Studio Grounding Question

[Bo Deadly]

Maybe check the wall socket to make sure neutral and earth are not accidentally reversed. Not sure if that would explain the problem but it couldn't be good.

 

[Gold]

This is OT but somewhat related. I've been working on a Barry Porter EQ. I ran coaxial cable for all the rotary switch connections. I have all the shields for each switch tied together into a ferrule. When I tighten the switch bushing the ferrule makes contact with the face plate. On the two channel version I already built this worked fine.

When I went to test the one I'm working on the positive gain worked as expected but negative gain would act up. It's some sort of oscillation but the IC's weren't getting hot. It was this way in each section. After tearing my hair out double checking everything I disconnected the rotary gain switch and used a pair 5K resistors to see what happened. That stopped the problem. My first thought was that the cables were too long from the switch to the connector. I tried the switch unmounted from the faceplate expecting it to oscillate. It didn't. Hmm. Then I tried connecting the shields to audio common instead of chassis. Bingo.

I am going to do some more digging and see if I can figure out what's happening. I'm using a remote power supply with separate 0V and Chassis runs. OV and Chassis meet in the remote PSU. My regular ohmmeter says less than a ohm between the two. The battery in my micro ohmmeter is dead but I'll check it with that. I guess it must have something to do with conductor to shield capacitance and the resistance between 0V and Chassis. Weird grounding stuff.

 

[radardoug]

Paul, one thing to watch is if you are connecting shielded cable directly to the output of an opamp, the cable capacitance can cause the opamp to oscillate. This can be cured with a 100 to 200 ohm resistor directly on the output of the opamp to isolate the cable capacitance.

 

[Gold]

These are the connections to the gain, frequency select and bandwidth rotary switches. I don’t think any of the connections are directly to an op amp output. The reason I did it was to prevent crosstalk. There are a whole lot of  cable connections stuffed in a small space. The bandwidth and frequency select are happy. I didn’t notice any gain inaccuracies or more noise so I may just not use the shields for the gain switches. Connecting the gain switch shield ferrules to audio common will be physically difficult.

 

[plumsolly]

I think it's a very important clue. Surely, the mains witch of the console PSU does not alter the earth circuit, so any ground loop is unchanged by the position on or off of the switch.
It means that turning on the PSU switch introduces significant leakage into the console ground. I would check for a defective Y capacitor. Y capacitors are between Line and Earth and between Neutral and Earth.

Thanks for your insight - that makes sense to me. According to the schematic, there are no Y capacitors in my console PS, but that doesn't mean they are not there. It is a two-person job to take the P.S. out of the rack, but I am going to try and have a look soon.

 

[plumsolly]

Maybe check the wall socket to make sure neutral and earth are not accidentally reversed. Not sure if that would explain the problem but it couldn't be good.

Makes sense - I have confirmed that it is the correct way around, though.

 

[plumsolly]

I made a video of the noise changing and reacting when other units are turned off, in case it reveals anything. Thanks again to all of you for lending me your expertise to try and solve this problem.

https://youtu.be/KsBz55ikozE

 

[scott2000]

According to the schematic, there are no Y capacitors in my console PS,

surge suppressors may have some as well....

 

[plumsolly]

surge suppressors may have some as well....

That's really interesting... I have a power conditioner in my system that might fit the bill.

 

[plumsolly]

Here is a scope shot of the noise on the positive rail in the phono preamp - It is the same on both rails. This noise is proportional to the noise on the output - when the noise on the output goes away, these rails look clean. Notice those little spikes at 120hz in addition to the more sawtoothy ripple. Those spikes are what I was seeing at the audio output.

 

[radardoug]

Is your power transformer putting out enough volts to allow for the drop across the regulators that they need?
This would be a classic problem, and as the mains varies, the regulators stop regulating. You haven't given a value for your picture, but the output from a 3 terminal regulator should be a straight line.

 

[Bo Deadly]

Maybe the mains wire at the wall is just some common 14 gauge stuff for lights and such but you should be using the 10 gauge BX stuff for air conditioners and such. Meaning you're taxing the circuit.

Also, I would not assume that the noise has gone away just because you don't see the harmonics. Make another wav like you did before with the console supply off, post it and I'll run it through my spectra.m script.

 

[Bo Deadly]

Also, how old is the console? If the caps are clapped out, the console might be drawing too much power, the transformer is struggling and it's radiating garbage backward into mains.

 

[plumsolly]

Is your power transformer putting out enough volts to allow for the drop across the regulators that they need?
This would be a classic problem, and as the mains varies, the regulators stop regulating. You haven't given a value for your picture, but the output from a 3 terminal regulator should be a straight line.

And Hot Dog - I think we have a winner ;D This particular module was set up for +/18V even though the rest of the console ran at +/-16V - I guess to gain just a little extra headroom because it's an output driver. I knew that and didn't really think anything of it. My transformer is putting out ~21V unregulated DC after the 10Ω series resistors, which is enough for the regulators to do their thing - until the console P.S. or power amp come on and start pulling enough current to drag the mains down just a little bit. Then it's right on the cusp and it is either not regulating or going in and out of regulation which I would imagine adds its own garbage into the mix. I just changed the resistors in the regulator to provide +/-16V which I think should allow for enough of a drop across the regulators in most cases - what's the conventional wisdom here - if I want 16V for example what should I feed the regulator in order to account for line sag but not make the regulators work harder than they need to?

I feel a little bit goofy for chasing down all these leads and having it turn out to be something so simple - but I learned a few things. Thanks again to everyone that chimed in.

 

[plumsolly]

Related power supply question: I have 18V secondaries and indeed, I measure ~18V RMS across them, so my DC should be about 18√2 - 2(0.7) = ~24VDC, but I am seeing more like 22VDC. What causes that?

 

[abbey road d enfer]

Related power supply question: I have 18V secondaries and indeed, I measure ~18V RMS across them, so my DC should be about 18√2 - 2(0.7) = ~24VDC, but I am seeing more like 22VDC. What causes that?

Have you checked with an oscilloscope?
The √2 factor is correct only for clean sinewaves. In contrast, square waves have identical peak and rms value. The voltage from the rectifiers is always in-between sine and square.

 

[plumsolly]

Have you checked with an oscilloscope?
The √2 factor is correct only for clean sinewaves. In contrast, square waves have identical peak and rms value. The voltage from the rectifiers is always in-between sine and square.

I have not - but that makes sense - so the rectified dc voltage will depend on how close your mains power is to an ideal sine wave?

 

[abbey road d enfer]

I have not - but that makes sense - so the rectified dc voltage will depend on how close your mains power is to an ideal sine wave?

The mains power is usually clean enough, but the interaction of various transformer losses and rectification seriously distort the secondary voltage.

 

[plumsolly]

The mains power is usually clean enough, but the interaction of various transformer losses and rectification seriously distort the secondary voltage.

That makes a lot of sense. Thank you - and happy new year

 

[Gold]

Paul, one thing to watch is if you are connecting shielded cable directly to the output of an opamp, the cable capacitance can cause the opamp to oscillate. This can be cured with a 100 to 200 ohm resistor directly on the output of the opamp to isolate the cable capacitance.

Doug, The problem turned out to be a close cousin to what you thought. This EQ will be M/S only. I am using Wayne Kirkwood's first version M/S board as a line receiver, M/S converter and line driver. It's easier to check things out in L/R. I was feeding the SVF directly after the Line Receiver on the EQ. The input impedance of the SVF is 5.1K. I was trying to drive that from a THAT 1246 with a too long unshielded wire. I shortened that wire up and shielded it. Problem solved. I didn't experience this in M/S mode. M and S get fed from THAT 1240's after the input 1246. Those drove the SVF fine even with the too long unshielded wire.

I tried raising the input impedance on the SVF to 10K but it doubled the available gain. I guess I'd have to halve the gain range set resistors. I gave up on that. Luckily shortening the cables worked.