Unbalanced output to balanced input?

[Req]

Hello, I've got quite a live looping setup in the works (MPC Key 61, seaboard rise 2, 61-key virus TI2, osmose, microfreak stellar, orba 2, TR-8S, TS7, push 3 standalone, MRCC/XpandR 4x1, RME multiface 2/digiface, 13900KS and various microphone stuff on 2 full K&M stands) but I've got 1 lingering problem. I need to run 5.1 from a Sound Blaster AE-9 (2x RCA, 2x stereo TRS) into 6 of the balanced TRS inputs on the Multiface. These pick up a huge amount of noise, especially with all 6 unmuted. It doesn't just sound like a ground loop, I can hear various patterns which I assume are coming from the many many devices those cables are running by. Unfortunately, the RME audio interfaces will not decode the 5.1 signal over SPDIF so I can't use the digital output on the SB AE-9 for 5.1 that way.

Is there anything I can put in-line, any kind of modification I can make to the SB AE-9 or the cables, etc to achieve a low noise floor on those inputs?

Alternately, does anybody know of a product that can decode 5.1 SPDIF to 6 ADAT channels with "no" latency?
Or, does anybody know of a product like this one, but with balanced outputs?

Amazon.com

I picked up a used Anthem AVM30 to see if it could give me the 6 balanced outputs, but I just couldn't get it to happen no matter how I configured it or the SB AE-9.

Thanks much!

Edit to add that I'm currently using:
2x "Mogami Gold TS-RCA 6FT 1/4" TS to RCA Male"
2x "Mogami Gold 10FT 3.5mm to Dual 1/4" TS"

 

[KSTR]

Is there anything I can put in-line, any kind of modification I can make to the SB AE-9 or the cables, etc to achieve a low noise floor on those inputs?

Yes, basically all you need is balanced cabling to the RME to make use of its balanced inputs.

However, this will mean custom cables in your case, notably for the 2x TRS stereo outputs as I'm not aware of any makers of ready-made cables for this specific scenario. Therefore, you'll probably need to make them yourself or find someone with soldering skill. You might get away with a bunch of (stacked) adapters but those just as costly as custom cables and may have issues.

The key point for any cable that converts unbalanced to balanced is that this conversion to balanced must happen directly at the source end.
The conversion itself is actually simple:
You have the hot conductor (like center pin for RCA, or T or R for TRS) and the shield conductor (RCA shell). The balanced cable has three conductors: hot(positive, XLR Pin2), cold(negative, XLR Pin3) and shield(XLR Pin1).
Hot goes to hot, obviously, and cold and shield both go to the shield of the unbalanced source connector.

I'd just by some moderate lengths of off-the-shelf TRS to TRS cables, cut them in half and solder the appropriate connectors to the open ends. Keeping cable lengths short is always a good idea, though.

 

[pma]

Generally speaking, there may be issues with balanced/ unbalanced conversion using cable with cold out connected to com (gnd). This depends on source circuitry. True, floating balanced output with transformers or circuit like DRV134 may have cold out connected with ground and hot will just move up. But, the consumer audio, almost everything tested at ASR, has simplistic output with 2 opamps, cold out just inverted and both hot and cold related to ground, fixed, not floating. Connecting cold to gnd equals to short circuit on cold, then, with elevated current injection to ground and rise of distortion.

I use INA217 or OPA1622 to convert balanced to unbalanced. Or a transformer.

 

[Zek]

@pma You're missing the point - it's not about connecting a balanced output to an unbalanced input, but vice versa - connecting an unbalanced output to a balanced input.

 

[Randyman...]

Generally speaking, there may be issues with balanced/ unbalanced conversion using cable with cold out connected to com (gnd). This depends on source circuitry. True, floating balanced output with transformers or circuit like DRV134 may have cold out connected with ground and hot will just move up. But, the consumer audio, almost everything tested at ASR, has simplistic output with 2 opamps, cold out just inverted and both hot and cold related to ground, fixed, not floating. Connecting cold to gnd equals to short circuit on cold, then, with elevated current injection to ground and rise of distortion.

I use INA217 or OPA1622 to convert balanced to unbalanced. Or a transformer.

He's going SoundBlaster unbalanced out to RME's electronically balanced inputs, so no worries about shorting cold/inverted out to ground. KSTR's suggestions will likely help - but you will not have any true galvanic isolation (PC's audio ground will still be tied to the RME ground). Floating the shields (sleeve) at the RME inputs will likely help (no reason they must be connected AFAIK). Basically driving a single-ended unbalanced line into balanced hot and cold - no shield connection at the TRS input. But the shoeld would still terminate at the unbalanced end to shunt noise there.

PS - I have a slew of Digifaces and Multifaces, along with a "big boy" MADIface-XT and two MADI HDSPe's and ADI-648 in my studio. RME is the bomb!!! Some of that gear is approaching 15 20 years old (time flies) and still works with modern OS's w/o a problem...

 

[Randyman...]

And just to mention it - the RME MF/DF can accept a DD/DTS SPDIF bitstream and pass it, it just can't decode it natively. If your "catch" machine with the Multiface can do software decoding and pass it back to Totalmix/ASIO - that might be an option...

 

[pma]

OK, sorry. Unbalanced to balanced is as simple as it was described, though not optimal. At least a balancing resistor with value of source output impedance should be inserted between cold and gnd, otherwise CMR is compromised.

 

[Randyman...]

OK, sorry. Unbalanced to balanced is as simple as it was described, though not optimal. At least a balancing resistor with value of source output impedance should be inserted between cold and gnd, otherwise CMR is compromised.

I'd assume most of the noise would be from ground contamination from the SB card - and not be reduced by CMRR of the balanced receiver? Lifting the shield/sleeve at the RME end and having the unbalanced "cold/shield" come into the Cold leg of the balanced receiver will at least get away from a 0 Ohm ground connection between the two, and lessen that ground-loop potential.

 

[pma]

I'd assume most of the noise would be from ground contamination from the SB card - and not be reduced by CMRR of the balanced receiver?

Yes but to suppress it you need to balance the sense voltage divider (sensing the buzz voltage on the shield ground wire).

Zsource/Zinput+ input divider to be balanced by that inserted resistor in the cold path, otherwise CMR is compromised. Calculate it, simulate it, measure it but please do not deny the laws of electrical circuits .

 

[KSTR]

Floating the shields (sleeve) at the RME inputs will likely help (no reason they must be connected AFAIK).

Normally, shield should be connected at both ends.
The idea is to let the disturbing current still have a path to flow, keeping the ground potentials at the sender and receiver roughly the same because they are "shorted together". This avoids taxing the common-mode rejection of the receiver which will be compromised when the source impedances on hot and cold leg not being exactly the same, as mentioned by @pma. The drawback is that the noise current is still flowing and it might be excessively high when both source and receiver have their grounds connected to mains earth, directly or via other devices in the setup.

In the end there will be always trade-offs and sometimes only experimentation will show what works best with a given amount of effort.
A ground lift might be beneficial but I would suggest doing at the source side, and I'd also try to match the source impedances (assuming known output impedance specs). There also is the variant of a partial ground lift, in the "lift" position the shield is partially connected (notably for radio frequencies) via a capacitor and paralleled resistor.

I do understand that it is an impertinence to deal with all that at the "soldering level" yourself and one has to wonder why the cable industry isn't picking up this terrain with a great variety of objective no-bullshit "problem solvers" cables -- made to order, for example -- paired with a competent user assistance which is paramount to avoid endless cycles of frustration.

 

[richardm]

I do understand that it is an impertinence to deal with all that at the "soldering level" yourself and one has to wonder why the cable industry isn't picking up this terrain with a great variety of objective no-bullshit "problem solvers" cables -- made to order, for example -- paired with a competent user assistance which is paramount to avoid endless cycles of frustration.

Hence my "FAFO box" idea. Perhaps something with a rotary knob to [hopefully] land on a wiring combination that works. Add a rheostat or two.

My unbalanced PC to balanced pro gear woes were fixed by sending sleeve into both tips, ring into right channel ring, tip into left channel ring, and leaving the cable shields connected to the sleeves on the balanced side.