John Curl's Blowtorch preamplifier part II - Page 1600 - 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 14th September 2011, 05:47 PM   #15991
BV is offline BV  Slovakia
diyAudio Member
 
Join Date: Sep 2005
Location: Trnava, Slovakia
Here are two pictures, one is simulation of IHD for my amp, second is measurement for the same , realized circuit. I found according this simulations quite reliable, if properly used with realistic models and all faults in realization are avoided (PCB, grounding, etc..).
2 PMA
What about posting here Your .cir file?
Attached Images
File Type: png 1k100Wsim.png (19.2 KB, 151 views)
File Type: png 100W-1k.png (11.2 KB, 151 views)
  Reply With Quote
Old 14th September 2011, 05:48 PM   #15992
jcx is offline jcx  United States
diyAudio Member
 
Join Date: Feb 2003
Location: ..
I’ve used the free Ltspice sim, developed, supported, used internally by a leading analog IC house
Linear Technology - Design Simulation and Device Models

anyone with a PC, Laptop can download run this – Mac users may need Wine

I posted my sims source files, included the Q models on the schematic

I have publicly requested PMA’s models, I think this would be “fair use” for copyright issues

My sims can be inspected for bias points, modified, any test waveform you want to try, tweaking Spice simulation/fft plot parameters



It seems a little premature to move to the exits as it were, muttering “inconclusive, nothing proved...”

when so far "2 out of 3 sims agree.." just not with the inner loop feedback R helping (and only if you count "full" circuit sims, dismissing my simplified model - we use simplified models because they are a lot less work)


The usual jc3full sim, green v(out)@1 is without inner feedback, yellow @2 trace is with inner feedback resistors

The pic is a little large but I wanted to include the 1K and still have everyone be able to see the higher frequency IMD skirts too

Now is someone going to point out that the inner feedback made the 1 kHz IMD product “better” by 1.5 dB while ignoring the faster decay of the higher order difference "skirt" envelope?
Attached Images
File Type: png jc3fullfft2.PNG (93.1 KB, 151 views)

Last edited by jcx; 14th September 2011 at 05:54 PM.
  Reply With Quote
Old 14th September 2011, 05:52 PM   #15993
PMA is offline PMA  Europe
diyAudio Member
 
PMA's Avatar
 
Join Date: Apr 2002
Location: Prague
I do not think so.
__________________
Pavel Macura
http://web.telecom.cz/macura/audiopage.html
  Reply With Quote
Old 14th September 2011, 05:58 PM   #15994
jcx is offline jcx  United States
diyAudio Member
 
Join Date: Feb 2003
Location: ..
Default its not my score keeping

Quote:
Originally Posted by ThorstenL View Post
Folks,

So let me summarise where this whole set of simulations regarding the JC-3's inner loop has left us:

1) We have several simulations that use more modern and arguably by far more linear parts and are more or less similar to the original design. Non have shown either 19 + 20KHz IMD spectrae, 76 + 80KHz Spectra or TIM diagnostic signal waveforms. All Show 1KHz or 10KHz HD only.

2) One Sim shows no appreciable difference between the inner loop connected and not at 1KHz, suggesting to me that the Output Stage transistor models are not very good, as even in Class A an EF2 stage is not THAT low distortion that 40dB+ change in feedback factor for the output stage have no effect whatsoever at lower order HD.

3) Another Sim shows that connecting the inner loop reduces higher order HD at the expense of more lower order HD (2nd to be specific) which would be case if the higher order terms are a mixture of multiplication products between multiple distorting stages and "re-entrant" distortion.

4) A third Sim using supposedly the same software, settings and models shows the opposite of the second sim (3) and was posted after the second one appeared to show that in fact the inner loop works as advertised. It shows that in fact things are as those promoting the superiority of the global loop claim...

Ciao T
or is it that you don't think that the models can be shared?

or that maybe we could work out sim setting, potential device bias, test levels, "leveling" assumptions issues if we were using the same models, even in different Spice sw

Last edited by jcx; 14th September 2011 at 06:11 PM.
  Reply With Quote
Old 14th September 2011, 06:17 PM   #15995
Previously: Kuei Yang Wang
 
Join Date: Nov 2002
Hi,

Quote:
Originally Posted by jcx View Post
It seems a little premature to move to the exits as it were, muttering “inconclusive, nothing proved...”
Why not.

Quote:
Originally Posted by jcx View Post
when so far "2 out of 3 sims agree.."
No, sorry, that is not what I noticed. Your sims showed basically "no real difference", based on you wrote...

Quote:
Originally Posted by jcx View Post
The usual jc3full sim, green v(out)@1 is without inner feedback, yellow @2 trace is with inner feedback resistors
This is 19 & 20KHz? Gosh this looks bad, what is wrong with your sim?

I can barely tell there are two signals!? Is that something that LTSpice does? I expect a rather different picture in my old (and now non-functional) copy of P-Spice. I was considering a switch to LTSpice, but looking at that I better buy an old 2nd hand Laptop to run Win XP and PSpice on...

Ciao T
  Reply With Quote
Old 14th September 2011, 06:20 PM   #15996
Previously: Kuei Yang Wang
 
Join Date: Nov 2002
Hi,

Quote:
Originally Posted by stinius View Post
In other words: incompetent design
Maybe, maybe not. You would not know...

Maybe I just do not use enough feedback?

Then again, how do I get a stable Amp with 60dB Loop Gain at 100KHz?

Ciao T
  Reply With Quote
Old 14th September 2011, 06:22 PM   #15997
BV is offline BV  Slovakia
diyAudio Member
 
Join Date: Sep 2005
Location: Trnava, Slovakia
Quote:
We have several simulations that use more modern and arguably by far more linear parts and are more or less similar to the original design. Non have shown either 19 + 20KHz IMD spectrae
With and without 1meg resistors, lower values are of course without those resistors.
Attached Images
File Type: png JC3 CCIF.png (29.4 KB, 146 views)
  Reply With Quote
Old 14th September 2011, 06:37 PM   #15998
jcx is offline jcx  United States
diyAudio Member
 
Join Date: Feb 2003
Location: ..
Default windowing in fft is a trade off, selectable in Ltspice

Blackman windowing - pushes down the "hair" of spectral leakge "floor", in exchange it widens the lines, takes up 4 "bins" for the windowing skirt to fall to the sim's numerical fft "noise floor"

the sim is only 5x periods of the 1 kHz difference so the Blackman "fattening" of the spectral lines means the overlap doesn't drop to the fft floor, the windowing skirt is ~ 20 dB down from the peak at 400 Hz offset = 2 "bins" for a 5 ms sim

the "points" of the spectral lines are what is relevant - there are no mechanisms generating less than 1 kHz differences in this sim

I chose Blackman rather than the defult becouse I wanted to "look lower" - I could have lengthed the sim but was pushing tol, t_minstepsize to the point of taking 10s of minutes already

Last edited by jcx; 14th September 2011 at 06:41 PM.
  Reply With Quote
Old 14th September 2011, 07:00 PM   #15999
diyAudio Member
 
scott wurcer's Avatar
 
Join Date: Jan 2004
Location: cambridge ma
Quote:
Originally Posted by jcx View Post
I could have lengthed the sim but was pushing tol, t_minstepsize to the point of taking 10s of minutes already
OUCH!

I point out that taking the input stage out of the circuit and estimating its output impedance has a problem, in situ the resistance seen by a current into the high Z node is ~1.6K vs ~1.8k in the two cases due to the global feedback. This can be estimated on the back of a small envelope.

Another problem I noticed is that this is one of those "re-entrant" distortion problems with a cascade of products and is VERY sensitive to the input level, 1dB at fundamental is 10 or so at the 9th harmonic.
__________________
"The question of who is right and who is wrong has seemed to me always too small to be worth a moment's thought, while the question of what is right and what is wrong has seemed all-important."
  Reply With Quote
Old 14th September 2011, 07:05 PM   #16000
jcx is offline jcx  United States
diyAudio Member
 
Join Date: Feb 2003
Location: ..
just a test - once nothing new is found I go back to default settings, min time step to give only a few 100k points

Last edited by jcx; 14th September 2011 at 07:07 PM.
  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



New To Site? Need Help?

All times are GMT. The time now is 04:21 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