Hello everybody,
I'm trying to do what the datasheet says I shouldn't: Use a TDA2030A as unity gain buffer. I did a simulation with no further compensation and the result seems to be a nice oscillator, as was expected. However, if I run the sim with the recommended snubber of 220nF and 1R on the output, I get a different result. If I understood it correctly, the phase margin is about 35 degrees, which should yield a stable operation. Not sure how 'fragile' this state is under actual conditions though. Does this look like it would start oscillating under certain conditions? Your opinions, please 😱.
The datasheet of the L165, which is basically the same as the TDA2030A, shows another approach for a unity gain buffer. That one is certainly inverting and only useable with split supplies (or else I'd need a virtual ground), but gives a much higher phase margin.
The model I used is the standard UniversalOpamp2 with the following parameters:
SpiceModel: level.3b
Value2: Avol=10k GBW=2Meg Slew=8Meg
SpiceLine: ilimit=3.5 rail=0 Vos=2m phimargin=0
SpiceLine2: en=2u enk=0 in=100p ink=0 Rin=500k
Open loop gain and phase look like the plot in the datasheet of the L165, so I assume the model is about right for this purpose. TDA2030A datasheet contains no phase plot.
Regards,
Lasse
I'm trying to do what the datasheet says I shouldn't: Use a TDA2030A as unity gain buffer. I did a simulation with no further compensation and the result seems to be a nice oscillator, as was expected. However, if I run the sim with the recommended snubber of 220nF and 1R on the output, I get a different result. If I understood it correctly, the phase margin is about 35 degrees, which should yield a stable operation. Not sure how 'fragile' this state is under actual conditions though. Does this look like it would start oscillating under certain conditions? Your opinions, please 😱.
The datasheet of the L165, which is basically the same as the TDA2030A, shows another approach for a unity gain buffer. That one is certainly inverting and only useable with split supplies (or else I'd need a virtual ground), but gives a much higher phase margin.
The model I used is the standard UniversalOpamp2 with the following parameters:
SpiceModel: level.3b
Value2: Avol=10k GBW=2Meg Slew=8Meg
SpiceLine: ilimit=3.5 rail=0 Vos=2m phimargin=0
SpiceLine2: en=2u enk=0 in=100p ink=0 Rin=500k
Open loop gain and phase look like the plot in the datasheet of the L165, so I assume the model is about right for this purpose. TDA2030A datasheet contains no phase plot.
Regards,
Lasse
Attachments
stability is a big subject, sims without good models may not tell you enough
"noise gain" compensation can be used with these power op amps with a minimum stable gain spec
but buffering with another amp is likely a good idea - otherwise it may go unstable if the input is disconnected/not powered
"noise gain" compensation can be used with these power op amps with a minimum stable gain spec
but buffering with another amp is likely a good idea - otherwise it may go unstable if the input is disconnected/not powered
Guess I'll use noise gain compensation, then. Same component count, much better looking simulation.
Thanks 🙂.
Thanks 🙂.
- Status
- Not open for further replies.