dimitry's paper in the September 1996 JAES would appear to disagree: http://www.aes.org/journal/toc/sept96.html
It also seems to put Hawksford's EC into a category of no better than NFB:
At least, that's what I take the summary to imply, as I'm too cheap to pay the $5 for the article. I've never come across a journal that charges even members for access to an article -.- ... if I had realized they do that, I'd have never subscribed to the AES.
Ordinary Miller compensation is sub-optimal
While ordinary Miller compensation is quite good for stability, I have found it to be less than the best for matters of high-frequency distortion. In fact, many amplifiers that suffer from soft or hard slewing distortion (or TIM) do so because of the way the simple Miller compensation works.
For those interested, my 1984 JAES paper on a MOSFET power amplifier with error correction (based on the Hawksford idea) is now posted on my fairly-new web site at www.cordellaudio.com.
That amplifier does not use simple Miller compensation. Instead, its main compensation is provided by what is effectively Miller compensation by feedback from the VAS output node back to the inverting input node of the input differential pair. The loop so formed does itself also need compensation, albeit at a much higher unity gain frequency, and that is provided by a simple series R-C shunt across the collectors of the cascoded input differential pair. This kind of compensation results in extraordinarily high slew rate capability and HF linearity without the usual compromises of classic simple Miller compensation.
I have also posted some other papers at www.cordellaudio.com that may be of interest on some of the subjects covered in this thread. Unfortuantely, a couple of those papers were scanned in with poor quality, for which I apologize. I'll try to fix that this weekend.
One of those papers is the one on Phase Intermodulation Distortion (PIM), what some have referred to as "fm distortion". That paper describes the origins of PIM, what it has to do with negative feedback, instrumentation for measuring it, and actual measurements on amplifiers with and without negative feedback.
Note that although it employs substantial amounts of negative feedback, the MOSFET power amplifier with error correction was measured to have less than 100 picoseconds of PIM.
Bob Cordell
While ordinary Miller compensation is quite good for stability, I have found it to be less than the best for matters of high-frequency distortion. In fact, many amplifiers that suffer from soft or hard slewing distortion (or TIM) do so because of the way the simple Miller compensation works.
For those interested, my 1984 JAES paper on a MOSFET power amplifier with error correction (based on the Hawksford idea) is now posted on my fairly-new web site at www.cordellaudio.com.
That amplifier does not use simple Miller compensation. Instead, its main compensation is provided by what is effectively Miller compensation by feedback from the VAS output node back to the inverting input node of the input differential pair. The loop so formed does itself also need compensation, albeit at a much higher unity gain frequency, and that is provided by a simple series R-C shunt across the collectors of the cascoded input differential pair. This kind of compensation results in extraordinarily high slew rate capability and HF linearity without the usual compromises of classic simple Miller compensation.
I have also posted some other papers at www.cordellaudio.com that may be of interest on some of the subjects covered in this thread. Unfortuantely, a couple of those papers were scanned in with poor quality, for which I apologize. I'll try to fix that this weekend.
One of those papers is the one on Phase Intermodulation Distortion (PIM), what some have referred to as "fm distortion". That paper describes the origins of PIM, what it has to do with negative feedback, instrumentation for measuring it, and actual measurements on amplifiers with and without negative feedback.
Note that although it employs substantial amounts of negative feedback, the MOSFET power amplifier with error correction was measured to have less than 100 picoseconds of PIM.
Bob Cordell
Miller compensation and Open Loop Bandwidth
While I am not a big fan of Miller compensation, there is no reason why Miller compensation cannot be used while achieving high open-loop bandwidth.
Indeed, high open loop bandwidth is usually achieved by getting rid of the extra loop gain that would normally be there at frequencies below 20 kHz. Such a reduction in low-frequency open loop gain is usually accomplished with either a resistive load on the VAS output, or a local feedback loop, or some combination of the two.
Bob Cordell
While I am not a big fan of Miller compensation, there is no reason why Miller compensation cannot be used while achieving high open-loop bandwidth.
Indeed, high open loop bandwidth is usually achieved by getting rid of the extra loop gain that would normally be there at frequencies below 20 kHz. Such a reduction in low-frequency open loop gain is usually accomplished with either a resistive load on the VAS output, or a local feedback loop, or some combination of the two.
Bob Cordell
It is no relation between OL bandwidth and Miller compensation. Lowering gain at certain frequency by cap/resistance to ground is just wasting the available OLG. Miller compensation exchanges reduction of OLG for low distortion inside Miller loop. Miller compensation allows to make the second pole time constant (associated with VAS load) smaller.
Re: Ordinary Miller compensation is sub-optimal
I think you'll find that, strictly speaking, your amp. uses series compensation of the lead variety.
The only difference between your approach and the usual phase lead approach (RC network directly shunting feedback resistor) resides in the fact that with the former the RC net (R13/C4) is connected to the second stage's output; output stage singularities are thus bypassed, and a trivial increase in gain margin accrues.
It is because of the later problem, i would suggest, that additional shunt compensation is required with respect to the input stage to augment major-loop stability margins.
I agree: single-pole Miller (or feedback) compensation is not ideal.
However, you could use double-pole Miller compensation to obtain stable loop gains at HF of the same order, at least, as those obtained with series compensation, with the additional advantage that the second stage is made more linear by the minor feedback loop at HF.
Bob Cordell said:
I think you'll find that, strictly speaking, your amp. uses series compensation of the lead variety.
The only difference between your approach and the usual phase lead approach (RC network directly shunting feedback resistor) resides in the fact that with the former the RC net (R13/C4) is connected to the second stage's output; output stage singularities are thus bypassed, and a trivial increase in gain margin accrues.
It is because of the later problem, i would suggest, that additional shunt compensation is required with respect to the input stage to augment major-loop stability margins.
PMA said:
I agree: single-pole Miller (or feedback) compensation is not ideal.
However, you could use double-pole Miller compensation to obtain stable loop gains at HF of the same order, at least, as those obtained with series compensation, with the additional advantage that the second stage is made more linear by the minor feedback loop at HF.
Re: Miller compensation and Open Loop Bandwidth
This, alas, is the electronic equivalent of robbing Peter to pay Paul.
The increase in forward-path bandwidth is at the very substantial expense of forward-path gain and second-stage linearity, with no compensating benefits.
Bob Cordell said:
This, alas, is the electronic equivalent of robbing Peter to pay Paul.
The increase in forward-path bandwidth is at the very substantial expense of forward-path gain and second-stage linearity, with no compensating benefits.
Yes, of course you are correct. I am certainly not an advocate of throwing away open loop gain to raise open loop bandwidth. I was just pointing out that that misguided approach to design is not inconsistent with Miller compensation. I don't think I have ever seen a case where doing that reduced ANY kind of measurable distortion, including TIM or "f.m." distortions.
Bob
Bob
Upupa Epops said:Gentlemen, is here anybody who have some experiences with Hawksford's circuit ( except Nelson, Charles and me ) ? Theory is theory, but tree of life is green.
I have. I have a power stage that uses Hawksford EC, driven by a front end voltage amp with H.EC. Works like a charm, no stability problems, no compensation, reasonably THD free. OL gain and CL gain are identical, only better linearity with H.EC.
Jan Didden
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.