The (high-cap.) unregulated PSU for chipamps

[carlosfm]

[Peter Daniel]

I see you've implemented Peranders suggestions, right?

[carlosfm]

Originally posted by Peter Daniel
I see you've implemented Peranders suggestions, right?

Not quite, as he was inserting the snubber on the chip's pins (after the caps), and he was using the same values as on the PSU.
And he doesn't use bypass caps.
Let's say that P-A knows this snubber circuit from me, no?
But well, I've been thinking in trying this, and it works.
Not a big improvement as the first snubber provides, but an extra degree of midband and treble detail.

The midband is gorgeous.

[homer09]

Can someone suggest a "hack" to integrate this snubber update onto Brian's LM3875 Rev3 or LM3886 boards? Im trying to figure it out, but im lost

also MKT caps? is this the same as metallized polypropylene, ie the blue BC caps included in Brian's kits?

[carlosfm]

Originally posted by homer09
Can someone suggest a "hack" to integrate this snubber update onto Brian's LM3875 Rev3 or LM3886 boards? Im trying to figure it out, but im lost

I don't see where's the problem...
On the film cap's place use the 100uf electrolythics.
Bypass the electrolythics directly under them, under the board.
On the bigger holes for the 1500uf caps insert the snubber.

Originally posted by homer09
also MKT caps? is this the same as metallized polypropylene, ie the blue BC caps included in Brian's kits?

Those are fine.

[moving_electron]

Originally posted by homer09
Can someone suggest a "hack" to integrate this snubber update onto Brian's LM3875 Rev3 or LM3886 boards? Im trying to figure it out, but im lost

also MKT caps? is this the same as metallized polypropylene, ie the blue BC caps included in Brian's kits?

MKT is polyester film.

Regarding using Brian's kit, the 3886 power supply and presumably the new LM3875 snubberized kit have the parts and boards to do the traditional snubber (Oh so January :-) ) that is the schematic that Carlos posted. So that part is in good shape.

The change that Carlos is recommending is adding a few things at the chip and is in the comments.

The 100Uf and 100nF is already there sort of. The 100uF caps are included in the 3886 kit for example. The 100nf is a polypropylene box in the 3886 kit. So you would need to use a 100nf MKT (polyester) instead that was not in the kit. So the holes are there but you would need to add this part ( quantity 4 - 2 for left, 2 for right amp )

The 300nF from V+ to V-. THis you would need to add. You would need 2 of these, one for each channel. I would expect to be able to just let the 3886 pins hange down below the board and solder to those underneath the board. Perhaps use a bit of insulation from some scrap wire to insulate each leg of the cap if it has to span across any traces.

For the snubber of 0.47/3w + 47nF: You have to add this. You would need 4 of these snubbers. Hmmm. What about going from the V+ in to CHG (?) underneath the board. Again some insulation from some scrap wire likely will be needed. Do the same for V- to CHG.

Edit: Carlos' method is cleaner, I would think for a new board that is not yet populated,

[homer09]

thanks guys, both methods make sense to me. guess i missed the obvious there....

[Nuuk]

OK some questions for Uncle Carlos (or anybody else who would like to answer them).

We now have two snubbers between the 10,000uF caps and the 100uF caps on the pins of the chip. Why is that and could these not be combined into one?

Why is the 0R47 resistor specified as 3W when the 1R is specified as 2W? Is this because you can only find the 0R47 in 3W wirewound? If so could we use a pair of 1R/1W resistors in parallel?

We seem to be adding a lot of polyester caps into the circuit and I have always understood these to be inferior to the polypropylenes. I presume that the polyesters are specified here for their comparitvely smaller size. Has anybody tried the mettalised polyphenylene sulphide caps?

[peranders]

Originally posted by Nuuk
Why is the 0R47 resistor specified as 3W when the 1R is specified as 2W? Is this because you can only find the 0R47 in 3W wirewound? If so could we use a pair of 1R/1W resistors in parallel?

The min rating of this resistor is << 0.6 watts => a plain 0.6 watt metall film will do and you will also get a "purer" resistance, only advantages.

I have used 3 x 1206 10 ohms and those aren't burnt yet.... see picture below, R9-R11

[peranders]

Originally posted by Nuuk
We seem to be adding a lot of polyester caps into the circuit and I have always understood these to be inferior to the polypropylenes. I presume that the polyesters are specified here for their comparitvely smaller size. Has anybody tried the mettalised polyphenylene sulphide caps?

Don't forget that high resonance frequency is one good property and big polyprops have lower resonance frequency than small polyester ones.

I have tested PPS caps but only as input coupling caps. The are too big for ordinary decoupling. For this purpose I use 100 nF X7R 0805 size.

C1 in the picture is PPS, C8, C9 are X7R.