Hi everybody
To address (potential) stability issues I'm testing a power amp (under construction) with different loads. Attached illustrations show the square wave response into 8R and into 8R with 100nF in parallel.
I hoped for a minor, and well-damped, overshoot for the capacitive load, but what I get more looks like some sort of "under-shoot", or over-damped over-shoot.....
Would anybody happen to have seen this before and know the cause?
The amp sounds and measures very well and my question is purely out of curiosity.
Thanks,
Nic
To address (potential) stability issues I'm testing a power amp (under construction) with different loads. Attached illustrations show the square wave response into 8R and into 8R with 100nF in parallel.
I hoped for a minor, and well-damped, overshoot for the capacitive load, but what I get more looks like some sort of "under-shoot", or over-damped over-shoot.....
Would anybody happen to have seen this before and know the cause?
The amp sounds and measures very well and my question is purely out of curiosity.
Thanks,
Nic
Check whether the period of the (damped) oscillatory ringing, happens to be similar to the LC resonant frequency of the amplifier's output inductor and the load capacitor.
Many amps are marginally stable into certain capacitive loads. Try smaller or larger capacitor values, and
it should visibly change.
it should visibly change.
You may want to do this with 20 feet of speaker wire between the amp and the load, and maybe a real speaker. Stability can be the cause of amp failure due to output shoot-through currents. You can improve stability either by increasing the compensation cap / dominant pole, at the cost of slew rate, or increasing IPS degeneration resistors at the cost of THD. So, your call on what's important to you. If you haven't already, consider two-pole compensation.
Period is around 1.3 µs (or 0.8MHz). There is no output inductor except for some 20 cm of "speaker" cable to the dummy load (non-inductive Caddock MP850 resistors).
I will try different cap values.
I will try different cap values.
Without the inductor, be careful with the capacitive loading, or there could be run away oscillation.
See post #5. If your normal length speaker cable (connected before the capacitor) kills the oscillation, it should be ok.
See post #5. If your normal length speaker cable (connected before the capacitor) kills the oscillation, it should be ok.
Here’s a few slides on why output L in feedback amps are important. You can decide not to use them, but you should then take precautions on your loop gain (reduce it) and phase.
https://hifisonix.com/output-l_1/
https://hifisonix.com/output-l_1/
Nic,
The ringing comes from the cascode.
An output L//R might help to isolate the load capacitance.
And / or a R-C to Gnd (Zobel).
Patrick
The ringing comes from the cascode.
An output L//R might help to isolate the load capacitance.
And / or a R-C to Gnd (Zobel).
Patrick
You need an output L.
You should post the schematic up so people can give you proper feedback.
Even non feedback amplifiers can show this problem. You have a complex load (even if it’s just a piece of wire to a resistor it still has L, R and C) and output devices, if operating in EF configuration, can reflect that back as negative base resistance which leads to instability. A LR output coupling network will help to solve this as will base stoppers.
You should post the schematic up so people can give you proper feedback.