Often I leave power ON always to just the dc servo circuit..... offers no delay in correcting offset when amp is turned on-off.
Reminder: The idea was just to correct for thermal drift in offset...... not really to be an auto-zero for production dc offset. Basically, the dc offset should be very low without servo operation. If not, a servo can increase (via unbalance) thd, esp 2H.
If you learn the range and amount of drift, a dc servo can be tailored to minimize the thermal offset drift.
THx-RNMarsh
Reminder: The idea was just to correct for thermal drift in offset...... not really to be an auto-zero for production dc offset. Basically, the dc offset should be very low without servo operation. If not, a servo can increase (via unbalance) thd, esp 2H.
If you learn the range and amount of drift, a dc servo can be tailored to minimize the thermal offset drift.
THx-RNMarsh
Last edited:
Richard, R8 could be made a parallel combination of 330R resistor and 10k trimpot and then, before the DC servo opamp was inserted to its socket, the output DC offset should be trimmed inside few millivolts. Now does it cover your requirement regarding a servo, the opamp output swing is lees then volte.
BR Damir
Does the offset stay put without servo? How far and how fast will it drift from cold (room temp) to max temp? Is the time constant of the servo compatible with the offset drift rate and yet not intrude into the audio?
A lot depends on heat sink area, build construction, thermal matching of input pairs, etc.
Yes, you need to trim (wamed up) before inserting servo opamp.
THx-RNMarsh
A lot depends on heat sink area, build construction, thermal matching of input pairs, etc.
Yes, you need to trim (wamed up) before inserting servo opamp.
THx-RNMarsh
Last edited:
I am not sure if this one satisfied your requirements, THD 0.001% at 200W, 500V/usec, as you did not say THD at what frequency. This amp do 0.0006% at 1kHz and 200W.
I tried here to introduce a bit different DC servo connection. It injects DC current in to VAS resistors, but those resistors are bypassed with 1 uF caps to widen open loop bandwidth at 20 kHz and in that way lower THD20k.
First step is to adjust DC offset with P1(best way is to use parallel combination of a fix resistor and a trimpot) with no DC servo OPA inserted . Next step is to adjust the OPS bias, and repeat the procedure if needed. In this case DC servo can cover temperature drift from 25 to 45 degree Celsius with its output going from +8 V to -8 V.
Richard tell me what is your opinion here.
BR Damir
Attachments
I decided not to experiment with the DC servo connection.
Here I put two lateral MOSFET 200 CFA latest schematic. Both use simple VAS.
First one uses diode in the VAS emitter, quite simple circuit with excellent data.
Damir
Here I put two lateral MOSFET 200 CFA latest schematic. Both use simple VAS.
First one uses diode in the VAS emitter, quite simple circuit with excellent data.
Damir
Attachments
Second one uses VAS with emitter resistor but shunted with quite big capacitor(polypropylene). It shows a bit better data. The layout can have possibility for both VAS types, so one can listening and decide what sounds better.
Damir
Damir
Attachments
....another PCB has born ....
Regards,Alex
Alex you are really fast, excellent work.
Just one thing, Vbe multiplier (Q18) does not need to be close to the heat sink, no thermal contact needed. You can put it in the middle and make connection shorter.
Cheers Damir
Hi Damir,
A very cool design
One "conceptual" question. In some of my designs, I also have this kind of phase response "sagging" at around 1MHz or a bit higher, and then going up again closer to ULGF. Sometimes I try making it less "curved", sometimes I leave it "as is", if it's not too aggressive.
What do you think about it?
Cheers,
Valery
A very cool design
One "conceptual" question. In some of my designs, I also have this kind of phase response "sagging" at around 1MHz or a bit higher, and then going up again closer to ULGF. Sometimes I try making it less "curved", sometimes I leave it "as is", if it's not too aggressive.
What do you think about it?
Cheers,
Valery
Hi Damir,
A very cool design
One "conceptual" question. In some of my designs, I also have this kind of phase response "sagging" at around 1MHz or a bit higher, and then going up again closer to ULGF. Sometimes I try making it less "curved", sometimes I leave it "as is", if it's not too aggressive.
What do you think about it?
Cheers,
Valery
Thanks Valery,
I am not sure what you mean that phase is going up closer to ULGF? In mine plot that is just a case, it can go over that ULGF line (gain).
I do it intentionally, as I am trying to have Two Pole compensation, more exactly Two Pole Output Inclusive Compensation. Shell I call it TPOIC? If you look how TPC behaves in a VFA where this phase behavior is more pronounced and phase "sagging" could go close to -180 degree and then go up to less then -120 degree, and it's still stable.
P.S. It looks that I am only one who advocate this kind of compensation, TPOIC.
Best wishes, Damir
Hi Damir
TPOIC performs better than OIC but I think it is still sub-optimal.
If the loop gain around the OPS is specified (maximised feedback to minimise distortion, within stability limits) then a two pole Miller Input Compensation should allow more feedback around the rest of the amplifier.
This should reduce total distortion, of course.
Probably doesn't make much difference in practice because OPS distortion dominates, so I am sure your amplifiers perform well.
But I think there are other benefits, the increased loop gain around the VAS effectively lowers its output impedance, which raises the frequency of one of the amplifier poles. This improves stability a little, at no cost to distortion.
Best wishes
David
Hi David,
I have tried different compensations on this simple amp, and TPMIC (as TPC or TMC connected) don't came close to TPOIC regarding THD or PM and GM. For this amp is ideal to deploy OIC(TPOIC exactly) in my opinion. Attached is the zip file with two last presented 200W LMOSFET amp and if you find the time to play with it, show me that I am wrong, the best thing of all in this forum is to learn from other people. You said, main distortion source is the OPS and OIC lowers exactly that type of distortion, I think, hope
Cheers, Damir
Hi Richard, I think Krisfr will soon order some board to be produced, we can wait.
BR Damir
I think that the C12 and C15 will have to be a little bigger, the 16mm dia. is a little small. I did not find any of that size on Ebay, maybe someone could suggest a source. I have 19mm caps of that spec. but they will fit only if I change the layout, which is very easy and doable there is room to do it.
Attachments
Last edited: