Datasheet BPA200 in practice

[Einric]

Is the datasheet BPA200 really a good design?
It looks like there is a buffer opamp of some kind on the input.
This design would be very easy to implement.

[Kindhornman]

Elnric,
I asked the same question the few days ago. I have yet to find somebody who has built this as drawn by National Semiconductor. I am still trying to find that person who has this exact circuit in operation.

Steven

[Einric]

I'm actually looking to bridge my two parallel LM4780's.
This looks like a very simple layout with a buffer to drive the inputs.

[Minion]

I believe that the buffer opamp is used because it gives a constant stable low impedance output to all 4 lm3886 inputs , half of the LM3886 are inverted so the input impedance affects the gain of these amps so a buffer is needed to make sure the gain is stable on the inverted amps no matter what the output impedance of the source is .....

Never built a BPA but of the paralell and bridge amps i have built i found the paralell amps quite tricky to get right and very unforgiving and the bridge amps much easier to get right .......

Cheers

[abraxalito]

Originally Posted by Einric

Is the datasheet BPA200 really a good design?

No, its not. But its not too hard to fix it up.

The primary reason why its bad is the 22uF capacitors. Electrolytics have tolerances such that the LF gain will not be sufficiently well matched in most instances - with the result that the paralleled amps will fight one another. Raising the value to 1000uF fixes this.

[Kindhornman]

abraxalito,
Two questions in one then. What would be better the higher value capacitor or a close match of the capacitors at the low value. If this was within a much closer range would it solve the problem. Can a film capacitor be substituted for the electrolytic at 22uF ? Couldn't all of these capacitors be film capacitors where they are in this portion of circuit?

[abraxalito]

Either would work fine - I just proposed the larger value because its cheaper and potentially smaller size. I have run a simulation on this in the past and I think you'd need better than 1% matching if using a 22uF, but my memory is hazy. No reason a film cap can't be used, no, just hard to find such close tolerances. Hand matching is an option too provided lack of stability over temperature and time doesn't compromise the matching.

[Kindhornman]

If the film cap is of a high voltage type, even though it would be larger would that take care of your concern about the temperature rise and stability over time? Wouldn't a film cap also have a better sound quality?

[abraxalito]

The temperature stability of a cap is mostly down to the kind of dielectric used, not the voltage rating. Stability over time I don't know anything about - though I'd guess that winding tightness has something to do with it, meaning boutique-style audiophile caps would be worse.

As regards sound quality - again not much experience listening to caps - the only comparisons I have done are between film and NP0 in active XOs, where the NP0s won easily. My guess would be that the lower the voltage across a cap, the less it can affect the sound. But also smaller is better because a bigger cap picks up more RF hash and has higher series inductance meaning more chances of RF resonances.

[Kindhornman]

So if I follow you correctly then the caps value should be close to the voltage rating of the circuit and not just some arbitrary high voltage. So a 50v cap would be good on a 35vdc circuit? And those boutique caps are so pretty......... I've seen normally produced caps made to look special by the fancy script printed and colored outside wrap while being exactly the same as a production film capacitor. Watched them being made once...