LM3886 schem, feedback cap.

[Ivarsson]

Hi!

In the (official and standard) schematic for the LM3886 there is an optional feedback capacitor with value 22µF. Most of the inofficial schematics write this as a polar (electrolytic) capacitor. Wouldn't this be kind of bad since it is connected to output through the feedback resistor and would see both positive and negative voltages compared to ground?

[KSTR]

In normal operation this cap (Ci on the datasheet frontpage) sees only AC little voltages (millivolts at best) accross it because its impedance at audio frequencies is low compared to the resistors. Only under a continuous DC signal with the wrong polarity applied this cap would suffer a little, although the resistors help avoid catastrophic failure (they limit the current to very low levels).

It is pefectly ok to use a polar electrolytic here when the circuit is properly designed and used.

- Klaus

[Ivarsson]

Thank you very much, that makes perfect sense!
Me and a friend are planning to build one dual channel amplifier each with this chip and I am now choosing components and designing a PCB:

[ionomolo]

Originally posted by Ivarsson
Hi!

In the (official and standard) schematic for the LM3886 there is an optional feedback capacitor with value 22µF. Most of the inofficial schematics write this as a polar (electrolytic) capacitor. Wouldn't this be kind of bad since it is connected to output through the feedback resistor and would see both positive and negative voltages compared to ground?

I agree with KSTR about the low signal levels at the cap, but i suggest a higher value (47uF - 100 uF) to get a better bass response. 22 uF + 1K has its corner frequency at 7.2 Hz, this seems fine but electrolytics distort near the cutoff frequencies (nothing to do with strange unmesurable effects) so it's better to place the corner at a lower value.

[KSTR]

I can back ionomolo's suggestion. Indeed, most designers choose an electrolytic cap about 5x...10x larger than the smallest meaningful design value with a "perfect" cap, say a good film type. Besides the distortion issue another reason is aging and less controlled tolerances of electrolytics. When the corner frequency is low enough, say about 1Hz, these kind of problems don't have much effect anymore.

- Klaus

[CFT]

The design in my amp has Ci (10 uF BG N) but no Cin.

Does anyone have experience whether this configuration is better or not in terms of sonic performance than having Cin but not Ci? (I understand that there will be no DC gain if Ci is in place). Thanks!

[AndrewT]

Originally posted by CFT
The design in my amp has Ci (10 uF BG N) but no Cin.

Does anyone have experience whether this configuration is better or not in terms of sonic performance than having Cin but not Ci? (I understand that there will be no DC gain if Ci is in place). Thanks!

Hi,
miss out both to create a DC coupled amplifier. But, include a DC servo and DC detect and cutoff.
or
include both to make an AC coupled amplifier.

But, don't miss out one to make a mixed AC & DC coupled amplifier.