Ground loops inside a single stompbox

[wcampagner]

Hello,

I live near a Radio Station and i've built a stompbox...
The stompbox is inside a metal box and the ground of the circuit is attached to the metal chassis... i've tested the chassis grounding and it is working...

I have a few problems... the first one is the bleed-through effect... i can hear some effect signal when the stompbox is in bypass mode... and the second problem is RF interference from the radio station...

I'm thinking it is a ground loop problem... because when i touch the strings of the guitar, the RF interference gets louder... or when i touch the stompboxes switches too... it seems that the stompbox is "processing" a little of the ground signal...

I've already tried some filters in the input... without success... i'm suspecting the problem is a bad layout of the signals VCC, GND and 1/2VCC... i've seen that my original Samsamp GT-2 have some kind of "star grounding" on the power capcitors... and the GT-2 doesn't have the bleed-through nor the RF interference problem in my house... it have also an area full of copper squares... maybe it is used for shielding some part of the circuit... i don't know...

I want to know if someone has faced it before... or if someone can help me solve this problem...
If the problem is really a ground loop problem due to a bad board layout... how can i solve it? Is there any guidelines to identify ground loops?

For information i'm trying to do an original tubescreamer... i'm using electronic bypass with CD4013 flip-flop...

Thanks a lot in advance,
Wagner.

[KMS]

your RF interference, how do you know it is Radio.  What exactly does it sound like?

Chassis grounding can stop or reduce AC hum, but the chassis as a grounding bus terminal is not good.  All your grounds should be isolated and then try grounding the chassis. Try no ground on the chassis and see if it helps because your body serves as additional ground to the chassis, which is making it worse, so the direct approach would be decreasing the ground on the chassis.

Are you using true bypass?  I expect a no for that since you mentioned an IC in the bypass.  Some part of your bypass is not fully choked when in by pass or you might have a wiring problem.

Are you using a battery for your PS ?   Wall wart with more than one effect on it?

I'm not sure I have the answers but folks will need to know a little more about this before we can help find the problem. My guess is that everything your doing needs to be up here for us to read. The fact that the noise gets worse with your body in contact with the box """"suggest""""  a ground problem, but the ground problem could be a PS problem, routing, a slight short to ground like a pot with solder blob shorting one of the leads to the pot chassis, lead dress etc or a combination problem. 

[niftydog]

what sort of noise are you getting? RFI and ground loop hum are very different sounds, so a description might help us narrow it down a bit.

As for the touching the strings thing, if it is RFI then all you're doing is acting like a big, squishy antenna and feeding RFI in via the ground path. There's not a whole lot you can do about that aside from using top quality cables, well sheilded guitars and stompboxes and perhaps using ferrite beads at the inputs to your stompboxes.

[wcampagner]

Thanks NiftyDog and KMS for the answers,

I'll try to explain a little more to see if we can find the problem...

About the grounding... i'm not grounding everything to the metal chassis... all the grounding path is in the PCB board and than there is only one point that grounds to the chassis...

I'm not using true bypass... i'm using electronic bypass with FET's for switching... so there is also digital components in the pedal... i read that is a good practise to separate the digital and analog grounds... i'll test it to see if it changes some thing... i also read that is also good to use small capacitors (0.047u for example), between the ground and VCC of every digital IC to minimize ground changes... i'll also try it too...

I've tested the pedal with batery and with PS too... both showed the same problem... i think it is not the power supply because, as i said, i used my Sansamp GT-2 with the same PS and i didn't have this interference...

I'm sure it is RF noise i'm hearing... because i can hear the radio station... it is the radio station that is near my house...

I'm suspecting of ground loops because when i touch the guitar string or when i touch the pedal actuators the RF interference gets worse... so i think somewhere in the pedal the ground is not right designed... and the actuators don't even touch any part of the PCB Board... they touch only the plastic pushbutton part... but they are grounded by the chassis...

And about the bleed-through problem? Isn't it related to ground loops too? i think it is...

Maybe i can solve both problems with a good traces design...

Is there any link or any information i can read about it? Any advices?

Thanks again in advance,
Wagner.

[R.G.]

I'm suspecting of ground loops because when i touch the guitar string or when i touch the pedal actuators the RF interference gets worse... so i think somewhere in the pedal the ground is not right designed... and the actuators don't even touch any part of the PCB Board... they touch only the plastic pushbutton part... but they are grounded by the chassis...

It's really quite difficult to exclude RFI when you are near a strong radio source. I'm guessing that you may have nothing any more wrong than most DIY stompboxes in your grounding, or only a trivial error. The fact that it gets worse when you touch the grounded actuators or guitar string only means that your body is acting as an antenna and conducting even more RF into the box.

The only really good way to prevent this kind of RFI is to make the circuit unresponsive to anything much above audio frequencies. This usually involves blocking the RF at the input to the circuit and using RF bypasses on the power supply, as well as possibly reworking a few places inside the circuit to cut RF response.

RF sensitivity is one of the unfortunate side effects of the generally excellent state of transistors and ICs today. When I first started messing with them, you had to worry about getting transistors that would go high enough in frequency. Now you almost can't buy a transistor that will not respond over 100MHz.

These may help, may not.
- Make sure that your signal ground is connected to your metal enclosure by one and only one contact.
- It sometimes helps if only one of the two jacks (in and out) is grounded to the box, the other being a plastic bushing type.
- Try a ferrite bead or 1K resistor in the incoming signal lead.
- Add a 0.01uF capacitor across your power supply where it comes onto your circuit board. This cap must be a monolithic ceramic or ceramic disk to be effective.

I wish I could offer you a definitive cure, but I can't. The answers are too application-specific.

[wcampagner]

Hello R.G.,

Thanks for your answer... i'll explain a little more to see if i can solve this problem... i can't even sleep anymore just think about solutions for it...

The signal ground is connected to your metal enclosure by one and only one contact.
I'm using normaled plastic jack's so... so none of the jack's are grounded to the box... they are grounded in the PCB.
I tried the 0.01uF capacitor across the power supply.
I tried also de 1k resistor in the incoming signal lead.

The only one i didn't tried was the ferrite bead... but i think i can find it here somewher and test.

All this tests and the problem is still there...

One test i did was to just tu plug the effect on the amp. without nothing on the Incoming... so by the normaled jack the incoming is automaticaly grounded... this way i didn't get any interference... i don't know if this can help...

What i'll try to do is another PCB layout... i thought the worst part of the project was to make it work on the breadboard... and then just put it on the Computer to auto route... but i was wrong... i started reading about PCB layout and i find a lot of "rules" to design audio systems like:

- separate digital and analog grounds and Vcc's (as i notice, my switching is very "Poping" when amp. volume is high.... i think it is happening because all grounds and Vcc's are mixed analog and digital).
- use small caps between grounds and Vcc's on the digital IC... it helps minimize sudden ground changes caused by digital IC's....
- make all grounds, Vcc's and Vcc's/2 possible like "star ground" style... the star should be at the electrolicts capacitors near the Power Supply pin....

Do you think this could help solve the problem?? Because i'll waste a lot of time to make a good PCB design...

Thanks again in advance,
Wagner.

[Paul Perry (Frostwave)]

If all else fails, and you have only ONE radio frequency to worry about, you can do what the old crystal set guys did to cut interference from a strong local radio, and that is to use a "wave trap".
The idea here, is to put a tuned circuit across the input. The trick is, to use a SERIES tuned circuit (that is, an inductance and capacitor in series), because that presents a very small impedance at radio frequencies.

It's going to be a hassle getting the tuning right, though! What frequency is the radio??

[wcampagner]

Hello,

The radio frequency is AM... it is 1,100kHz.

I think i'll try first the new PCB layout... because i think it'll solve the problem... if it does not solve i'll post here again.

Thanks,
Wagner.