The additional filtering was to isolate the input stage from big current bursts in the OPS.
But I'm no expert! 😎
But I'm no expert! 😎
It is a good buffer.
I tested with a good opamp as input and I let the output be 10Vp.
At 1kHz THD was 0.0003% ... that is -110dB.
I tested with a good opamp as input and I let the output be 10Vp.
At 1kHz THD was 0.0003% ... that is -110dB.
Optimized the resistors' value around the bias network.
Added degenerative resistors R29, R32 to Q5, Q6. This solves the transient current spikes. It also limits the current of Q7 and Q8.
Bode plot
Added degenerative resistors R29, R32 to Q5, Q6. This solves the transient current spikes. It also limits the current of Q7 and Q8.
Bode plot
Last edited:
Hi again!
Spent some time with layout... first one unsuccessful, but here is a finished board.
Latest schematic for this PCB:
PCB:
Spent some time with layout... first one unsuccessful, but here is a finished board.
Latest schematic for this PCB:
PCB:
Last edited:
I bought some 10uF ceramic caps before, primary for audio input coupling.
I realize the voltage here is pretty low. They are perfect here.
The price before tariffs. The 16V version.
The 50V version is about double the price. The 16V version is enough here.
Keen to see you build working and any changes that result. Nice design
My modelling of this circuit shows the current in the output stage shows a strong negative temperature characteristic. I use Microcap and I'm not sure its best for this purpose. Keen to know if the standing current is stable in practice. I assume that the MJE's are on the heatsink with the IRF's.
Hi, just small point, you have chosen X7R MLCC for signal path, while they are very useful in PS, they are very nonlinear. Better use electrolytic there, and larger value as Morten did.
Ideally, a 100uF electrolytic cap paralleling with 0.1uF ceramic cap should be used here. The caps only pass Error Correction signal. The amp would work even without those caps. At low frequency, the globe NFB would dominate. Thus, the bandwidth here is not critical. Choosing ceramic type cap is major for its small footprint.
Thanks for mentioning the nonlinearity of X7R. I overlooked this portion.
Hi @m0rten, how is your board doing? My board has arrived. To my surprise, the Error Correction loop is not fully stable. Removing C311/C312 (#25) fixed oscillation (about 50mVpp @30MHz). That means the oscillation is caused by Error Correction alone. Yea..I can reduce the Error Correction feedback ratio, but it is not longer holding the advantage to other solution such as CFP.
I would try resistors in series with C311 + 312. Begin with trimpots of 1k and see how far it is possible to reduce the resistance.
Increasing R323 + 324 could also help. If you have access to Bode plot use it.
A few pF in paralell with R310 is standard. I was surprised when i saw the concept the first time without it.
Increasing R323 + 324 could also help. If you have access to Bode plot use it.
A few pF in paralell with R310 is standard. I was surprised when i saw the concept the first time without it.
I did investigate the stability based on the schematic in #1. The phase margin is about 30 degrees. I thought it should be unconditional stable as there is only 2 active components in the Error Correction loop.
Adding resistors in series with C311/C312 is a good way to compensate. I would keep the value small to ensure the Error Correction NFB ratio is about 20dB. I would also add caps in parallel with R320/R322. They would create a dominate pole. The beauty of this pole is that it only shows up in the Error Correction feedback. It won't show up in the global negative feedback loop.
Like this.
Its bode plot.
BTW, the global NFB unit bandwidth of the entire amp is about 1MHz, which is very conservative for me. Almost 90 degrees phase margin. No need to add small cap to the feedback resistor in this case.
Last edited: