Yes, it would be great. Very interesting. My build is rock solid 🙄
I will check a couple of ideas later today...
I hope, oscillation is not too violent? Not full amplitude swing?
Yeah, very interesting...
Terry does great job on testing our stuff. Still a bit surprised with the small differences between the way our builds behave (for example, he needs more R between the gates for biasing).
Good thing - after we fix it with Terry, it's normally bullet proof
🙂
Yeah, very interesting...
Terry does great job on testing our stuff. Still a bit surprised with the small differences between the way our builds behave (for example, he needs more R between the gates for biasing).
Good thing - after we fix it with Terry, it's normally bullet proof
I checked and with the TubSuMo IPS attached the trimmer on the OPS is at 47R with 16mV across the emitters.
I tried adding another 47pf cap to C26 and them C25. Very little change. I am attaching a pic of the oscillation at 1khz. It starts oscillating at about 2vac output.
I will attach it to a Slewmaster OPS next.
Blessings, Terry
Attachments
OK, I just tried a Slewmaster OPS. It oscillates badly. Through a 150W light bulb the most voltage I can get on the rails is +-41V. Again, D15, D17 pulse with the oscillation. It won't level out, even with the load removed on the output.
I have double and triple checked the parts and they are all per schematic. I'm not sure why it is stable for you. Maybe I'm missing something.
Thanks, Terry
I have double and triple checked the parts and they are all per schematic. I'm not sure why it is stable for you. Maybe I'm missing something.
Thanks, Terry
95 degrees phase margin, 20db gain margin
Here we go - two 2.2pF caps marked with green circles (C27, C28).
Not very convenient - need long pins.
They make the loop gain/phase responses just perfect.
Optimal value for C18 becomes 4.7pF - I've got no such value in stock, so left it unchanged so far, no problem.
Here we go - two 2.2pF caps marked with green circles (C27, C28).
Not very convenient - need long pins.
They make the loop gain/phase responses just perfect.
Optimal value for C18 becomes 4.7pF - I've got no such value in stock, so left it unchanged so far, no problem.
Attachments
Have you got 4 of them?
Then put 2 in series instead of one 2.2pF - it will be a very close value...
Then put 2 in series instead of one 2.2pF - it will be a very close value...
Made it worse, a lot worse. I reconnected it to the new tubsumo OPS and without the new caps and no load there is no oscillation. Sine waves and square waves look beautiful. As soon as I connect a load to the output I get oscillation. For sure the new caps you suggested are not the answer. Still won't work at all with the Slewmaster OPS
Thanks, Terry
That's why I call this 2 stage gain amp the "beast". Hard to tame.
(below 1) is the old sansui 2 stage (like our designs).
You will note sansui just uses a simple 2 device VAS.
This design does not perform as it is simulated.
We might be "embellishing" this particular design too much.
It might be better to "muntz" the design 😀 - like (below2)
- drop the baxandall/cascode.
Very little overall performance drop , but much greater margin.
In this form the lag/lead/local compensation works as it should.
OS
Attachments
I don't at all understand why Valery's amp is stable and mine isn't. I followed the as built to the letter. Also, why is it so much worse with the Slewmaster OPS than with the lateral OPS?
It's the super pair 🙁 ... I just got mine to go sub ppm (without the baxandall).
I will update on my thread. Thimios will test. 🙂🙂
OS