Bob Cordell Interview: Error Correction - Page 297 - 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 2nd December 2007, 02:04 PM   #2961
syn08 is offline syn08  Canada
Account disabled at member's request
 
Join Date: Aug 2005
Location: Toronto
Default Re: Re: Re: hec != hoax ?

Quote:
Originally posted by Edmond Stuart


I really hate bias circuits that cripple the VAS

I was thinking about those 5 little VDC symbols sitting like thorns in my eyes :-))
  Reply With Quote
Old 2nd December 2007, 02:13 PM   #2962
diyAudio Member
 
Edmond Stuart's Avatar
 
Join Date: Nov 2003
Location: Amsterdam
Quote:
Originally posted by lumanauw
Hi, Edmond,

Discussing about VAS drive, what is your opinion about this kind of VAS drive? This is from SKA topology. It boost the gain of the VAS drive, (making it high impedance) by using bootstrap (R2+R3+C2). Then feeded to emitor follower Q5. From here, it drives the VAS stage (Q7-Q8, but here Q7-Q8 are already power devices).
Hi lumanauw,

What else could I say about this circuit, as most ins and outs you have already figured out by yourself.
To be honest, I don't like this circuit at all, because the bias of the output stage is not well controlled, for example.

Cheers, Edmond.
Attached Images
File Type: gif ska.gif (7.3 KB, 698 views)
  Reply With Quote
Old 2nd December 2007, 03:25 PM   #2963
andy_c is offline andy_c  United States
Banned
 
Join Date: Apr 2003
In Greg's actual design, the input stage current sources are a function of heat sink temperature to compensate the output stage bias. I had some email conversations with him about this a couple of years ago. He was trying to get me to tweak the level 3 SPICE models of his output devices so the simulated and measured tempco agreed. It looked too time consuming, so I gave up on it.
  Reply With Quote
Old 4th December 2007, 07:29 AM   #2964
KSTR is offline KSTR  Germany
diyAudio Member
 
KSTR's Avatar
 
Join Date: Jul 2007
Location: Central Berlin, Germany
Hi,

@Edmond:
Thanks for elaborating on the VAS details... I get a clearer picture now (still it's not intuitive to me... I need to sim it to really get it, it seems). Also, that 40ppb THD20 @300mA is really stunning, though I assume this figure is for the complete amp, not the OS in isolation.

@all, FWIW: Here I played a little with that autobias class A OS, and as it deals a little with error correction / feedback in two regards (#1: error cross-feed, #2: separating voltage and current terms in the feedback loops) you might want to take a look there (comments welcome). As of now, I abandoned the output stage concept I posted here a while ago, since this autobias thing does the same performance with lower parts count and simpler topology, all cheap discrete stuff.

- Klaus
  Reply With Quote
Old 4th December 2007, 04:44 PM   #2965
diyAudio Member
 
Edmond Stuart's Avatar
 
Join Date: Nov 2003
Location: Amsterdam
Default Re: Re: Re: Re: hec != hoax ?

Quote:
Originally posted by syn08

I was thinking about those 5 little VDC symbols sitting like thorns in my eyes :-))
Hi Ovidiu,

Thorns are gone.

Cheers, Edmond.

NB: This is not a complete amp: no protection, no Zobel network, no servo etc.
Attached Images
File Type: png amp7-pub.png (56.1 KB, 822 views)
  Reply With Quote
Old 4th December 2007, 07:56 PM   #2966
syn08 is offline syn08  Canada
Account disabled at member's request
 
Join Date: Aug 2005
Location: Toronto
Default Re: Re: Re: Re: Re: hec != hoax ?

Quote:
Originally posted by Edmond Stuart


Thorns are gone.

NB: This is not a complete amp: no protection, no Zobel network, no servo etc.
They are :-)

Now, how would you compare this design complexity with respect to the PGP amp?

But after considering each amp simulated performance?
  Reply With Quote
Old 5th December 2007, 11:36 AM   #2967
diyAudio Member
 
Edmond Stuart's Avatar
 
Join Date: Nov 2003
Location: Amsterdam
Default Re: Re: Re: Re: Re: Re: hec != hoax ?

Quote:
Originally posted by syn08
They are :-)

Now, how would you compare this design complexity with respect to the PGP amp?

But after considering each amp simulated performance?
Hi Ovidiu,

Of course the front end is less complex, as it lacks a NDFL stage. Nevertheless, due to TMC, the distortion is about at the same level.
OTOH, the offset voltage is quite large (0.5V), so a servo needed, which adds to complexity (an op-amp, +/- 15 PSU, etc.).

Notice that this is not just a TMC amp, rather a double TMC amp (DTMC). Not only the take off point is 'TMCed' (see: R26-27 and C7-8), but also the feedback input for the Miller compensation is 'TMCed' (see: R17 and C4-5). As a result, at AF the Miller cap (C6) is effectively connected from the output right to the inverting input, while at HF the Miller cap is effectively connected between VAS input and VAS output. This arrangement ensures an extreme low distortion of the front end and a very modest drive requirement of the input stage (biased at a mere 1mA).

BTW, DTMC is only compatible with a CFB input stage, thus no LTP, let alone configurations with an additional gain stage between the IPS and VAS.

Cheers, Edmond.
  Reply With Quote
Old 5th December 2007, 04:39 PM   #2968
diyAudio Member
 
Join Date: Sep 2006
Default Re: Re: Re: Re: Re: Re: Re: hec != hoax ?

Quote:
Originally posted by Edmond Stuart


Hi Ovidiu,

Of course the front end is less complex, as it lacks a NDFL stage. Nevertheless, due to TMC, the distortion is about at the same level.
OTOH, the offset voltage is quite large (0.5V), so a servo needed, which adds to complexity (an op-amp, +/- 15 PSU, etc.).

Notice that this is not just a TMC amp, rather a double TMC amp (DTMC). Not only the take off point is 'TMCed' (see: R26-27 and C7-8), but also the feedback input for the Miller compensation is 'TMCed' (see: R17 and C4-5). As a result, at AF the Miller cap (C6) is effectively connected from the output right to the inverting input, while at HF the Miller cap is effectively connected between VAS input and VAS output. This arrangement ensures an extreme low distortion of the front end and a very modest drive requirement of the input stage (biased at a mere 1mA).

BTW, DTMC is only compatible with a CFB input stage, thus no LTP, let alone configurations with an additional gain stage between the IPS and VAS.

Cheers, Edmond.

Hi Edmond,

Although I don't call it "DTMC", the approach of taking the TMC path from the output of the amplifier back to the input is the way I have applied TMC to my input/VAS architecture. That is a fairly obvious extension of my favorite type of feedback compensation to TMC. Indeed, TMC inclusive of the input stage is quite compatible with an LTP input stage as shown in the schematic below.

That simulation yielded 0.001% THD-20 out to 200 kHz at 100 watts into 8 ohms with a simple and straightforward version of my amplifier that did not use EC.

The TMC path includes C5, C2 and R36.

Note also that the input cascode is bootstrapped with a replica of the input signal derived from the output signal via R3 and R27. I have found that doing so reduces a significant contributor to distortion that was in my original MOSFET amplifier design.

No Early effect mitigation is used in the VAS of this simple design.

Cheers,
Bob
Attached Images
File Type: png tmc amp.png (97.8 KB, 585 views)
  Reply With Quote
Old 5th December 2007, 04:57 PM   #2969
syn08 is offline syn08  Canada
Account disabled at member's request
 
Join Date: Aug 2005
Location: Toronto
Default Re: Re: Re: Re: Re: Re: Re: hec != hoax ?

Quote:
Originally posted by Edmond Stuart


Hi Ovidiu,

Of course the front end is less complex, as it lacks a NDFL stage. Nevertheless, due to TMC, the distortion is about at the same level.
OTOH, the offset voltage is quite large (0.5V), so a servo needed, which adds to complexity (an op-amp, +/- 15 PSU, etc.).

Notice that this is not just a TMC amp, rather a double TMC amp (DTMC). Not only the take off point is 'TMCed' (see: R26-27 and C7-8), but also the feedback input for the Miller compensation is 'TMCed' (see: R17 and C4-5). As a result, at AF the Miller cap (C6) is effectively connected from the output right to the inverting input, while at HF the Miller cap is effectively connected between VAS input and VAS output. This arrangement ensures an extreme low distortion of the front end and a very modest drive requirement of the input stage (biased at a mere 1mA).

BTW, DTMC is only compatible with a CFB input stage, thus no LTP, let alone configurations with an additional gain stage between the IPS and VAS.

Cheers, Edmond.
It would be interesting to describe the pros and cons for each approach (NDFL vs. TMC). If the performance (not only THD20, but also PSRR, SR, IMD, DIM, etc...) is the same, then why/how would one chose between NDFL and TMC?

You know I am a little skeptic regarding TMC (in particular about delivering under 1ppm THD20 with a reasonable phase margin). I have to build and measure such an amp to find if it really delivers to te expectations and better understand what's the price of this approach and how it would compare to NDFL.

As a rule, I think it would be always a good idea to show both sides of a certain new design approach. As you know, there is no free lunch... Maybe defining a consistent set of parameters against each design should be evaluated, both from a simulation and measurements perspective, would help?

BTW, another issue that must be addressed to make such designs more appealing is to make the simulation results reproductible. Unfortunately, it seems that PSpice (at least the Cadence/OrCad version that I am using) is unable to always reproduce your numbers, even for simpler designs and even if the device models are identical. I think it would be a good idea to define a set of simulation benchmarks, to provide consistency across platforms. I think this is important in particular for "sensitive" topics like THD, IMD, etc... and I don't think it's realistic to expect that everybody should use and trust Micro-Cap.

I know, unfortunately there's nothing sexy in such work, but I don't think there's other way to raise the level of confidence in a certain design. Of course, I am assuming you find confidence and results consistency important.
  Reply With Quote
Old 5th December 2007, 07:13 PM   #2970
diyAudio Member
 
Edmond Stuart's Avatar
 
Join Date: Nov 2003
Location: Amsterdam
Default Re: Re: Re: Re: Re: Re: Re: Re: hec != hoax ?

Quote:
Originally posted by Bob Cordell
Hi Edmond,

Although I don't call it "DTMC", the approach of taking the TMC path from the output of the amplifier back to the input is the way I have applied TMC to my input/VAS architecture. That is a fairly obvious extension of my favorite type of feedback compensation to TMC. Indeed, TMC inclusive of the input stage is quite compatible with an LTP input stage as shown in the schematic below.

That simulation yielded 0.001% THD-20 out to 200 kHz at 100 watts into 8 ohms with a simple and straightforward version of my amplifier that did not use EC.

The TMC path includes C5, C2 and R36.

Note also that the input cascode is bootstrapped with a replica of the input signal derived from the output signal via R3 and R27. I have found that doing so reduces a significant contributor to distortion that was in my original MOSFET amplifier design.

No Early effect mitigation is used in the VAS of this simple design.

Cheers,
Bob
Hi Bob,

Perhaps you should have a closer look at my latest schematic. I insist that it is DTMC, not just TMC, because:
"also the feedback input for the Miller compensation is 'TMCed' (see: R17 and C4-5). As a result, at AF the Miller cap (C6) is effectively connected from the output right to the inverting input, while at HF the Miller cap is effectively connected between VAS input and VAS output."

Before making any comment on your schematic, please increase the font size, as the component values are hard to identify.


Cheers, Edmond.
  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 11:09 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