RF and noise issues in opamp design

[ricothetroll]

Hi,

I designed and built this a few months ago :
http://img855.imageshack.us/img855/2692/livelooperinterfacesche.png
http://img35.imageshack.us/img35/5262/livelooperinterfacelayo.png

It's a balanced FX loop with the return mixed to the original signal. I use it with my band to use Ableton Live as a synchronised, multitrack looping station.

The schematic is pretty classic. The 3rd entry of the summer UC1B adds the target's (amp) ground to the output signal in order to avoid hum : the internal ground is canceled by the differential amplifier and the output signal is referenced to the target's ground.

It works pretty good, we've been using it for a while now, but I'd like to improve it to make it more silent. Indeed, it has a significant noise contribution, that can be heard in high volume (rehearsal or live) conditions, or when it is plugged in the a drive channel of an amp. The noise I'm talking about is electronic noise (hiss), mixed with some RF interference (I can hear "La Première" radio station quite good when the volume is high enough )

I guess the RF interference doesn't come from the input, as it is filtered for that purpose by R3/C23. Unplugging the input jack doesn't reduce it either. I also tried different values for C3/C11 to lower the cutoff frequency of IC1B, without any success. I tried to add a bigger output series resistance by increasing the value of R19, without success. The only thing I did that reduced the RF interference was adding a cap between "amp out" and the amp's ground, but it wasn't very significant without altering the audio high frequency response.

For the electronic noise, I don't know what to do either. The values I used for the series resistors to IC1B (R1, R2 and R7) are already pretty small, and the values for the feedback network seem reasonable to me.

Any help would be appreciated ! I must admit I've come to the limit of my (small) knowledges concerning those issues.

Best regards.

Eric

[merlinb]

Looks like most of the hiss in that design is coming from R3. I would drop that down to 1k, and increase C23 if necessary. You could afford to make R5/6/9 smaller too, though I don't think they're the problem.

[R.G.]

In my experience, any time you hear a hiss and anything which might be RF, you need to check for RF oscillation in the circuit. Internal RF oscillation sometimes sounds like a subjectively angrier sounding hiss than thermal hiss. Modern opamp are pretty tolerant, but if the power supplies and grounds to them don't make them happy, they will go off and count their toes at units of megahertz. You'll probably want to either get the power and grounding pristine and/or get a good scope on it to see if it's singing into the ether.

[ricothetroll]

Thanx a lot for your answers !
I'll check those things out. In case I would have RF oscillations, what could break it ? IC1B is already filtered, and lowering the hi-cut frequency doesn't help... Hiss does sound pretty angry !
Best regards.
Eric

[R.G.]

What stops RF oscillation - unfortunately - depends on what is causing it. Common causes are poor power supply and ground wiring and routing, or wiring issues letting positive feedback happen through the small capacitances between wires. 

It used to be de rigeur to put a 0.1 or 0.01 ceramic cap from the power supplies to ground on every opamp package to keep the impedance and inductive nature of the power wiring at bay. Hardly anyone does that today, but it can still help sometimes. My boss is always wondering why I put all those ceramic caps across the power supply in my layouts. 

You could try soldering a 0.1uF ceramic (not film) from the  (+) to (-) pins on your opamps. If it's power supply related that will usually either fix or change what's happening. It's important that these caps be as physically close to the IC pins as you can reasonably get them. Even long lead wires are a problem.

Grounding and signal routing issues can let it pick up external RF and amplify it. I think you mentioned radio pickup. Is it inside a metal box?

[ricothetroll]

Hi R.G.,

Thanx a lot for your help !

I tried what you told me with the ceramic cap, and I finally could get rid of the radio by putting it between one of the supply pin of the opamp and the target's ground (on the output jack, to the amp). Putting it between both supply pins of the opamp, and between the supply pins and the internal ground had no effect, though the internal ground has a compliance of 100R-0.1u to the target's ground. Then, do you have an idea of where the problem cames from ? That's just for my personal culture 

Concerning the hiss, there's no improvement. I tried to shunt the series resistor by replacing it with a wire without any effect. I also replaced the gain network of IC1B, originally 33k/16k5 with 3k3/1k65, but no effect neither.

Anyway, I'm already pleased by having the radio issue solved !

Any idea concerning the hiss is still welcome... 

Best regards.

Eric