I came across an issue where a spice model isn't made for a part (the one in question is the TPA3255). I'm not really trying to do any distortion measurements as much as just checking for signal integrity from input to power stage. Is there a way to simulate a chip amp like that using the universal op amps? If not, is there a way simulate the power uplift from the output stage? (I was also considering just doing a simple transistor class A or something just to complete the simulation, but I figured I would ask just in case)
Oh man. So since I posted this I've been on a bit of an odyssey to figure this one out. So after asking around I actually found someone who knew one of the folks over at TI. The short explanation is that they don't have a Spice model for it because if they did it would apparently give away trade secrets. They don't even have a model internal to them for it. This led me deeper into the spec sheet. I took a closer look and found that the amplification factor for it was 21.5db at 7Vp-p. I used an ideal op amp and set the output maximum to be 600 watts at 2 ohms (maximum possible output for the TPA3255). Funny thing is, when I tried to implement the PFFB circuit in spice it couldn't stay stable. Apparently TI recommends you only test it under specific conditions which kinda tells me it's not a terrific product overall so I'm considering a design rework.
Pretty straight forward class D amp
with good ratings.
Zillions of pre made boards.
Turn on the power switch, 2 wires
to the speaker. And done.
Having fun with simulation
and amplifier topology is not found with
Class D.
Plenty more fun to just model
classic known topologies. Too learn
and diagnose amps in sim
with good ratings.
Zillions of pre made boards.
Turn on the power switch, 2 wires
to the speaker. And done.
Having fun with simulation
and amplifier topology is not found with
Class D.
Plenty more fun to just model
classic known topologies. Too learn
and diagnose amps in sim
I won't deny that. On paper it definitely seems like a good chip. So I was told, the issue is with the fixed gain of the chip itself. With only 21.5db of gain, anymore feedback than that of the PFFB app note circuit and it destabilizes.
Why specifically class D in that regard?
Yes and no. Imo at some point if you only stick to classic circuits, you miss out on the chance of stumbling on to something new. Great thing about sim is you don't have to worry about summoning the magic smoke lol
Full system would have a preamp
or opamp stage somewhere.
So fixed gain not much a issue.
And modeling would be
helpful for designing pre gain
stage.
You can build basic self oscillating
class D amps in sim.
And even such simple models
can crash the software pretty often.
Full model of commercial chip.
Can only guess would be a
nightmare for most desktops.
or opamp stage somewhere.
So fixed gain not much a issue.
And modeling would be
helpful for designing pre gain
stage.
You can build basic self oscillating
class D amps in sim.
And even such simple models
can crash the software pretty often.
Full model of commercial chip.
Can only guess would be a
nightmare for most desktops.
This actually kind of explains some of the experiences I've been having while trying different things. The INA851 sticks out for me in that regard. Very cool op amp, but definitely has a lot of little idiosyncrasies to account for like the clamp voltage vs common mode and both both sets of gain stages.