Bob Cordell's Power amplifier book - Page 349 - 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 11th June 2013, 11:22 PM   #3481
diyAudio Member
 
keantoken's Avatar
 
Join Date: Aug 2006
Location: Texas
Blog Entries: 2
Quote:
Originally Posted by michaelkiwanuka View Post
The major feedback loop has no impact, at least to a first order, on the output impedance of the TIS. You can confirm this by disabling the major loop by means of a large capacitor shunting the inverting input of the amplifier to ground or using a large inductor in series with the inverting input of the amplifier.
I did, and again got a different result.
Attached Images
File Type: png Screenshot from 2013-06-11 18:21:50.png (118.5 KB, 99 views)
  Reply With Quote
Old 12th June 2013, 01:17 AM   #3482
diyAudio Member
 
Join Date: Sep 2006
Quote:
Originally Posted by michaelkiwanuka View Post
The major feedback loop has no impact, at least to a first order, on the output impedance of the TIS. You can confirm this by disabling the major loop by means of a large capacitor shunting the inverting input of the amplifier to ground or using a large inductor in series with the inverting input of the amplifier.

Further, Bob, I have also demonstrated that the minor loop has no effect on the loop gain of the ANF current source. Why do you find this so difficult to appreciate?

Tomorrow I'll do as you've asked; viz. run a sim of the loop gain of the current source with a resistive load.
Thanks, Mike. I'll look forward to your result.

Did you disable both loops?

Cheers,
Bob
  Reply With Quote
Old 12th June 2013, 02:52 AM   #3483
diyAudio Member
 
Join Date: Nov 2010
Quote:
Originally Posted by Bob Cordell View Post

Can you refer me to where Stochino and Self discussed this (I did not see any reference to this in Self's 5th edition).
Self 5th Edition pp 94 - 95.

Quote:
I'm very interested in learning more about any disadvantages in the replica cascode driving technique as compared to the other technique.
Hope this helps until Manso's time zone is awake.

Best wishes
David
  Reply With Quote
Old 12th June 2013, 10:03 AM   #3484
Banned
 
Join Date: Oct 2010
Location: Chigwell, Essex
Quote:
Originally Posted by HarryDymond View Post
Why even simulate the current source with all the other amplifier guff if you're going to isolate it? Why not just simulate it by itself? And in order to show off the "intrinsic" behaviour, surely it would be best to present it with the "kindest" load, which for a current source is a short circuit from DC to daylight (the dual of the "kindest" load for a voltage source being an open circuit from DC to daylight).
Harry, you're a genius! Of course in order to obviate the load on a voltage source its output must be open circuit. Conversely, in order to eliminate the load on a nominal current source its output must be short circuit. Why I didn't remember this most elementary facet of electrical circuit theory I haven't the faintest idea!

Now, using a large inductor in series with the current source output does isolate it from the rest of the circuit, but puts maximum loading on the current source because its output is then a virtual open circuit at the frequencies of interest. This is the exact opposite of what is required to establish the intrinsic loop transmission of the current source, and explains the relatively low loop gain I obtained with the inductor in situ.

Therefore, what is required in principal is a large capacitor shunting the output of the current source to ground. I have tried this, and, although loop transmission increases from roughly 12dB with the series inductor to 59dB across the audio band with the shunt capacitor, there is virtually no difference in loop gain with and without the shunt capacitor. I am yet to identify a reason for this behaviour. Any ideas anyone?

Nevertheless, Harry, you're right in suggesting that extraneous circuitry need not be included to establish the intrinsic loop gain of the ANF current source, unless it is desired to quantify the effect of such circuitry on the loop gain, which, as I have indicated above, is negligible at least in the case of the amplifier topology at issue.

The figure below shows the loop gain of the amplified negative feedback current source in isolation with its output short circuited to ground. The asc file and transistor models are also included.

Bob, I think it is now apparent why loading the current source resistively is not a good thing, so I shall not be running that simulation I promised you after all.
Attached Images
File Type: png 01.png (11.0 KB, 79 views)
File Type: png 02.png (48.3 KB, 79 views)
Attached Files
File Type: asc ANF_loop_gain3.asc (1.5 KB, 1 views)
File Type: txt BJT_models.txt (7.4 KB, 1 views)
  Reply With Quote
Old 12th June 2013, 10:17 AM   #3485
diyAudio Member
 
Join Date: Jan 2013
Hi Bob

Dave Zan beat me to it.

It will be interesting if you could confirm the same result as Self with the blameless as it might show exactly why your design overcomes this or that maybe your design will also perform better if those cascodes are biased from the LTP.
  Reply With Quote
Old 12th June 2013, 10:18 AM   #3486
Banned
 
Join Date: Oct 2010
Location: Chigwell, Essex
You'll recall that I indicated somewhere in this thread that placing a resistor in series with the control transistor in the amplified negative feedback (ANF) current source reduces loop gain. This remains true as shown below.

Of interest, however, inclusion of the 2K2 resistor also reduces the unity loop gain frequency drastically; in this case from 50Mhz (phase margin~44 degrees) to just under 2Mhz (phase margin~83 degrees).
Attached Images
File Type: png 01.png (20.2 KB, 78 views)
File Type: png 02.png (52.2 KB, 75 views)
  Reply With Quote
Old 12th June 2013, 10:33 AM   #3487
diyAudio Member
 
Join Date: Jan 2013
Quote:
Originally Posted by michaelkiwanuka View Post
You'll recall that I indicated somewhere in this thread that placing a resistor in series with the control transistor in the amplified negative feedback (ANF) current source reduces loop gain. This remains true as shown below.

Of interest, however, inclusion of the 2K2 resistor also reduces the unity loop gain frequency drastically; in this case from 50Mhz (phase margin~44 degrees) to just under 2Mhz (phase margin~83 degrees).
Isnt that a good thing ?? Although I wouldnt make such a drastic reduction.
  Reply With Quote
Old 12th June 2013, 12:05 PM   #3488
Banned
 
Join Date: Oct 2010
Location: Chigwell, Essex
Quote:
Originally Posted by manso View Post
Isnt that a good thing ?? Although I wouldnt make such a drastic reduction.
As far as I am aware nobody has reported instability local to the ANF current source in the absence of the resistor in series with the base of the control transistor in amplifier applications.

Therefore since the 2K2 resistor reduces loop gain at 20KHz by 20dB its inclusion may not be a good thing.

Presumably this reduction in loop gain is due to the low pass filter effect that occurs with the resistor in conjuction with the input capacitance of the control transistor.

Last edited by michaelkiwanuka; 12th June 2013 at 12:08 PM.
  Reply With Quote
Old 12th June 2013, 02:30 PM   #3489
diyAudio Member
 
Join Date: Sep 2006
Quote:
Originally Posted by manso View Post
Hi Bob

Dave Zan beat me to it.

It will be interesting if you could confirm the same result as Self with the blameless as it might show exactly why your design overcomes this or that maybe your design will also perform better if those cascodes are biased from the LTP.
Thanks for pointing me to this, manso and Dave.

Self discusses the driven cascode and concludes that direct drive at AC from the tail works well, and he was unable to get satisfactory results from replica drive derived from the output signal. His setup was using a very high common mode voltage of 5V at an output level of 10V and a gain of 2 in a test amplifier without an output stage, but Miller compensated. He observed that CM distortion increased at high frequencies. Distortion was 2nd harmonic at 0.002% at 15kHz. He observed that the CM distortion percentage (2nd harmonic) was proportional to the square of the CM voltage (seems odd) and speculated that in an amplifier with a closed loop gain of 23 and operating at the same output level the CM distortion would be less than 0.00002%.

He observed that the distortion was worse at low Vce and speculated that the effect was either Early effect or Ccb nonlinearity, both of which can be expected to get worse at low Vce. He said that the CM distortion effect cannot be found in SPICE simulation and opined that this may be due to the linear modeling in SPICE of the Early effect. This would not account for the effect if the CM distortion was mainly from Ccb. He then noted that changing the input devices did not change the effect, so then concluded the effect is not from Early effect. But then he subsequently concluded that it was Early effect. These conclusions are a bit confusing to me when taken together.

He then bootstrapped the cascode from the LTP tail with AC coupling, placing the cascode DC bias at about half the positive rail voltage. In an arrangement with gain of 1.3 and output level of 5V, the cascode reduced CM distortion from 0.003% to 0.0015% at 20kHz and concluded that the cascode completely eliminated the HF distortion effect.

Curiously, at lower frequencies below 5kHz in this arrangement the cascode version produced 20% MORE CM distortion than the non-cascode version.

When instead he used a replica of the feedback signal whose amplitude was controlled by a variable attenuator he got only a small reduction in CM distortion, and at a drive level of only 131mV, way off from what should have been the results if the replica feed was operating as hoped. Details of the replica feed circuit or distortion plot were not shown, so it is unclear why the replica feed did not work. It was speculated that the replica feed did not work because the output signal is phase shifted compared with the input signal. However, the closed-loop phase shift of the amplifier at 20kHz should have been quite small, perhaps only 5 degrees.

Bottom line is that I should do some SPICE simulations with both types of cascode drive and see how each performs when I get some time. I should also verify whether or not SPICE is blind to this effect, as Self asserts.

Finally, in my MOSFET amplifier where replica drive of the cascode appeared to work, the differences that might account for the good performance of the replica drive could be due to the lower (normal) CM voltage level, the use of MIC compensation, or the use of JFETs in my LTP. I should note that I did not try driving the cascode bases from the LTP tail, so I cannot say that that would not have been even better.

Cheers,
Bob
  Reply With Quote
Old 12th June 2013, 02:44 PM   #3490
diyAudio Member
 
Join Date: Jan 2013
I didnt quite understand the reason for the analysis in the first place is it to obtain the highest output impedance ?? This is important for CMRR and PSRR. If one wants very good performance the anf current source is the way to go; to include low noise tempco and better Zout the better option is the one shown by Cordell with the zener, the de facto CS I like too use. For better youll have to use cascoded and Self shows the improved results obtained by its use. The choice of transistors is also important if one wants high performance.

One can obtain a near 7 times better Zout with the inclusion of another active part, opamp companies seem to think its worth the complexity and use it to obtain the very high CMRR and PSRR figures they do.
  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
Another realization of Bob Cordell's THD Analyzer giulianodes Equipment & Tools 37 9th October 2013 02:48 AM
best audio amplifier book?? Bouvett Everything Else 30 13th August 2012 02:43 AM
Amplifier Design Book pixie Everything Else 27 11th June 2010 08:36 PM
Project 11.1 from Slone "High-Power Amplifier" Book Karl71 Solid State 46 6th October 2008 03:47 AM
book-audio power amplifier by Douglas Self mikee12345 Solid State 8 16th November 2003 01:16 AM


New To Site? Need Help?

All times are GMT. The time now is 03:55 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