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

[MisterCMRR]

Perhaps I misunderstood the original post, but "basic ground loops" come in lots of flavors - the most predominate being common-impedance coupling in unbalanced interfaces. If disconnecting the shield in a balanced interface reduces hum, it means there's either a classic "pin 1" (shield pin routed to signal ground) problem in the gear at either end - or a SCIN (shield-current-induced-noise) issue in the interconnect cable. Lifting the shield connection just trades one problem for another. But I also realize that once doing something works, most user don't care what the problem was.

 

[MisterCMRR]

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).

I'm a member of AES standards committee working group SC-05-05 (grounding & interfacing) that's currently working on project X152, whose goal is to generate a "best practices" document to persuade equipment manufacturers to be more forthcoming about just what their balanced output and input stages are (by category is where things seem to be headed) so that users don't have to play guessing games and experiment to get good results. Anyone, AES member or not, can have a voice in the creation of this document. See AES Standards participation for details.

 

[JohnRoberts]

Good luck with that seriously.

I recall back in the 80s/90s when things were even less orderly. I shepherd Peavey through their Pin 3 hot to pin 2 hot transition. Peavey had standardized on the wrong pin, years before there were standards for that (When in doubt do what's right).

I also made a point of silkscreening legends on the back plate of my mixers and active boxes to leave no questions for the lads trying to connect up their systems.

Large manufacturers get embarrassing wake up calls when their own SKUs don't play nice with their own sibling products (I saw that happen exactly once at Peavey but once was too many times).

Small companies and junior engineers are more likely to make bad choices.

JR

PS; Using differential inputs everywhere is relatively inexpensive (often just a couple extra resistors), and 3 circuit wiring for all interfaces, balanced (cough) or unbalanced can avoid many embarrassing silences, or bad sounds.

 

[Newmarket]

I recall back in the 80s/90s when things were even less orderly.

Well that's the wonders of 'Standards'. They are so great that we have loads of them IIRC thjere was universal acceptance of 'XLR: socket = Input ; Plug = Output' except for somewhere - I think it was broadcast in Singapore but somewhere anyway - where they went with the alternative.
And weren't Shure Mics reversed wrt pin 2/3 ? And have they changed ?

I also made a point of silkscreening legends on the back plate of my mixers and active boxes to leave no questions for the lads trying to connect up their systems.

Nice ! (Fans of the (UK) comedy show "The Fast Show" may appreciate this more )

PS; Using differential inputs everywhere is relatively inexpensive (often just a couple extra resistors), and 3 circuit wiring for all interfaces, balanced (cough) or unbalanced can avoid many embarrassing silences, or bad sounds.

Indeed. And can still be combined with a great big '0V' busbar for brute force backup.

 

[JohnRoberts]

Well that's the wonders of 'Standards'. They are so great that we have loads of them IIRC thjere was universal acceptance of 'XLR: socket = Input ; Plug = Output' except for somewhere - I think it was broadcast in Singapore but somewhere anyway - where they went with the alternative.
And weren't Shure Mics reversed wrt pin 2/3 ? And have they changed ?

I was following this more closely back in the early 80s. I got RE/P magazine to mail out a questionnaire to all their advertisers asking about pin 2/3 hot. I published the survey results in my "Audio Mythology" magazine column of which manufacturers used pin 2 hot and which used pin 3 hot. There was no clear pattern to which companies were in which group, while even by then the new companies were pretty much already on the pin 2 bandwagon. Several older established companies were saddled with pre-standard choices.

Speaking of microphones my research back then turned up an older IEC standard declaring pin 2 hot for mics, i.e. a positive pressure wave results in a positive voltage. In fact I saw consoles with pin 2 hot for mic inputs but pin 3 hot for line I/O.

Pin 2 or pin 3 hot was generally a minor irritation when interfacing gear together. "Absolute polarity" (I wrote about that too in my column back then) was not considered a thing until the 80s. The most common evidence that polarity mattered at all was when multiple mics picked up the same source (mostly LF) but with opposite polarity causing a suck out or cancellation. One of the required disciplines related to console design is that all the sundry routing options or intermediate I/O (like insert jacks) need to maintain correct polarity. Even EQ bypass switches or HPF circuits were an opportunity to inadvertently flip polarity. I've seen simple inverting stages added to bring polarity back in agreement.

"Absolute polarity", while not very audible (if at all) rose in importance for recording archival purposes back in the 1980s.

Nice ! (Fans of the (UK) comedy show "The Fast Show" may appreciate this more )

I don't know that one but appreciate many UK comedy shows, that got airtime over here. A number of popular US comedy shows were lifted from UK originals. I never thought that the US version of "Office" held a candle to the original.

JR

Indeed. And can still be combined with a great big '0V' busbar for brute force backup.

 

[isophase]

Hello,

Signal is present on both pins hot and cold at the stereo cue outputs.

I made a pair of special TRS/TS cables (+/- >> tip/sleeve/open shield)
The noise is worse with these cables!
And more quiet with the instrument cables.
I also connected one of cue outputs to a small synth amplifier in the room and also get an awful noise out.

i did not have my isobox with me to try that, but it would be ideal to avoid installing one permanently if possible.

All tests were done with the main XLR outputs from the sSL mixer connected to the Apollo interface (XLRF>TRS)
I should have tested disconnecting these cables to check if it would kill the noise from the Manley DI but it didn’t occur to me on the moment. Duhhh

what to do next ?

thanks

 

[Newmarket]

Connect screen via 100n capacitor in parallel with a 100 Ohm resistor.

 

[Bo Deadly]

what to do next ?

What did the noise sound like? More hum or different now?

What I would really like to know is if the SSL SIX cue outputs have no ground sense and they're just driving the cold to ground. I would be shocked if that were the case but I otherwise cannot really explain the noise.

One way to check if the output has ground sense might be to put out a 100Hz tone and measure the cold AC RMS voltage with a meter. Make the level high enough to get a good reading on the meter (like exactly 2 VRMS = 5.63 Vpp). Then, without adjusting the level, put a 100 ohm resistor between the cold pin and the shield and measure the RMS voltage again. The SSL SIX documentation claims the output impedance on each leg is 100 ohms. So if the cold output drops to only 1 VRMS, that means that the full voltage is being applied to the output and that the outputs have no ground sense.

Another thing to check would be to measure what sort of current you're getting between grounds. To do that, put a 10 ohm resistor between the shield of the cable and the sleeve of the DI input and measure regular DC voltage across the resistor. If you get more than 5 mV or so, that means that you have a significant current running between the two devices.

I realize this doesn't solve your immediate problem but if you want to properly fix a problem, you need to understand it.

And I really want to know if SSL cannot bother to use ground sense outputs or if Manley's power supply is junk or both.

 

[ccaudle]

I would really like to know is if the SSL SIX cue outputs have no ground sense

What are you referring to as a ground sense circuit? If you mean either the traditional cross-couple output circuit, or the API version licensed by ThatCorp, an easy way to tell is by measuring the output of the hot leg, then grounding the cold leg. If the hot leg doubles in amplitude then it is a cross coupled circuit, if the hot leg amplitude stays the same then it is not cross coupled.
As pointed out in the app notes from ThatCorp, the point of the API circuit is to solve some shortcomings in the original design. If the SSL does have a cross-coupled output, but using the original configuration, that can sometimes oscillate when the cold leg is connected to ground at the receiver end, and must be connected at the transmitter end.

In that sense it isn't really a transformer replacement, which is why you really need to know the output circuit configuration to know the best way to connect from balanced to unbalanced.

 

[Bo Deadly]

What are you referring to as a ground sense circuit? If you mean either the traditional cross-couple output circuit, or the API version licensed by ThatCorp, an easy way to tell is by measuring the output of the hot leg, then grounding the cold leg. If the hot leg doubles in amplitude then it is a cross coupled circuit, if the hot leg amplitude stays the same then it is not cross coupled.

Yes. Cross-couple, dual-feedback, ground-sensing, whatever you wanna call it. I was thinking something like DRV134 but just as 3 amps and some resistors with 100Rs on the outputs (sense before maybe prevents oscillation?). Although surprisingly THAT 1646 is actually cheaper than DRV134 and cheap enough to justify using it in this case.

I guess I would actually not be shocked if the SSL SIX was driving cold ground. Most folks are not going to tap a DB25 cue output to drive singled ended loads and I suppose the cross-coupled output might add a little noise. But for $1500 and an SSL badge, it would not guess that any outputs were driving cold to ground.

 

[ccaudle]

Cross-couple, dual-feedback, ground-sensing, whatever you wanna call it.

That is the reason I asked, Soundcraft used a scheme on some of their small mixers that they called "ground compensating" in which the cold output (if the manual description is to be believed) is actually an input. I always thought it was something like a buffer driving the ref input on a diff-amp configuration, so that if there was voltage difference between the downstream equipment signal ground and the mixer signal ground, sensing on the cold "output" would then turn that difference from differential mode to common mode. I don't have the schematics though, so I am reading between the lines of the description, could be something different. The output connection is labeled as "Hot," "Ground Sense" on ring, and "Screen" on the shield connection. The manual glossary describes the "ground compensation" outputs as "a technique used on unbalanced outputs to cancel out the effect of ground loops." (bolding mine, not in the original)

I just wanted to make sure that what you were calling "ground-sensing" was not what Soundcraft used to call "ground compensating" since that is obviously a lot different than the traditional cross-coupled balanced output stage.

Most folks are not going to tap a DB25 cue output to drive singled ended loads

The cue output is on a separate 0.25"/6.3mm phone connector, if you notice the OP talked about connecting the cue output to the Manley DI with a quarter-inch patch cable.

I looked up the manual for that SSL SiX model small mixer, and the outputs in question are described as "balanced TRS outputs" but no recommendation on how to connect to an unbalanced input. Also no specifications for performance into an unbalanced input. As Mr. CMRR pointed out, users are really left guessing at the best way to interconnect their equipment with unbalanced gear.

One thing not discussed so far (unless I missed it) is gain structure. The Manley DI is intended for sending instruments to a microphone preamp, there is a switch which changes between -6 dB and -23 dB. Either one is going to require some makeup gain on the Focusrite A/D converter, but if the DI is in -23 dB mode that seems like asking for noise problems when you have to make up that 23 dB in the downstream Focusrite device. Oddly enough the -6 dB setting is labeled "boost" and the -23 dB setting is labeled "unity" so it is not a guarantee that someone would set that up correctly without the manual in front of them.

 

[isophase]

Thank you for your help GDIY!
I won’t be returning to the studio soon as I’m quite busy at the moment, I will do some more tests next time I’m there and report back. Thanks a lot