LM386 Reverb Driver

[loogie]

I'm using the simplest circuit from the data sheet and it drives the reverb tank nicely -- or at least it sounds good to me.  I plug it into my effects loop. 

Its on a breadboard and I'm using a wall-wart for power.  At first I just fed it a signal and plugged in a speaker.  It hummed so I put the biggest cap I could find, about 1000uf, across the wall-wart's output.  That helped tremendously, but it still hums a bit.  I wonder if that much capacitance could blow up my wall wart?  Its just a leftover from a cordless phone.

I was thinking I would add an RC section or two.  Or should I put in a voltage regulator?  Anybody ever done this sort of thing?

[FYL]

Most consumer grade DC wall-warts are very noisy. They are designed to a price point for general apps, not for the lowest noise in audio or MI applications.

The LM386 in std x20 gain mode shows very poor PSSR, around 6 dB if I remember well. That's OK with batteries - pure DC, no 60/120 Hz - but not with most power supplies. You need a dedicated PS, preferably regulated.

[loogie]

Well I took your advice (sort of).  I bought a 7805 regulator from RS (that's what they had) and added it and another big cap.  Its very quiet now and the reverb sounds real nice -- there's more than enough of it. 

I'm sure you're right about the dedicated power supply and if I pursue this idea further I'll do that.  Highly recommended for anyone who is short a tube for a reverb driver.

[PRR]

> using the simplest circuit from the data sheet

Yeah, it "works".

You might look at the next several plans. All have an optional "bypass" cap. They don't say (you were supposed to know from LM380 app-notes) that a cap from pin 7 to ground shunts power crap away from the more-sensitive input stage. The data is top-center page 4. No-cap, power crap gets to speaker at -6dB. A mere 10uFd from 7 to ground will cut 120Hz crap to -40dB, a BIG improvement from the "minimum parts" plan.