Sound Quality Vs. Measurements - Page 395 - diyAudio
Go Back   Home > Forums > Member Areas > The Lounge

The Lounge A place to talk about almost anything but politics and religion.

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 24th March 2012, 04:26 PM   #3941
dvv is online now dvv  Serbia
diyAudio Member
 
Join Date: Sep 2004
Location: Belgrade, Serbia
Quote:
Originally Posted by ThorstenL View Post
Hi,



Then consider buffering the inputs.



I would probably suggest to instead keep simple ones and bootstrap the CCS AND the cascodes to the common mode signal. It can be done using a minimal number of extra parts and they are all resistors.

The CCS's should do much better (as you effectively keep the voltage conditions under which they work constant) and the cascodes now rigidly lock the voltages across the amplification transistors without common mode related swings.

Ciao T
Thorsten, I do have an active, fully complementary buffer for it. It should be on the "Input board", which is also to contain XLR sockets and associated electronics (optional, i.e. in two versions, just an RCA Cinch buffer, or the same plus XLR).

It's as vanilla as they come, so I didn't bother posting it. I can if there's an interest for it.

As for the CCSs, as you say, since any I use is fed off regulated voltages in all cases, I am probably not likely to not encounter quite a bit of the usual problems with them. Those shown I picked from a Siemens engineer's project from somewhere, I think (but am not sure) it was from an Elektor MOSFET amp article many years ago. I've since used it and it seems very good to me, but I could be wrong.

Regarding your suggestion, I would once again beg for a schematic (hand jotted is just fine), as I am not sure what exactly do you have in mind.
  Reply With Quote
Old 24th March 2012, 07:18 PM   #3942
tvrgeek is offline tvrgeek  United States
diyAudio Member
 
Join Date: Dec 2009
Location: Md
Well, re-ran all the various IPS and VAS bias and current source options. My conclusion is that the distortion is dominated by the output so not much of anything I do has more effect than trading half a dB between even and odd harmonics. The sim puts things at about .003% second and .001% third. Quite a bit better than the original on odd. Need to order a few resistors and NPO caps. This has been a lot of fun, but I still don't know what makes the Rotel sound better to my wife!

I am going to play with all the other ideas on the MX-50 boards as they are a lot easier to work on and don't have expensive outputs.
Attached Images
File Type: png 120-done.png (87.8 KB, 88 views)
  Reply With Quote
Old 24th March 2012, 07:34 PM   #3943
bcarso is offline bcarso  United States
diyAudio Member
 
Join Date: May 2005
Location: Canoga Park, California
Quote:
Originally Posted by tvrgeek View Post
Well, re-ran all the various IPS and VAS bias and current source options. My conclusion is that the distortion is dominated by the output so not much of anything I do has more effect than trading half a dB between even and odd harmonics. The sim puts things at about .003% second and .001% third. Quite a bit better than the original on odd. Need to order a few resistors and NPO caps. This has been a lot of fun, but I still don't know what makes the Rotel sound better to my wife!

I am going to play with all the other ideas on the MX-50 boards as they are a lot easier to work on and don't have expensive outputs.
Try the split compensation as described by Self in LA Vol. 0. One more cap, one more resistor, and the output stage distortion is reduced by being partially inside the integrator loop. I see lots of benefits and so far no drawbacks.
  Reply With Quote
Old 24th March 2012, 08:25 PM   #3944
jcx is offline jcx  United States
diyAudio Member
 
Join Date: Feb 2003
Location: ..
are you referring to "Transitional Miller Compensation" Edmond Stuart re-invented, promoted here? – the R leg of “2-Pole” “T” connected to the output instead of AC gnd as in “2-pole” compensation?

both “TMC” and “2-pole” compensation really require the added gain of the “Beta enhanced”, input buffered VAS to show large benefits

It took several of us lots of posts to get across the fact that "Two Pole" compensation can give near identical results to "TMC" in the Cordell Book thread

which seems at 1st glance to leave little reason to choose between them - Except for the misinterpretation that "TMC" loop gain as naively measured makes it look like you still have "single pole" global loop gain roll off - giving the False impression of greater loop stability than 2-pole comp

my post: Bob Cordell's Power amplifier book shows “TMC” apparent gain/phase margin in global loop measurement falls apart with the added delay sim ( “CMC” = conventional Miller comp )

to properly measure, compare "TMC" and "2-pole" compensation you have to measure total feedback loop gain around the output devices: "inside" the TMC inner loop - both compensation techniques can be adjusted to give very similar loop gain curves, gain, phase margins

when trimmed to "the same" equivalent loop stability margins, loop unity gain intercept frequency they are seen to be nearly equivalent

I prefer 2-pole on the grounds that the added gain is visible in the global loop and therefore additionally linearizes the input diff pair vs TMC

Last edited by jcx; 24th March 2012 at 08:48 PM.
  Reply With Quote
Old 24th March 2012, 08:40 PM   #3945
bcarso is offline bcarso  United States
diyAudio Member
 
Join Date: May 2005
Location: Canoga Park, California
Quote:
Originally Posted by jcx View Post
are you referring to "Transitional Miller Compensation" Edmond Stuart re-invented, promoted here? – the R leg of “2-Pole” “T” connected to the output instead of AC gnd as in “2-pole” compensation?

both “TMC” and “2-pole” compensation really require the added gain of the “Beta enhanced”, input buffered VAS to show large benefits

It took several of us lots of posts to get across the fact that "Two Pole" compensation can give near identical results to "TMC" in the Cordell Book thread

which seems at 1st glance to leave little reason to choose between them - Except for the misinterpretation that "TMC" loop gain as naively measured makes it look like you still have "single pole" global loop gain roll off - giving the False impression of greater loop stability than 2-pole comp

Bob Cordell's Power amplifier book shows “TMC” apparent gain/phase margin in global loop measurement falls apart with the added dalay sim ( “CMC” = conventional Miller copm )

to properly measure, compare "TMC" and "2-pole" compensation you have to measure total feedback loop gain around the output devices: "inside" the TMC inner loop - both compensation techniques can be adjusted to give very similar loop gain curves, gain, phase margins

when trimmed to "the same" equivalent loop stability margins, loop unity gain intercept frequency they are seen to be nearly equivalent

I prefer 2-pole on the grounds that the added gain is visible in the global loop and therefore additionally linearizes the input diff pair vs TMC
Yes, as you describe, transitional Miller (I'd first seen it in the Self reference, probably breezed past it in Cordell...). Haven't followed the thread in here.

I was skeptical that it would help here as much as it does, but I see a ~25dB improvement (!) in sims. I do not have precise models of the output devices tvrgeek is using, but the improvement effect should be more-or-less independent of those.

With a more elaborate design, JFET cascoded input devices*, better transistors all around, beta enhancement stage inserted in second stage, much better current mirror of the "e-follower" type for balancing load currents, bootstrapped cascode for second stage .... I see ridiculous performance from a meter-reader perspective, clearly unbelievable but intriguing nonetheless as it doesn't entail a lot of effort or expense. Of course the input stage and its cascode parts are assumed perfectly matched. I also beefed up the second stage current and as well raised the emitter resistor value so that the bootstrapping would be more effective (to ~55 ohms, 10mA stage current, output CB stage an MPSW92 for better dissipation capability).

Oh, and transient response looks very clean indeed. Frequency response is about -3dBr at 550 kHz at the output (not looking after the L-R decoupling).

*that is, bipolars with the 47 ohm degen Rs and JFETs on top, with the gates tied to the bipolar emitters. Some voltage drop is needed in the JFET drains to avoid exceeding breakdown voltages.

Last edited by bcarso; 24th March 2012 at 08:45 PM. Reason: clarification
  Reply With Quote
Old 24th March 2012, 10:16 PM   #3946
tvrgeek is offline tvrgeek  United States
diyAudio Member
 
Join Date: Dec 2009
Location: Md
Quote:
Originally Posted by bcarso View Post
Try the split compensation as described by Self in LA Vol. 0. One more cap, one more resistor, and the output stage distortion is reduced by being partially inside the integrator loop. I see lots of benefits and so far no drawbacks.
Waiting for payday to send a pile of my money to Jan for his mini-books. I have always found him worth reading. This was car insurance month.
In the mean time, there are about the next 200 pages of the Spice manual to digest.

Gasp, has the price of copper gone up. $200 for the coils for a speaker prototype I am working on. It is almost cheaper to go electronic.

I will probably go ahead and order a handful of 2SA970 and 2SC1085's for the half or so positions that are within their limits. (input and top of cascode) I have not found anything still in production that would seem to better the existing for the rest.

Exicon models attached. I found them on one of these threads. The models ending in "c" thanks to Mr. Cordell.
Attached Files
File Type: txt Ecf10n20.txt (2.5 KB, 12 views)
File Type: txt Ecf10p20.txt (2.5 KB, 9 views)
  Reply With Quote
Old 24th March 2012, 10:19 PM   #3947
tvrgeek is offline tvrgeek  United States
diyAudio Member
 
Join Date: Dec 2009
Location: Md
My models if anyone wishes to play.
Attached Files
File Type: asc DH-120-Repaired-Red-Red.asc (12.7 KB, 7 views)
File Type: asc DH-120-Original.asc (13.2 KB, 6 views)
  Reply With Quote
Old 24th March 2012, 10:40 PM   #3948
tvrgeek is offline tvrgeek  United States
diyAudio Member
 
Join Date: Dec 2009
Location: Md
I have the diagram of the two pole. I have been a little timid as it was not easy getting the real amp to be stable on the bench. 33p CMC with the 3.3p in the feedback did the trick.

I see models of 2SC1775 and 2SC992. I would love to have them.

I did try extending the CMC cap to the driver stage. Sang like a bird. So there is clearly something I don't understand about that. The above link I am going to have to print to study.

That I did not notice much difference adding degeneration on the IPS CM I am assuming is that with perfect parts, they have little effect. The benefit is with real parts? Same with adding a cap across the bias spreader.

The thing I could not improve on in balancing the idle current between the P and N outputs. Once I got the sim to behave, I found whatever I did did not cause much difference, so is 10% "close enough"? , 110mA and 121 mA. Upping to 125 or so in the sim did not do well, but I sure plan on doing it for ear.

the .01 resistors on my drawing are there just so I have a quick place to measure idle current. They are not real.
  Reply With Quote
Old 24th March 2012, 10:53 PM   #3949
tvrgeek is offline tvrgeek  United States
diyAudio Member
 
Join Date: Dec 2009
Location: Md
Modeled the 2 pole. 20dB improvement in even order, 3 dB worse on odd. Now to look at those things that seemed to help odd.
  Reply With Quote
Old 24th March 2012, 11:41 PM   #3950
tvrgeek is offline tvrgeek  United States
diyAudio Member
 
Join Date: Dec 2009
Location: Md
Finally, a change that was dramatic. Upped the IPS bias to 2mA, VAS to 5.7.
Don't understand how the two pole works, but it was a 10 to 20dB improvement and then showed results from the increased bias that it did not before.

Now, it if works on the real amp!
  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
Quality Control differences = variations in sound quality? KT Class D 3 4th June 2014 12:02 AM
Sound Card for Measurements Marik Solid State 2 2nd January 2012 08:59 PM
Sound Card Recommendations (For Audio Measurements) dchisholm Equipment & Tools 5 16th July 2011 09:40 AM
How to protect sound card during amp measurements? okapi Everything Else 13 2nd September 2008 03:06 PM
Sound cards - test and measurements jackinnj Everything Else 2 5th July 2003 03:02 PM


New To Site? Need Help?

All times are GMT. The time now is 10:39 AM.


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