using 386 to LDR as sidechain input. getting bleed through... how to fix?

[big bustle]

i'm using an lm386 to drive a vtl5c3 to modify existing circuits. like using an iphone with a simple sequencer as an input to vary repeat time on a delay pedal. so far it's pretty awesome. except i'm getting audio bleed through.

how can i isolate the 386 from the rest of the circuit so that i don't actually hear the input audio.

something tells me this question could be answered if i understood amp design a little better.

[petemoore]

i'm using an lm386 to drive a vtl5c3 to modify existing circuits.
 Doesn't reveal what the 386's input is, or anything much about the actual circuits, somewhere in this mist is a noise issue, clearing the fog away with a schematic or diagram or more clearly described circuit would fascilitate the ability to provide better solutions or likely sources of the problem.
  like using an iphone with a simple sequencer as an input to vary repeat time on a delay pedal.
 I like the concept...
 so far it's pretty awesome.
 I bet !
 except i'm getting audio bleed through.
 Without this you'll hear no audio signal...in other words I'm not sure I've read the meaning of this wording as it was intended.
 how can i isolate the 386 from the rest of the circuit so that i don't actually hear the input audio.
  Again I don't understand...the source' is another word for audio input, audio input is generally where the phonograph or guitar is plugged in so the amp can amp it or whatever, then the speaker makes it so we Can hear it.
 something tells me this question could be answered if i understood amp design a little better.
 probablee...
 There are various ways noises [identifying the source by way of type of noise is one thing, elimination process where applicable is another excellent sure fire way to at least locate it]. Ie where stuff 'can' be signal = ground or unplugged, it is simple matter to locate by process of elimination. That works great...unplug or zero the guitar volume for a way to eliminate...better yet if the cable is taken out of the next thing...you get the idea...by process of increasing elimination, noise source can sometimes be located.
 Runs into trouble, ie replacing a PS with a battery is super simple way of applying a floating "Super DC'' [no ripple] power supply, and disconnecting [ie why we want floating] all grounds. If battery improves the noise rejection, it could be the battery is a no-noise type of DC, or it could be that it doesn't require a ground connection other than the signal ground.
 Throwing a BFC as RG calls it, large value capacitence across the power rails, and in the ccase of 386 which can draw large 'gobbles' of current, having say V+ through a 100ohm then BFC between the +/- rails, cheap try and one other elimination process that can be experimented with [ie put the BFC on is cheapeasy...can just stay there.
 LFO's are oscillators by name and nature, oscillator means what it means, this one may be connected to a chip that draws large currents and hence induces [like signal transformer] or injects [by way through powr supply lines] 'noise' into the audio signal path.  

[brett]

Hi
first thing that comes to mind is that the input impedance of the 386 is low.  To isolate the input from other circuits, you'll want a JFET buffer (or similar) in front.
Next most likely is transmission via the power supply line.  Low impedance, high capacitance between the line and ground is the solution.  This implies using both film caps (e.g. 0.1uF) and electro caps (100uF).  They must be as close as possible to the 386.
If you've got those issues covered we might consider less obvious possibilities (capacitive coupling, etc).
cheers