Hunting Down System Noise, Looking For Clues

[Matt C]

I've noticed some noise in my studio setup that I'm trying to hunt down and minimize, but I'm a bit baffled.  I'm hoping someone here can provide some clues to point me in the right direction. Mainly I'm having trouble even sorting out whether the noise is coming from an actually noisy output, or if it's ground noise finding its way into the system through improper wiring.

Without going into great detail, the studio consists of an old analog console, a multi-track tape machine, a computer, digital converters, and some misc. outboard gear. Obviously it's a complicated enough situation that I'm not looking for anyone here to solve it for me, but ideas or clues are welcome.

Here's what I've observed:
- On some level, the noise seems to be coming from the computer and its screen.  If they are both totally off, the noise goes away.
- There are a few different layers of noise - one hum coming from the computer screen, some HF screeching coming from the computer, and another higher hum that appears only when an audio application is active (my DAW, windows media player, etc)
- The noise is appearing at the console's tape return inputs. The noise is loudest when the tape returns are connected to the outputs of my D/A converter.  If I connect the tape returns to the tape machine instead (even if it's powered down), the noise is still there but noticeably quieter.  If the tape return is disconnected, the noise is gone.
- If I connect only the converter output cable's shield to the tape return's signal input (with no other connections), I hear the same noise as if the signal leads are connected normally.
- If I plug the computer and screen into a different power outlet the noise may increase or decrease but doesn't go away.

 

[gswan]

How is the DAC/ADC connected to the computer?

 

[ruffrecords]

How is the DAC/ADC connected to the computer?

And where does the DAC.ADC gets its power?

Cheers

Ian

 

[scott2000]

Here's what I've observed:
- On some level, the noise seems to be coming from the computer and its screen.  If they are both totally off, the noise goes away.

- If I plug the computer and screen into a different power outlet the noise may increase or decrease but doesn't go away.

When you plugged into a different outlet, was that outlet on a different circuit breaker?

 

[JohnRoberts]

Troubleshooting noise requires a methodical and logical approach...  Rearranging plugs in outlets is a little optimistic.

When you have multiple audio boxes connected in series, you want to determine where the noise is coming from (there may be multiple noise sources) so we can reduce them one at a time.

How does the noise respond to volume/gain controls...  If the noise is modulated by the volume control and goes away completely when volume is reduced the noise is coming from an earlier stage, if the noise does not change it is later.

Noise often comes from interfaces between boxes. Sometimes a different wiring approach helps, sometimes the gear has a design problem (search pin #1 problem).

Good luck but try to be systematic... or you could find yourself chasing your tail.

JR

 

[Matt C]

How is the DAC/ADC connected to the computer?

A firewire cable running from the converter to a PCIe card in the computer

And where does the DAC.ADC gets its power?

There's a daisy chain of Furman rackmount power strips that are powering everything, and ultimately get plugged into the same wall outlet. So right now the ADC/DAC is plugged into a different place in that chain than the computer, but it all leads to the same wall outlet.

When you plugged into a different outlet, was that outlet on a different circuit breaker?

I think so, but I am not 100% sure

Troubleshooting noise requires a methodical and logical approach...  Rearranging plugs in outlets is a little optimistic.

I figured as much, but as I started trying stuff out I realized I didn't know of the proper way to differentiate between a noisy signal output vs. ground noise. That's part of what prompted my post here

 

[scott2000]

I have a separate line to the outside  service panel for my gear.. I had a computer monitor causing the same problem. I took it off of my main feed and plugged it into a different outlet that was part of another circuit in the house and everything cleared up. I'm not sure why it happened as the noise never bothered me before so, I suspect this was a band aide for another issue but, it's all good so, I'm not worried yet.....

If you cut the breaker to the outlet in question, you can tell what  everything on that circuit is connected to....Like if some of your gear shuts off.....

Guess you could run an extension cord to your neighbor if you are not sure????.....

Not sure if having a separate line is the only reason it worked for me though.....

 

[nielsk]

What converters?
It is very likely a ground loop issue, it reminds me of a big Pro Tools install that was causing endless weird noise chasing until I isolated all the rack mount components (I/O, clock, rack mount computer, and so on...((all the audio cabling was already correctly single source shield terminated))  by installing isolation shoulder washers on the rack screws, all the problems went away .
Despite what many will say, if there is more than ONE ground path for each and every piece of gear, there will be ground loops. The only question is when they will cause problems...

 

[Matt C]

What converters?

MOTU 24i/o, with corresponding PCIe card (although I'm almost scared to mention it since last time I did it apparently drove someone insane!)

I guess it just comes down to disconnecting everything and see what happens as I reconnect things?  At first I was concerned about things being mis-wired, because the console's tape returns are unbalanced.  The wiring scheme still might need fixing but while briefly testing it out, no other possible wiring scheme from DAC out-->Console Tape Return solved the problem.

 

[Audio1Man]

Get an FFT Analyzer (Audio Precision,  Dscope, QA401 or other) connect it to the amplifier outputs/speakers. Run the test in repeat  mode and  methodical  go through turning things OFF, ON, UP & DOWN. Watch the spectrum for the changes.  Sometimes BEATING SIGNALS can be hard to understand, however with the FFT spectrum it can give you a clue.  I have several time where CRT MONITORS have made extra noise.

Do as John said "Troubleshooting noise requires a methodical and logical approach"

Good hunting
Duke

 

[Matt C]

Did some testing this evening and found a couple interesting bits that maybe you can help me understand:

- I took the converter physically out of the rack and unplugged all its usual analog connections (the firewire cable to the PC was still in place). This made the noise in the console disappear.  If I made any connection from the console to the converter's inputs or outputs, the noise would come back.

- Since the connection from converter output to console tape return is a balanced to unbalanced situation, I tried every permutation of reasonable wiring schemes, but it did not remove the noise.

- Isolating this signal connection (from converter to tape return) using a 10k:10k transformer worked to remove the noise.I'm assuming this means the issue is ground-noise related, but I'm still having trouble seeing exactly how.  There's a path to ground through the console's chassis (which is tied to signal 0V), and there's a path to ground through the converter outputs' shield/0V which is also tied to it's chassis.  Is it this dual path that's creating a ground loop?  Or is there some other mechanism by which the computer can be contaminating the ground?

Maybe I'm overlooking something but it seems like my only option is adding differential receivers to the console's tape returns, so it doesn't depend on the noisy ground coming from the converters?  Or can something be done with the computer itself to stop it from generating all this noise?  New power supply?

 

[gswan]

Did some testing this evening and found a couple interesting bits that maybe you can help me understand:

- I took the converter physically out of the rack and unplugged all its usual analog connections (the firewire cable to the PC was still in place). This made the noise in the console disappear.  If I made any connection from the console to the converter's inputs or outputs, the noise would come back.

- Since the connection from converter output to console tape return is a balanced to unbalanced situation, I tried every permutation of reasonable wiring schemes, but it did not remove the noise.

- Isolating this signal connection (from converter to tape return) using a 10k:10k transformer worked to remove the noise.I'm assuming this means the issue is ground-noise related, but I'm still having trouble seeing exactly how.  There's a path to ground through the console's chassis (which is tied to signal 0V), and there's a path to ground through the converter outputs' shield/0V which is also tied to it's chassis.  Is it this dual path that's creating a ground loop?  Or is there some other mechanism by which the computer can be contaminating the ground?

Maybe I'm overlooking something but it seems like my only option is adding differential receivers to the console's tape returns, so it doesn't depend on the noisy ground coming from the converters?  Or can something be done with the computer itself to stop it from generating all this noise?  New power supply?

That's what a ground loop is. When there are two paths to ground from different parts of the system where there is a different in ground potential.

A transformer coupling between them is a truly balanced and galvanically isolated system, so no ground current can flow between them.

Adding electronic balancing probably won't help because it needs a DC return path, and that is usually done using the shield connection.

 

[Matt C]

Is it reasonable to think that there's a way to kill the ground noise at the source?  Adding 48 line level transformers to my console isn't out of the question, but really doesn't seem practical.  Or do we just assume that ground noise is always gonna be there and we just find ways to protect our signals from it?

 

[gswan]

It can be done, but requires careful examination of the system grounding and wiring.
First thing is to make sure that both devices are well grounded to the same ground point (star wired).
Sometimes running a heavy ground cable from the rack to the console can also help, as you want all equipment to be well grounded to the same potential, so no stray ground currents will flow. It only takes a tiny stray current through the shield of an audio cable to end up as an audible noise.
Is your PC also well grounded (it is not a laptop is it)?

 

[Newmarket]

Adding electronic balancing probably won't help because it needs a DC return path, and that is usually done using the shield connection.

I don't see that. I think if you're using the screen for DC return current then you've got something wrong in the circuit or wiring.
The point of differential connection is to take the 'Ground'  line out of the equation. After all you can AC couple the signals and break the ground line and the differential function remains. (Althoughthere are other good reasons for keeping the screen connection intact)

 

[ruffrecords]

Have you tried lifting the grounds on the connections to/from the converter?

Cheers

Ian

 

[zamproject]

MOTU 24i/o, with corresponding PCIe card (although I'm almost scared to mention it since last time I did it apparently drove someone insane!)

I still using the same MOTU PCIe system, no shame.
Just a question, is your suspected connection (tape return) hooked at output 23-24 ?
Because the far most left hand (view from back) produce non negligible more hum than other channels in this unit,
due to proximity of IEC connector as main trafo.
Hope this help...
Best
Zam

 

[mrclunk]

Sounds like a ground loop caused by the PC, firewire cable and interface.

I take it the Motu is not powered via the firewire bus, so if you have spare cable u could try cutting the power ground on the firewire cable? Or covering the pin? (I have no idea if that ground is needed as a ref between devices!)
Do firewire  isolators exist?

I'd also defo plug your PC and motu into the same power outlet and make them the first bit of kit on the incoming  mains supply to room.  See if you can measure/hear a difference?

Also, doubt this is your problem but i often see dodgy mains plug wiring in project studios. Check all the terminals are screwed down tight.

 

[Matt C]

Have you tried lifting the grounds on the connections to/from the converter?

I did try this while connecting the converters TRS output (balanced) to the console's tape returns (unbalanced, directly driving a 10k fader). The results being:
- connecting only the TIP and leaving the shield floating (produced loudest noise)
- connecting converter shield to console chassis (hum was reduced but was still significant)
- connecting converter shield to console signal 0V (very minor reduction in hum)

First thing is to make sure that both devices are well grounded to the same ground point (star wired).

Is your PC also well grounded (it is not a laptop is it)?

I'm a little unclear on this, you mean just plugging the power cords in at the same place?  If so, yes, when testing all this out I had the computer, converters, and screen all plugged into the same power strip, and that power strip was in turn connected to the same set of outlets as the console.

It's a desktop PC with a 3-prong cord so I'm assuming well grounded, but I can't speak to any other internal details that might be problematic.

Do firewire  isolators exist?

seems like that would be ideal

 

[scott2000]

I did try this while connecting the converters TRS output (balanced) to the console's tape returns (unbalanced, directly driving a 10k fader).

How did you go from the balanced TRS converter output to the unbalanced console?

TS to TS? This would short pin3 and ground at the converter which I'm guessing is ok with the Motu....

I'm sure you already know but......

Some balanced outputs want pin 3 floating some want it shorted to ground when interfacing to unbalanced.  Depends on the circuit ......