HOWEVER … this (and all available cables) are intended to pass signals in the other direction - to split a Stereo unbalanced signal into two balanced outputs. That means that, on both the XLR connectors, both pin 1 (shield/ground) and 3 (negative signal) are connected to the Sleeve of the TRS connectors, as shown in the circuit diagram (from ProX XC-SYXF05 5FT 1/4" TRS-M Stereo Y Cable | IDJNOW).
I’m not sure what happens in this scenario, but isn’t the sleeve of the TRS connector being driven with the negative (pin 3) of both left and right signals? Which would cause … I’m not sure what at the receiving end.
So I tried it by routing the output ¼” TRS stereo signal into a stereo channel of a Mackie Mix8 mixer (using Y splitter) and listening on headpones. What I heard was … perfectly normal. Left and right were separated and seemed normal (although I had no way to tell if the whole signal was phase-inverted).
Can anyone explain why the cable as-is works?
Should I cut the strap between pins 1 and 3 and will it make any difference??
The difference between the two methods is transformer/cross-coupled or not which is dependent on the RME output. If I didn’t know, I’d try option 3 first.
Essentially, when you connect pins 1 and 3 on a symmetrical XLR output, you “pull down” the pin3 output to ground, so the negative part of the output simply gets discarded. All you are left with is the “positive” part of the signal against ground, so you’ll have an output signal that is 6 dB lower than the symmetrical output, but otherwise the same.
So your experiencing “perfectly normal” output is to be expected, except the signal through your connection cabling should be 6db lower than connecting your XLR output directly to the mixer. And of course the “interference rejection” capabilities of the symmetrical setup are gone, since the signal is measured against ground and not against negative…
With a decent output converter, shorting a pin to ground shouldn’t hurt it - there will be an output resistor in the circuit that avoids anything nasty happening from shorting pin 3 against ground.
I am not sure what happens when you disconnect the pin1-pin3 connection - ideally, things should work out nicely, but depending on the specific output / input circuits, the positive signal might be losing its “zero” reference, if the shield isn’t connected to ground on one side. So I don’t see a reason to start fiddling with this cable if it is currently working.
Quick addendum: I just checked the Babyface manual - RME recommend to NOT connect output pins 1 and 3 on outputs. Due to the implementation (low-resistance, non-servo-balanced) of their output stage, pulling pin 3 to ground will increase distortion levels and power consumption. Having pins 1 and 3 connected won’t hurt your interface (they are short-proof, according to RME), but it’s better to not have them connected.
Makes sense to cut the jumper. I also have a concern about connecting XLR Pin3 from different channels when wiring to the TRS (sleeve). Cutting the jumper between Pin1 and Pin3 breaks the connections. I have no concerns about connecting Pin1 from different channels since those pins should be chassis ground unless the ground has been lifted.
There should be no change to the levels, and I doubt that you would be able to hear the quality impact - audio quality of the Babyface is such that the THD (harmonic distortion) is super-low. An increase at that level should only be audible to the vampires among us…
I set up a white noise file and set the levels so the mixer (a Mackie Mix8 with only -20, 0. +6 and OVL LEDs) to flicker right at 0dB.
I then cut the pin1-to-pin3 strap on both the XLR connectors and restored the same setup. Here’s what happened:
The levels rose substantially. Both channels showed an increase of about +4dB (figured from interpolating between the 0 db and +6 db LEDs with some careful knob twisting).
When switching back to the original recording I was using (a Jazz backing track with separated bass, drums, and acoustic electric guitar), the channels now seem much more separated. I think with the straps in place, I may have (somehow) been hearing a portion (half volume?) of the left channel in the right side and a portion of the right channel in the left side. The separation is now (without the straps) more dramatic.
There is now a massive buzz on both channels - somewhere around -30 dB. The buzz is independent of the source volume, and does not change when I move the cable around (i.e. closer to or further from potential interference …
This is all really confusing. I love the complete and authoritative nature of the Rane documents, but they only seem to cover the more common situations. I’m wondering if there is (or should be) a carefully curated reference for some of the less common situations - it might be immensely valuable to the community. This scenario (converting stereo balanced to stereo unbalanced with a common sleeve), cannot be all that uncommon …
Just checked the cable, which turns out not to be wired like the diagram (ugh). When I had cut the straps, that tied pin 3 (negative signal) rather than pin 1 on each of the XLR connectors to the TRS sleeve.
I will re-solder (this could get messy) so only pin 1 (Shield / ground) on both XLR is connected to the TRS Sleeve.
Cable is now correctly wired (only burned one finger). Results are:
Signal level (using the White Noise test) is now exactly what it was with the strap in place.
Sound is now pristine, properly separated.
The Buzz is Gone! (… “The Buzz is Gone, You know you done me wrong baby, and you’ll be sorry some day …”)
Interesting what happens when you replace ground with the negative signal … I’ll be an experienced engineer could immediately identify the issue having heard it once (the hyper-separation is particularly wierd).
THANKS for the assistance … I would never have gotten here without the assistance of y’all …
My thoughts… Original wiring was pins 1-3 shorted but the wire was going to pin3. When ground jumper was cut to Pin1, the output (pin3 and a floating pin1) became ungrounded and a ground loop developed (Buzz). Also, the original shorting of Pins 1-3 was eliminating 1/2 of the balanced signal, so it was already at the -6dB. When Pin3 was cut, the output ground became floating with the other half of the XLR, so the signal level increased as the wiring was converted back to balanced by ungrounding Pin3. Could have started clipping, too, with the additional 6dB.
- without proper grounding, there’s a lot in limbo…