Help with ground loop problem : synth > small mixer > tube DI

[isophase]

Hello GroupDIY,

I would like some help to solve a ground loop problem I discovered today while I was installing some equipment at my friends studio.

setup as follows:
OB6 synth > SSL Six compact mixer stereo input >> stereo main output into Apollo interface (XLRF/TRS)
In parallel, a stereo cue send output from the mixer (manual says Balanced TRS output) feeds Manley dual mono preamp direct inputs. Manley outputs feeds focusrite ADAT converter line inputs (XLRF/TRS) back into the Apollo through optical cable.

The TRS stereo cue output from the mixer is connected to the Manley DI inputs using short unbalanced instrument cables.
Everything sounds really great, but I can hear a very slight hum/buzz from the Manley DI that I’m pretty sure is a basic group loop.

i will bring an iso box to the place tomorrow to test and confirm.

I’m thinking of making special TRS/TS cable with hot and cold connected to Tip and sleeve at the other end with the shield unconnected.

Looks like it should work and to avoid the ground loop? Mabe this is the correct way to do it?
I never had to make a cable like this before.
Thank you

 

[ccaudle]

I’m thinking of making special TRS/TS cable with hot and cold connected to Tip and sleeve at the other end with the shield unconnected.

You need to look at the documentation for the mixer to see what is recommended for connecting the cue outputs to unbalanced.

There are about three ways to make balanced outputs (not counting transformers, which I assume is not relevant for most small mixers).
One way has the cold leg impedance balanced for noise cancellation, but no active signal drive. That is pretty safe to connect up whichever way is convenient.
Another way has the cold leg driven by an op-amp with no cross coupling to the hot leg, if you connect the cold leg to ground from an output like that (e.g. by using an unbalanced TS to TS cable) you risk blowing up the op-amp, or at the very least causing problems for the power supply as the op-amp tries to drive into a short circuit to ground.
The third way has cross-coupled resistors on the output op-amps to make it behave more like a transformer, but depending on the circuit specifics you may need to connect the cold leg to ground right at the output as opposed to at the end of the cable. You should just be able to leave it unconnected as well.

So without knowing the specifics of which output is in the mixer you have, the safest is TRS to TS with the TS side having no connection to the cold leg coming from the R connection (tip to tip and shield connected all the way through).

Are you doing this just because you like the sound of the Manley DI and want to compare the sound of the synth straight through the mixer to the sound of the synth through the tube DI?

Bill Whitlock has written lots of good articles on troubleshooting balanced to unbalanced connection problems, but the short of it is unbalanced connections are inherently susceptible to noise, sometimes you have to put a transformer in line to isolate noise current into an unbalanced input.
Here are a few of the white papers Bill authored when he owned Jensen:
balanced and unbalanced interconnection
hum and buzz troubleshooting guide

 

[isophase]

Are you doing this just because you like the sound of the Manley DI and want to compare the sound of the synth straight through the mixer to the sound of the synth through the tube DI?

Yes, my friend would like to record both signals (straight+tube DI) at the same time.
Both sound really good to me. But my friend likes to complicate things a lot

Another way has the cold leg driven by an op-amp with no cross coupling to the hot leg, if you connect the cold leg to ground from an output like that (e.g. by using an unbalanced TS to TS cable) you risk blowing up the op-amp, or at the very least causing problems for the power supply as the op-amp tries to drive into a short circuit to ground.

I hope thats not the case, I will check the mixer manual again and do some more tests tomorow.

but does my special cable make sens?

Thank you!

 

[Bo Deadly]

The TRS stereo cue output from the mixer is connected to the Manley DI inputs using short unbalanced instrument cables.
Everything sounds really great, but I can hear a very slight hum/buzz from the Manley DI that I’m pretty sure is a basic group loop.

I’m thinking of making special TRS/TS cable with hot and cold connected to Tip and sleeve at the other end with the shield unconnected.

What is the impedance of the DI? If it's really high like for a guitar, you're asking for hum.

Connecting cold to the DI 0V and leaving the shield open (aka ground lift) could definitely reduce hum.

 

[Bo Deadly]

Another way has the cold leg driven by an op-amp with no cross coupling to the hot leg, if you connect the cold leg to ground from an output like that (e.g. by using an unbalanced TS to TS cable) you risk blowing up the op-amp, or at the very least causing problems for the power supply as the op-amp tries to drive into a short circuit to ground.

True. Something to watch out for. But the SSL SIX is a fairly newish piece of gear. I don't think any decent modern gear would drive the cold leg. And even if they did, there would be a resistor on it so at least it wouldn't damage anything.

 

[Chrisfromthepast]

The SSL user manual says:
“All of SiX’s inputs and outputs are fully balanced (with the exception of the Phones output). This means professional equipment with balanced connections can be properly interfaced allowing longer cable lengths without noise/hum pick-up penalties and the best signal to noise performance from the whole signal chain.”

I ASSume the use of the term “fully balanced” describes an output that the marketing team is trying to differentiate from impedance balanced.
The OP could check to see if there was signal present on pin3, and that could simplify the discussion?

http://kb.inmusicbrands.com/media/images/rane/n110fig4a.pngYou’ve probably seen it, but I refer people to this guide because it usually works. Use case 8 from this page is the solution you are talking about, and that is the solution we usually use in an A/V install over short distances.
If that don’t work, a simple iso transformer would be the ticket. Jensen makes a good one that comes in an enclosure. For a synthesizer signal that can be extend well below 100hz, beware of cheapo 1:1 transformers. That sort of saturation is no good.

 

[isophase]

What is the impedance of the DI?

Manual says 300kohms

Use case 8 from this page is the solution you are talking about,

Yes that’s the cable I want to try.

No time yesterday to do some tests as the studio was busy recording some stuff..
I will try on Monday or Tuesday and report back.

thank you all for your help

 

[Bo Deadly]

Manual says 300kohms

When you make your cable, also put a 10K resistor across the hot and cold at the DI end.

 

[JohnRoberts]

"ground loops" are a catch all term describing multiple ground related noise issues.

Can you isolate the source of the "noise" to one of the several interfaces.

If you are getting hum from the DI output does shorting it's input make a difference?

JR

 

[Newmarket]

"ground loops" are a catch all term describing multiple ground related noise issues.

This ! Can't be said enough.

 

[scott2000]

This ! Can't be said enough.

A ground loops loop.

 

[Newmarket]

A ground loops loop.

Loops within Loops ! Inception !

 

[scott2000]

Loops within Loops ! Inception !

This ! Can't be said enough.

 

[isophase]

Haha! yes i know this has been discussed in looping loops here. thanks a lot for your patience group DIY I really love this place and was away for too long!

When you make your cable, also put a 10K resistor across the hot and cold at the DI end.

Ok, I get it.
Make the DI input resemble more as an line input impedance.

 

[Bo Deadly]

Make the DI input resemble more as an line input impedance.

It's not so much about what the input resembles to the source and more about the DI seeing a low impedance. A 300K input is more likely to pickup noise from the line than a 10K. However, in practice it probably won't make a difference in noise since you said the cable was super short and the source is only 100 ohm per leg. So don't bend over backwards to get the 10K in there. Just making a balanced cable with the shield disconnected will have a much more significant impact on noise.

 

[MisterCMRR]

What is the impedance of the DI? If it's really high like for a guitar, you're asking for hum.

Connecting cold to the DI 0V and leaving the shield open (aka ground lift) could definitely reduce hum.

The impedance of the line driving the DI input is likely quite low. Since the impedance of the line is source and load impedances in parallel, having a high-impedance input will have no effect on hum.
Carrying balanced high and low to the unbalanced input without a shield connection will have the same risk of possible damage to the driver as with a shield, but it the driver has voltage drive on the low side (no cross-coupling), it may make matters even worse as the high short-circuit currents find another pathway back to the driver (rather than directly back via the cable shield). This is often the source of serious crosstalk issues in systems or even inside consoles.

 

[MisterCMRR]

When you make your cable, also put a 10K resistor across the hot and cold at the DI end.

There is nothing inherently noisy about a high-impedance input ... unless you leave it unconnected to any source and it's not well shielded. But what would be the point of that?

 

[Newmarket]

This ! Can't be said enough.

Ha! I think we'll have to stop now before we are banished !

 

[Newmarket]

Just making a balanced cable with the shield disconnected will have a much more significant impact on noise.

If considering that might want to think about a ground lift network rather than simply not connecting. 100n // 100R.

 

[Bo Deadly]

The impedance of the line driving the DI input is likely quite low. Since the impedance of the line is source and load impedances in parallel, having a high-impedance input will have no effect on hum.

If you read the original post isophase said "I can hear a very slight hum/buzz from the Manley DI that I’m pretty sure is a basic group loop". So the concern is current in the shield polluting grounds. No one has suggested the problem is hum from the signal lines. So using a balanced cable with the shield disconnected at the DI could be right solution in this particular case.

Carrying balanced high and low to the unbalanced input without a shield connection will have the same risk of possible damage to the driver as with a shield, but it the driver has voltage drive on the low side (no cross-coupling), it may make matters even worse as the high short-circuit currents find another pathway back to the driver (rather than directly back via the cable shield). This is often the source of serious crosstalk issues in systems or even inside consoles.

Appreciate the expert input but this was also discussed. The source is a modern SSL small format mixer. So I think it's safe to assume that a company like SSL would not do something as dumb as use a driver that can't handle driving cold to ground. Even the cheap pro-sumer mixers don't do that anymore. The only vaguely modern piece of gear that I've found that has that, so called "tascam problem" as it is sometimes referred to, is a Livewire Solutions Active DI. I had to hack mine to be impedance balanced.