Gain Blockwith Variable Loop Gain - diyAudio
Go Back   Home > Forums > Amplifiers > Solid State

Solid State Talk all about solid state amplification.

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 7th July 2008, 09:50 AM   #1
Bonsai is offline Bonsai  Taiwan
diyAudio Member
 
Bonsai's Avatar
 
Join Date: Jul 2003
Location: Taipei, Taiwan
Default Gain Blockwith Variable Loop Gain

Here is an idea I am trying. I want to create a gain block where the loop gain can be varied continuosly between 0db (zero loop gain) and some arbitrarily high value (loop stability notwithstanding).

The idea is that you can change the loop gain on the fly and then see if an audible difference can be heard.

Here is my 1st shot at it. A pretty conventional CFA topology using a Hawksford VAS stage and diamond buffer.

The loop gain is varied by adjusting the value of R18 (c.f. Hawksfords paper where he mentions this). With R18 set to 1 meg (about 65db OLG), the 20k THD simulates at <1ppm and with R18 set to 1k the 20k THD is about 0.015% - so this would I guess discount distortion as a key reason for the difference in sound, if any is detectable. Of course, some work is needed to ensure the circuit is stable across such a wide loop gain range.

Edmond, since you are the resident CFA king, maybe you can comment on the approach or offer some insights.

BTW, if you insert a cascode in the collectors of Q13 and Q14, you can get the 20k THD down to about 250 parts per billion (with R18 set to 10k - so about 40db OLG at 20kHz).
Attached Files
File Type: zip cfa_1.zip (1.7 KB, 84 views)
__________________
bonsai
Amplifier Design and Construction for MUSIC! http://hifisonix.com/
  Reply With Quote
Old 9th July 2008, 10:26 AM   #2
diyAudio Member
 
Edmond Stuart's Avatar
 
Join Date: Nov 2003
Location: Amsterdam
Hi Bonsai,

Me, the CFA king? That's too much honor.

Anyhow, of course you can do it that way, but some people might argue that your circuit has a different topology than most amplifiers have, so listening tests are not representative.
For example, your circuit is symmetrical, therefore (almost) no even harmonics will show up, while asymmetrical designs will produce more even harmonics. In other words, the distribution of the harmonics will be quite different.

Perhaps, you may also build a second circuit, that is, a asymmetrical one, or just start with an existing high NFB amplifier and load the VAS output with a pot.

Probably you know already what I think about the low vs high NFB controverse, if not, look here:
http://www.diyaudio.com/forums/showt...88#post1540288

As for the additional cascodes, your right, they will lower the distortion even further. BTW, this is exactly what I have done with my latest design.

Cheers,
Edmond.
__________________
Een volk dat voor tirannen zwicht, zal meer dan lijf en
goed verliezen dan dooft het licht…(H.M. van Randwijk)
  Reply With Quote
Old 9th July 2008, 11:28 AM   #3
Bonsai is offline Bonsai  Taiwan
diyAudio Member
 
Bonsai's Avatar
 
Join Date: Jul 2003
Location: Taipei, Taiwan
Edmond, thanks for the feedback.

I am specifcally trying to remove the 'distortion' part of the sound - this is why I mae the statement about 20kTHD distortion with heavily loaded VAS being 0.015% and therefore not a consideration.

I guess, I would like to see if, in a circuit with inherently very low open loop distortion, what, if any, changes to the sound will be by applying feedback.

I know your views on ZGNF designs, and personally I have no beef with feedback. I use it myself.

However, if we can conduct an experiment where we can vary GNF, it would help illuminate the discussion and perhaps settle a few arguments don't you think?
__________________
bonsai
Amplifier Design and Construction for MUSIC! http://hifisonix.com/
  Reply With Quote
Old 9th July 2008, 02:31 PM   #4
diyAudio Member
 
Edmond Stuart's Avatar
 
Join Date: Nov 2003
Location: Amsterdam
Hi Bonsai,

I will appreciate all your effort to shed more light on this topic. But frankly speaking, I'm lacking the enthusiasm (and time!) to discuss the matter with the 'other camp', because I'm afraid that they will be completely deaf to arguments from the high NFB camp.

Besides, I'm hopeless behind schedule with other things like the PCP amp and my website, sorry.

Nevertheless, please go on with your experiments and I will follow them with great interest.

Cheers,
Edmond.
__________________
Een volk dat voor tirannen zwicht, zal meer dan lijf en
goed verliezen dan dooft het licht…(H.M. van Randwijk)
  Reply With Quote
Old 9th July 2008, 02:45 PM   #5
jcx is offline jcx  United States
diyAudio Member
 
Join Date: Feb 2003
Location: ..
Loading the VAS is a little more complicated than it looks at 1st sight

it's "obvious" that you reduce the gain by heaver loading - it is less obvious that you also reduce the output impedance - which then causes lower "interface distortion" with the next stage which presents a nonlinear dynamic load to the VAS

for the common situation where the input current nonlinearity of the output/driver stage is the larger distortion component the reduced drive impedance can be nearly as effective as the unloaded VAS higher (Voltage) loop gain in reducing this distortion

at a minimum a "fair" comparison would require a gain trimmable block with fixed input and output impedances regardless of gain setting

and then you have to decide what is "fair" when it comes to loop gain compnsation as you vary gain


I don't see the critical ".options plotwinsize=0" directive in your schematic - this turns off Ltspice' default lossy data compresson that makes distortion measurements futile - I always include the directive on each schematic .asc rather than try to remember to change control panel settings

I find it simpler to attach Ltspice .asc files by just copying them, renaming the copy with a extra ".txt" extension, maybe
newer defaults in Vista make this too difficult - I'm avoiding Vista machines at work as long as I can
  Reply With Quote
Old 9th July 2008, 03:38 PM   #6
diyAudio Member
 
Edmond Stuart's Avatar
 
Join Date: Nov 2003
Location: Amsterdam
Hi JCX,

Although I generally agree with you, one remark about the VAS output impedance vs the input impedance of the OPS and the effect of loading the VAS:
Looking at the blameless amp, for example, VAS-Zo= 500 Ohm and OPS-Zi = 200kOhm at 1kHz. So I don't think that, at least in this case, the "interface distortion" will be much lower if the VAS is loaded by an external resistor of say 5kOhm. The DC open loop gain however, is decreased by ~20dB.

@Bonsai,

What do we do with unity gain frequency of the NFB loop? Keeping it constant or varying it with the same amount as the DC loop gain?


Cheers,
Edmond.
__________________
Een volk dat voor tirannen zwicht, zal meer dan lijf en
goed verliezen dan dooft het licht…(H.M. van Randwijk)
  Reply With Quote
Old 9th July 2008, 04:10 PM   #7
jcx is offline jcx  United States
diyAudio Member
 
Join Date: Feb 2003
Location: ..
granted there are lots of possible scenerios,

I was thinking more about C loading of the VAS to establish the dominant pole vs Miller C compensation, another example exagerating these effects is directly driving Mosfet output Q without intervening buffers

also I'd consider 20 KHz as the "design" frequency since I believe available negative feedaback at 1 KHz can be so high that resloving any effects there is a measurement challenge
  Reply With Quote
Old 9th July 2008, 05:09 PM   #8
KSTR is offline KSTR  Germany
diyAudio Member
 
KSTR's Avatar
 
Join Date: Jul 2007
Location: Central Berlin, Germany
I've found that this circuit needs a Miller cap of 100pF or more to have enough phase margin at low gains. And since it's all pretty much class A there is not that much distortion to be reduced by feedback anyway. At higher signal amplitudes and/or heavier load the picture changes when the outputs are starved of current or even enter class B.

- Klaus
  Reply With Quote
Old 9th July 2008, 05:32 PM   #9
diyAudio Member
 
traderbam's Avatar
 
Join Date: Jan 2002
Location: Earth
Bonsai,
If your objective is to vary the magnitude of the feedback, whilst keeping the loop GBP the same, why not make a servo and vary the servo gain? So make a fixed gain amp. Add a servo. Vary the servo gain.
That way you don't mess with the intrinsic distortion in the amp.
Brian
  Reply With Quote
Old 9th July 2008, 05:44 PM   #10
diyAudio Member
 
traderbam's Avatar
 
Join Date: Jan 2002
Location: Earth
I don'tmean you need a fixed gain amp, I mean you don't need to vary the forward gain of the amp. You can vary the gain of the feedback instead. This way you are not altering the amp's characteristics when you change the feedback gain.
  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
hi-loop gain amp mfc Solid State 49 10th January 2009 06:10 AM
open loop gain,how many is enough? Leolabs Solid State 28 17th May 2006 09:00 PM
Loop gain question lumanauw Solid State 55 16th December 2005 09:59 PM
Gain loop formula fscarpa58 Everything Else 2 16th January 2004 11:19 AM
Variable Gain preamp? Nisbeth Solid State 8 5th June 2002 11:46 AM


New To Site? Need Help?

All times are GMT. The time now is 01:13 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