strange way to see the things ;-)
What else than 'distortions' can modify the "sound of a signal" and make an amplifier not 'transparent' ?
An other what ?
Yes it is better called as "another things", it will stay simple with that.

Sorry for late replay, I have problem with my switching AC regulator, the mosfet hit the limiter at certain loads.
  Reply With Quote
Old 30th November 2013, 12:49 PM   #2278
AndrewT is offline AndrewT  Scotland
R.I.P.
 
Join Date: Jul 2004
Location: Scottish Borders
Quote:
Originally Posted by Bob Cordell View Post
.......An interesting half-crazy idea is to go far non-conventional when one does this, not using big wires to the PSU, but rather small-tightly coupled ones, including ground. The little bit of resistance may actually help matters and even give a little additional filtering from the PS at HF. ............
Bob C.
I give you permission to promote my half-crazy idea.
It works.
__________________
regards Andrew T.
  Reply With Quote
Old 30th November 2013, 01:03 PM   #2279
ontoaba is offline ontoaba  Indonesia
diyAudio Member
 
ontoaba's Avatar
 
Join Date: Nov 2009
Location: Kudus, Malang, Dieng
Quote:
Originally Posted by RNMarsh View Post
What is a basic topology for cancellation and what variations exist? We can cover what others have done to help get the creative juices flowing but would like some fresh ideas as well.


Thx-RNMarsh
Most of them are added (predicted or measured) modulating error to the output(like wavebourn did, also that jfet), and adding measured error back to the input(like bob Cordell did).
  Reply With Quote
Old 30th November 2013, 04:50 PM   #2280
MarcelvdG is offline MarcelvdG  Netherlands
diyAudio Member
 
Join Date: Mar 2003
Location: Haarlem, the Netherlands
Quote:
Originally Posted by Edmond Stuart View Post
/OT

This is a funny remark for several reasons:
1. Why would a 'harmonic mean rule' be much nicer than a 'product rule'?

2. Contrary to what the LT1166 datasheet says, this chip does keep the minimum OP current at about half the quiescent current (see: Comparison of a few techniques).

3. Marcel's version of the auto-bias circuit (see: ‘Audio power with a new loop’, EW, Feb. 1996, pp.140-143),
behaves more or less the same as the LT1166.

Cheers, E.
Hi Edmond,

I never actually used the LT1166, my comments are based on the information I read about it when it was introduced. If Linear Technology implemented a harmonic mean rule or something similar in the LT1166 but state in its data sheet that they used a product rule, then what I've written is clearly not valid (and Linear Technology then obviously has an internal communication problem).

In the literature you see that many authors prefer non-switching output stages, that is, output stages where the current through one output device approaches some nonzero value (like half the quiescent current) when the current through the other gets large. The supposed advantage is that the inactive device can turn on quickly when it is already biased at a reasonable current rather than switched off entirely or almost entirely. This behaviour is easy to achieve with a class AB control loop by choosing a control rule that lets the current through one device approach some nonzero minimum value when the current through the other output device tends to infinity.

I don't know what effect this has on the distortion measured with sine waves or with actual music, I never compared the distortion of otherwise similar amplifiers with exp(-I1*R)+exp(-I2*R), harmonic mean and product control rules.

By the way, I like the fact that you put a Dutch quote below your posts.

Best regards,
Marcel
  Reply With Quote

Reply


CFA Topology Audio AmplifiersHide this!Advertise here!
Thread Tools

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

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
nx-Amplifier: 100 W CFA topology Amp Bonsai Solid State 35 9th July 2013 04:34 PM
Cambridge Audio 740A topology lduarte1973 Chip Amps 5 27th June 2011 07:32 AM
crossfire cfa-602, car audio amp problem tuneman Car Audio 29 23rd September 2010 05:08 AM
crossfire cfa-602, car audio amp problem tuneman Solid State 1 11th September 2010 01:51 PM
pro amplifiers vs. home audio amplifiers bonsai171 Solid State 26 25th February 2004 08:24 PM


New To Site? Need Help?

All times are GMT. The time now is 10:43 PM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2021 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 14.29%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2021 DragonByte Technologies Ltd.
Copyright ©1999-2021 diyAudio
Wiki