hiss with LT1054 charge pump

[Dimitree]

Hi guys
I've made a little opamp circuit, and I'm powering it with stable 9V DC. Internally, 9V DC goes to a LT1054, arranged, like on the datasheet, to get a negative voltage doubler, so theoretically +18V and -18V, but obviosly there is the loss. So, when I connect the circuit, I get +16.4V on the positive rail, and -15.0V on the negative rail. It's fine for me, but, unfortunately there's a high pitch noise, I believe this is the infamous whine. I tried to add a 10pf capacitor between pin 2 and 7 to boost the switching frequency, but nothing noticeable happened, the noise is still there.
When I remove the power supply and use the pedal in bypass, it is fine, but then when I plug in the power supply the whine comes, especially when it is in bypass (but maybe because there's lot of gain when the effect is on, so it is less noticeable). So the fact that the whine is there even in bypass, could give me a hint about the problem? maybe it is not the LT1054 circuit? what could it be?

many thanks
Dimitri

[Dimitree]

I tried to remove the LT1054 when I play in bypass and the hiss disappear, so I would say it is that part 
what could I do? where could it be the problem? I tried the charge pump circuit on breadboard before doing the pcb..it worked perfect with no hiss. Now the negative voltage is sagging a lot (not the first time I connected, I was reading the same voltages as on breadboard), and the hiss is really audible.

[liquids]

What circuit?  I've never really had this issue in that regard.  I'd still be assuming it is probably related to wiring and the audio portion of the circuit.  Be clean with your wiring if it's a high gain circuit....that or you have hearing up to 25khz.   

Also, when you do a charge pump, use the biggest caps you can manage all over the place.    The 10uF on the data sheet can be increased.  And use schottky diodes.  With my 1 spot (9.4v) just early this week I was measuring 18.6 on the positive and -18.5 without load, and not much different under load.  Most op amps (etc) max out at +/- 18v anyway.

[Processaurus]

-edit: didn't read too well.

Is the LT1054 star grounded back at the power supply, and are you avoiding the stereo jack input battery turn on trick?  I've learned grounding is important with those chips.

[Dimitree]

all the grounds in my circuit go to jack ground and power supply ground (both isolated from chassis) and then from the pcb there's a cable going to chassis.
no battery mate!

[TELEFUNKON]

Your stable 9v DC power supply happens to be a switchmode unit?

Also, when you do a charge pump, use the biggest caps you can manage all over the place.

A couple of smaller caps here and there can work wonders with high frequency oscillations! (.1uF ceramic disks)

[Dimitree]

I tried the same LT1054 power supply circuit, and with the same DC power supply I use, on a breaboard before doing the pcb, and it worked perfect, no hiss. In this other project is giving me hiss 

[ibodog]

Recently I converted a couple of fuzzes to work with a similar charge pump circuit but with MAX1044.  The idea was to plug them into a Pedal Power or OneSpot with just the normal power cables that come with those supplies.  Each pedal alone was fine on the charge pump circuit.  But when I connected them both to the same OneSpot I began to get a high frequency whine.  I think that the (inaudable) higher frequency whine is there for both pedals, but probably each very high frequency whine is a slightly different frequency.  So I get to hear a lower sideband modulation artifact in the audible band.  Sound like a plausible theory?  If I separate them on the Pedal Power then there is no similar whine.

I, too, have a buffer I was building that has the LT1054 providing ~+/-9V to the opamp and it is hissier than I'd expected.  I have not yet gone back to troubleshoot...