Re: VAS buffer
OK, with 2 resistors you loose some headroom, but tying them to two different points, between which some (DC) voltage exists, you gain some headroom (+DRV and -DRV, for example, see post 2826 and 2827).
Maybe because not everybody gives away his secrets, or is it because the majority of designers is so terribly conservative?
Cheers,
Edmond.
Terry Demol said:
OK, with 2 resistors you loose some headroom, but tying them to two different points, between which some (DC) voltage exists, you gain some headroom (+DRV and -DRV, for example, see post 2826 and 2827).
Maybe because not everybody gives away his secrets, or is it because the majority of designers is so terribly conservative?
Cheers,
Edmond.
Re: Re: Re: Re: Re: VAS buffer
I always thought that it was #2, as implemented by Tom Holman, for example.
However, to get rid of the capacitance effect of the Baker clamp diode, I like the idea of what I call a flying Baker clamp. This is similar in concept to the "flying catch diodes" I used in the MOSFET power amplifiers I used in my Athena active loudspeakers. I use the term "flying" when both sides of the diode nominally have the same signal swing, so diode capacitance has little or no effect. One then can apply a Baker clamp to the signal source for the end of the diode not connected to the output of the VAS. That stops it from flying when the time for clipping occurs.
Note that the concept of flying clipping diodes can also be used in implementing soft clipping circuits ahead of the NFB amplifier.
Cheers,
Bob
Edmond Stuart said:
I always thought that it was #2, as implemented by Tom Holman, for example.
However, to get rid of the capacitance effect of the Baker clamp diode, I like the idea of what I call a flying Baker clamp. This is similar in concept to the "flying catch diodes" I used in the MOSFET power amplifiers I used in my Athena active loudspeakers. I use the term "flying" when both sides of the diode nominally have the same signal swing, so diode capacitance has little or no effect. One then can apply a Baker clamp to the signal source for the end of the diode not connected to the output of the VAS. That stops it from flying when the time for clipping occurs.
Note that the concept of flying clipping diodes can also be used in implementing soft clipping circuits ahead of the NFB amplifier.
Cheers,
Bob
Baker clamp
Hi Bob,
After reading:
http://www.ieeta.pt/~alex/docs/ApplicationNotes/Rectifier Applications Handbook.pdf
and
http://www.zetex.com/3.0/appnotes/apps/an22.pdf
do you still think it's #2?
Was it Tom Holman, perhaps, who caused the confusion?
Cheers,
Edmond.
Bob Cordell said:
Hi Bob,
After reading:
http://www.ieeta.pt/~alex/docs/ApplicationNotes/Rectifier Applications Handbook.pdf
and
http://www.zetex.com/3.0/appnotes/apps/an22.pdf
do you still think it's #2?
Was it Tom Holman, perhaps, who caused the confusion?
Cheers,
Edmond.
Re: Baker clamp
I don't know what you guys are talking about. In my electronic texts and back in switching powersupplies 101, a Baker clamp is/was a diode connected in parallel with the base-collector junction of a bipolar transistor - typically used in switching applications.
For a Darlington transistor, this diode can be a silicon device. For a non Darlington transistor, it must be a schottkey - so long as its forward voltage drop is less than Vbe-VceSat.
It is there to prevent saturation of the transistor, allowing it to switch faster.
Edmond Stuart said:
I don't know what you guys are talking about. In my electronic texts and back in switching powersupplies 101, a Baker clamp is/was a diode connected in parallel with the base-collector junction of a bipolar transistor - typically used in switching applications.
For a Darlington transistor, this diode can be a silicon device. For a non Darlington transistor, it must be a schottkey - so long as its forward voltage drop is less than Vbe-VceSat.
It is there to prevent saturation of the transistor, allowing it to switch faster.
lumanauw said:
You're right. An Si diode baker clamp would only work on a BJT driven with an emitter follower (as in either a buffered VAS in an audio amplifier or a flyback switching transistor in a SM power supply). Its been a while since I have looked at my obsolete TV power supply circuits
Anyway, from memory, here is how it is done on a un-buffered BJT switch with a si diode:
Attachments
Re: Re: Baker clamp
Hi Glen,
You're perfectly right and in a perfect world, above discussion should be totally unnecessary. But the previous posts clearly demonstrate that not everybody attributes the same meaning to a Baker clamp, or even worse, doesn't recognize that this kind clamping -no matter of what application- is based on NFB.
Therefore, I think that this discussion was valuable and will prevent further confusion in the future.
Cheers,
Edmond.
G.Kleinschmidt said:
Hi Glen,
You're perfectly right and in a perfect world, above discussion should be totally unnecessary. But the previous posts clearly demonstrate that not everybody attributes the same meaning to a Baker clamp, or even worse, doesn't recognize that this kind clamping -no matter of what application- is based on NFB.
Therefore, I think that this discussion was valuable and will prevent further confusion in the future.
Cheers,
Edmond.
Re: Re: Re: Baker clamp
I recognize the Baker clamp simply as a circuit technique to avoid transistor saturation.
What would be the reasoning and advantages of looking at a Baker clamp as a negative feedback loop?
For a particular implementation as Glen illustrated above, how would you define the loop gain and how would you use it to analyze the circuit behaviour?
Edmond Stuart said:
I recognize the Baker clamp simply as a circuit technique to avoid transistor saturation.
What would be the reasoning and advantages of looking at a Baker clamp as a negative feedback loop?
For a particular implementation as Glen illustrated above, how would you define the loop gain and how would you use it to analyze the circuit behaviour?
Re: Baker clamp
Hi Edmond,
Thanks for pointing this out. I stand corrected. It is probably both. Tom may have been taking some liberty in the use of the term, although I think that National uses the term in the same way in describing their line of IC power amplifier drivers.
Cheers,
Bob
Edmond Stuart said:
Hi Edmond,
Thanks for pointing this out. I stand corrected. It is probably both. Tom may have been taking some liberty in the use of the term, although I think that National uses the term in the same way in describing their line of IC power amplifier drivers.
Cheers,
Bob
Miller compensation
You're welcome.
Hi Bob,
Now my next point of confusion: the Miller compensation.
If the Miller cap encompasses only one tranny, it's clearly Miller.
If this cap encompasses two trannies (a Darlington or so), it's still Miller.
If this cap encompasses four trannies, as in your EC amp (C4=20pF), should we still call it a Miller compensation, or just lag compensation?
But if we call it lag compensation, we create another point of confusion. According to Mike KS (you certainly remember him) it's lead compensation. You (and me) disagree on that, but he has a point: C4 sees a voltage that is (almost) exactly equal to the output voltage. If C4 was directly tied to the output, than it certainly was lead compensation.
Comments invited.
Cheers,
Edmond.
Bob Cordell said:
You're welcome.
Hi Bob,
Now my next point of confusion: the Miller compensation.
If the Miller cap encompasses only one tranny, it's clearly Miller.
If this cap encompasses two trannies (a Darlington or so), it's still Miller.
If this cap encompasses four trannies, as in your EC amp (C4=20pF), should we still call it a Miller compensation, or just lag compensation?
But if we call it lag compensation, we create another point of confusion. According to Mike KS (you certainly remember him) it's lead compensation. You (and me) disagree on that, but he has a point: C4 sees a voltage that is (almost) exactly equal to the output voltage. If C4 was directly tied to the output, than it certainly was lead compensation.
Comments invited.
Cheers,
Edmond.
Re: Miller compensation
You do not call it Miller or lead-lag compensation based on the number of trannies the loop encompasses.
A lead-lag compensator has a transfer fuction of
H(s)=[(s+z1)*(s+z2)]/[(s+p1)*(s+p2)]
The circuit behaviour is depending of the absolute and relative positions of the zeros and poles. To save a long discussion take a look here: http://eceserv0.ece.wisc.edu/~cobb/ECE409/Exp4.pdf
The Miller compensation, also known as 'dominant pole compensation', always introduces a dominant pole (one which masks the effects of other poles) into the open loop frequency response. The particular case of a Miller compensation encompassing two trannies (a Darlington) is from this perspective identical with the one tranny case.
Edmond Stuart said:
You do not call it Miller or lead-lag compensation based on the number of trannies the loop encompasses.
A lead-lag compensator has a transfer fuction of
H(s)=[(s+z1)*(s+z2)]/[(s+p1)*(s+p2)]
The circuit behaviour is depending of the absolute and relative positions of the zeros and poles. To save a long discussion take a look here: http://eceserv0.ece.wisc.edu/~cobb/ECE409/Exp4.pdf
The Miller compensation, also known as 'dominant pole compensation', always introduces a dominant pole (one which masks the effects of other poles) into the open loop frequency response. The particular case of a Miller compensation encompassing two trannies (a Darlington) is from this perspective identical with the one tranny case.
Re: Re: Re: Re: Baker clamp
Use of Schottky diodes was widely seen in S, LS, AS, and F logic families:
http://www.computerhistory.org/semiconductor/timeline/1969-Schottky.html
Pete B.
Use of Schottky diodes was widely seen in S, LS, AS, and F logic families:
http://www.computerhistory.org/semiconductor/timeline/1969-Schottky.html
Pete B.
Miller compensation is the cap, connected between input and output of inverting voltage-controlled current source
miller compesation in not equal to 'dominant pole compensation'
learn by heart Solomon paper:
http://www.ee.unb.ca/Courses/EE3111/DFL/AdditionalMaterial/NSopamptutorial.pdf
dimitri said:
I'm afraid there's a confusion between the "Miller effect" and "Miller compensation".
In your reference, I can't find anything beyond mentioning the Miller effect, that is, multiplying the capacitance connected across a voltage gain stage.
Quote from your reference: "We note at this point that p1, which represents the dominant pole of the amplifier, is due simply to the familiar Miller-multiplied feedback capacitance gmR2Cp combined with input node resistance, R1."
From this perspective, "Miller compensation" doesn't exist at all, there's only a frequency compensation using the Miller effect. Others (including Wikipedia and apparently E.G. Stuart) are calling the dominant pole compensation "Miller compensation".