Oscilloscope noise

[Pdavis68]

I was testing this breadboarded amp (tube) yesterday and I had the two channels of my scope at two different points in the signal chain. The signal on the scope was a mess unless I wrapped my hand around the probe wires, in which case it would clean right up. Both probes have ground clips attached and grounded What's going on?

[AudioMan]

Perhaps oscillations in your circuit, or feedback?

[Pdavis68]

Why would wrapping my hand around the probe wires clean up oscillation in the circuit it's probing?

[DrGonz78]

I was testing this breadboarded amp (tube) yesterday and I had the two channels of my scope at two different points in the signal chain. The signal on the scope was a mess unless I wrapped my hand around the probe wires, in which case it would clean right up. Both probes have ground clips attached and grounded What's going on?

Sounds a bit like stray capacitance.

[Pdavis68]

In the circuit or the scope? The noise is in both lines in the scope.

What would the fix be (besides holding the probe wires while I'm busy probing and dodging high voltage)?

[AudioMan]

Are you using 10X probes?
The 'scope cable/probe capacitance can be just enough to cause your circuit to burst into oscillation in high impedance circuitry.
The solution is not easy to discern at this point because it may be the way your circuit is laid out, and whether there are decoupling components for even order stages.
You did say it was a "breadboard" circuit that could have enough stray capacitance between input and output.

[Enzo]

Try using one ground clip, instead of both. Once the scope is referenced to ground, the other added clip can cause ground loops in the test setup.

Also, this is a breadboard, so every wire in it is a big antenna.

[DrGonz78]

Not sure this one was answered? Does it behave this same way when you only scoped with one single probe at a time? Are there two ground points leading to the scope or just one, as Enzo pointed out?

Edit: Ignore my underlined question, obviously, since I re-read your original post and you already said so.

[Pdavis68]

Yes, AudioMan, 10x probes. I did not realize that. It's not a huge issue, but it does make it difficult to test the circuit. You may be right about it inducing oscillation since I did notice that with the probes hooked in, the voltages seemed to go up quite a bit in a few places if I didn't have the volume and gain zeroed.

Thanks, Enzo. I didn't think about ground loops and it didn't occur to me to try just one probe at a time. I was trying to see post 1st stage & secondary side of the OT at the same time. And they were grounded on opposite sides of the breadboard.

[AudioMan]

Another trick that has been used, was to add a 100K resistor at the end of your probe and touch the circuit with the loose end of the 100K. This was to isolate the shunt capacitance somewhat. It does create a bit of error reading the voltage levels, but may get around your "probe induced" oscillations - if that is what we are dealing with.

[gearhead63]

Does your scope have a bandwidth limit switch? If it does then use it, or you'll be picking stray capacitance from your lights, other circuits in the room etc.

[loudthud]

Try this: Find yourself a big ferrite donut or snap on EMI bead. Loop the scope probe lead through it three or four times near the scope end.

[Mick Bailey]

Scope or meter leads can do odd things to a circuit. Even a DMM can cause oscillation. Some small EL84 Fender amps can hum badly and draw excessive current just off measuring voltages at the tube socket due to oscillation. They're borderline to begin with and don't take much to tip them over.

Years ago I was a Prime computer systems engineer and part of the engineer's kit was a set of extender boards (to enable PCBs to be pulled out of the backplane to work on) and a really decent 'scope. I got my kit before the first training course and the first thing the tutor said was to dump the scope and extenders because the capacitance upset the timing so much they were useless.

[nickb]

Try this: Find yourself a big ferrite donut or snap on EMI bead. Loop the scope probe lead through it three or four times near the scope end.

^^^^ Wot he said.

If this is the cure, the problem is poor common mode rejection in your 'scope - a very frequent issue. What this means is that is the same signal is present on the shield and center of the scope input the 'scope should measure zero. In practice it will always show something and it's ability to reject the common signal is called 'common mode rejection'. The common mode rejection will vary with frequency, getting worse with higher frequencies is the norm. That you see this effect is an indication that unwanted HF is present - probably from instability.

When you wrap your hand around the lead, you are capacatively coupling the shield to earth via your body. That capacitance attenuates the noise that is in the shield and helps the 'scope out.

The instability in your breadboard is something we already talked about in your other thread on the breadboard.

[Pdavis68]

I don't currently have an amp on the breadboard, but I'm going to start putting one together this week, so I'll be back to the oscope stuff soon. It definitely seemed better only using one probe, so I suspect having both probes hooked up, but grounded in different places was the actual issue. I'll do more testing when I get this next amp on the board. With Thanksgiving this week, though, it may be weekend after next before I have something I can test.