Balanced to Unbalanced

[Clint Goss]

I frequently need to send a balanced XLR signal to an unbalanced destination. This needs to be done on the fly in my performance rig (Ableton Push 3 and a Sylphyo wind controller rendered through several hardware modules), depending on the house system at the venue and the gear of other performers that are running through my rig.

My major confusion is: How do I tell whether to tie pin 1 to pin 3 at the XLR source end?

Some manuals I have warn that doing it wrong can cause distortion or even damage the output stage of the source device. An example of this warning is in the manual for my RME Babyface Pro FS:

The short circuit protected, low impedance XLR line outputs do not operate servo balanced! When connecting unbalanced equipment, make sure pin 3 of the XLR output is not connected. A connection to ground might cause a decreased THD (higher distortion) and increased power consumption!

OK, so that implies that if the manual says that the XLR output is servo-balanced, then Pin 1 should be tied to Pin 3.

All other gear I have encountered, at best, describes the output stage variously as:

• Cross-coupled
• Transformer-balanced
• Active
• Passive
• Electronically balanced
• Servo-balanced

... or are simply silent on the design of the output stage (maybe this is "none of the above"??)

Could someone identify, for the above terms (and maybe any others that I've missed), whether to tie Pin 1 to Pin 3 at the source end??

Also: What would you suggest if the documentation is silent on the design of the output stage?

For reference, I think these are my choices for wiring adapter cables:

As I understand it, Case C, using a twisted pair cable and routing the negative / cold / Pin 3 to the shield of the destination will get at least part of the Common Mode Rejection (20-30dB) that a fully balanced connection would (60dB). I am hoping I could use Case C in any situation where Pin 3 can be tied to Pin 1.

The main reference for this is: Rane Technical Staff, 2015, Rane Note #110, Sound System Interconnection, Figure 4, Cases 3 and 5.

Thanks for any feedback you could provide!

 

[somebodyelse]

If the documentation is silent then don't use a cable to do it. The terms aren't a definitive guide because even for the same term it may depend on the implementation.

The safest and universally effective way is to use a purpose built balanced to single ended converter. It could be a box with a transformer in it, or an active one.

 

[Clint Goss]

Thanks for the feedback!

> then don't use a cable to do it.

i.e. "use a transformer" I think is what you are suggesting?

That's not so practical for me, since this is road gear and needs to be a minimal / lightweight as possible. I might need zero of these in one venue, and three of them in another venue, depending on the venue and what other performers are routing their hear through mine ...

I'm really looking for a cable adapter solution ...

 

[sergeauckland]

In that case you need to find out how the balanced output is implemented, preferably from the manufacturers. If that's not possible, the safest is to connect only pins 1 and 2, leaving pin 3 floating. If that works, then all well and good. If it doesn't, then connect pin3 to pin 1.

Note that by leaving pin 3 floating, you will only get half the signal, so don't be fooled by it being a little quieter than with pin 3 grounded.

As Somebodyelse suggested the only safe and reliable way is to use a transformer or active box. My preference is for a transformer as it's passive, requires no power, is smaller and gives galvanic isolation.

S

 

[Clint Goss]

Cool. Thank you for that ...

> how the balanced output is implemented,

If they say cross-coupled, then should I tie Pin 1 to pin 3?

If they say transformer, then should I tie Pin 1 to Pin 3?

If they say active, then should I tie Pin 1 to Pin 3?

If they say passive, then should I tie Pin 1 to Pin 3?

If they say electronically balanced, then should I tie Pin 1 to Pin 3?

If they say servo-balanced, then should I tie Pin 1 to Pin 3?

Any other terms that might be listed??

 

[MaxwellsEq]

I don't think you can ever come up with a safe way of cabling this unless you know the internals OR the manuals have explicit instructions on how to do it. Selecting a wiring solution, based on descriptive wording is unlikely to be reliable. Recording studios encounter this on a daily basis, which is why they have a cupboard of DI or balancing boxes.

 

[sergeauckland]

Exactly this ^. I do this all the time, and always ask the manufacturers or consult the Service Manual. If neither is possible, then I keep some transformers for just such use. They also stop ground loops with unbalanced equipment, so doubly useful.

S

 

[joeren]

I would think that a transformer would be the best solution. The reasoning is that they are passive, relatively inexpensive, and when used closed to the single ended destination, you have the advantage of using the balanced line and shield for the majority of the length of run. A good 2:1 ratio, shielded audio transformer, connecting the balanced input GND, Plus & Neg. On the output side tie Positive to the signal, the Neg to the destination equipment common and leave the GND open. I saw this on eBay. https://www.ebay.com/itm/334340331699

 

[Clint Goss]

Ok folks, this is all very, very helpful. Thank you all!

I though I needed a wiring solution, and I really needed a new mindset on this subject ...

I would prefer to have a complete product rather than the DIY board suggested on EBay ..

I do use Radial DI boxes with Jensen transformers (JDI - passive, and J48 - active) for performers sending unbalanced into my system. I also have several EBTech Hum eliminators, although I'm not sure how they are built (?transformer based?).

Would my EBTech Hum eliminators fit the bill?

Or another commercial product with transformers?

 

[theREALdotnet]

My major confusion is: How do I tell whether to tie pin 1 to pin 3 at the XLR source end?

I don't understand why pin 1 would be connected at all? With XLR the signal is between pins 2 and 3, those should be connected to S and T below.

View attachment 430226

I'd go with C, but don't connect pin 1 to pin 3. You can connect pin 1 to the shield at the XLR plug only.

The problem remains that in a stereo (or multi-channel) setup, the XLR pins that are connected to their respective S contacts at the RCA end may potentially be shorted together, because RCA S contacts often are. This can cause issues at the XLR source, which is not normally designed for pins 3 of the stereo channels to be shorted together. Without seeing the schematics of both connected devices it's hard to predict what will happen, and this is the reason XLR-to-RCA cables are usually recommended against (not just by RME, also by Topping, for example).

 

[Clint Goss]

Also looking at the Radial J-ISO unit (https://www.radialeng.com/product/j-iso) ... which I'm guessing does a similar job to the EBTech Hum Eliminator, but with higher-quality transformers ... is that close to correct??

 

[sergeauckland]

Also looking at the Radial J-ISO unit (https://www.radialeng.com/product/j-iso) ... which I'm guessing does a similar job to the EBTech Hum Eliminator, but with higher-quality transformers ... is that close to correct??

Sadly, can't tell from the spec.

SPECIFICATIONS Pro-ISO J•ISO Freq. response 20Hz to 18kHz (+0dB, -3dB) 20Hz to 20kHz (+0dB, -1.5dB) THD @ 1k 0.005% @ 1kHz 0.001% @ 1kHz Stereo Inputs Balanced, low-Z, +4dB line-level, female XLR, pin-2 hot

The frequency response is hardly stellar, although probably OK, Also, distortion is quoted only at 1kHz, so not difficult to achieve. They claim 0.1% at +20dBu which would be fine if at 50Hz, not great if only at 1kHz.

Nevertheless, these items look well made and very rugged for life on the road.

S.

 

[somebodyelse]

Cool. Thank you for that ...

> how the balanced output is implemented,

If they say cross-coupled, then should I tie Pin 1 to pin 3?

If they say transformer, then should I tie Pin 1 to Pin 3?

If they say active, then should I tie Pin 1 to Pin 3?

If they say passive, then should I tie Pin 1 to Pin 3?

If they say electronically balanced, then should I tie Pin 1 to Pin 3?

If they say servo-balanced, then should I tie Pin 1 to Pin 3?

Any other terms that might be listed??

There's more than one way to implement most of those descriptions, and depending on how they did it the answer may be either yes or no, or some other pinout. Getting it wrong might cause damage in the worst case, hence the advice not to even try it if you don't know. You don't want to be damaging other peoples' gear.

 

[naitsu]

When I was dealing with the same dilemma, I found this resource the most helpful, because it covers and explains several different scenarios and doesn't just tell to do XYZ period like many others do:

Dealing with unbalanced gear in the studio - Crookwood

Sigh. Dealing with unbalanced gear in the studio is a pain and always results in a compromise. Why? Well because: The tiny audio signal return goes down the same wire as the mains ground currents The levels are often lower than pro levels The output impedances are often stupidly large However...

crookwood.com

From there I gathered the instructions for my case:
Connecting hard balanced outputs to mains grounded unbalanced inputs
- Always use balanced cable (two internal wires, a hot wire and cold wire, plus a separate screen).
- At the unbalanced end, connect the hot wire to the output or input pin, and the cold wire to the ground pin.
- At the balanced end, connect the hot wire to the hot pin (XLR pin 2), and the cold wire – when it’s a hard balanced output - connect it to the shield pin (XLR pin 1).
Screen:
- At the unbalanced input, connect the screen wire to the input ground pin.
- At the balanced output, connect the screen wire to the screen pin (XLR pin 1).
- At the balanced output, leave XLR pin 3 unattached – the cold wire should be attached to the screen pin (XLR pin 1).
Note: leaving pin 3 floating lets the hard balanced circuit work properly, and by connecting the grounds at both ends, you maximise RF rejection and minimise hum.

And this is the result.

This looks to correspond your A option, but using a shielded twisted pair cable.

Actually I made TRS-XLR and XLR-TS adapters so that I can use any length XLR I need for the run, but the principle was the same.

I can't remember how I determined that my balanced output was a "hard balanced" type. But at least I know that it is a true balanced output because both hot and cold provides a signal (opposite polarity). Or maybe I just figured or read that it's safer to let cold float, if you don't know the exact type. It just might not be optimal in some cases. (See the RME warning in the original post). But what contradicts with this is that all pre-made XLR to TS adapters that I have seen connect cold to and ground together. Go figure.

My device is a Motu 624 and here's the specs.

 

[DVDdoug]

Radial has an excellent reputation. That's probably the best and most foolproof solution. Note that most transformer "direct boxes" are also impedance converters for guitars. That Radial box is correct for your use.

There are active converters that might have better performance (than a transformer) but they require power and the extra power connections is a disadvantage if you are frequently moving things around and changing connections. And you probably don't want to rely on a battery either. Good transformers can be audibly perfect and a transformer will last forever!

Transformers can normally also be use the other way around (unbalanced to balanced) but with the Radial the volume control on the unbalanced output might mess that up, and Radial doesn't say it can be used that way and it's clearly marked "input" and "output".

 

[Lambda]

That's not so practical for me, since this is road gear and needs to be a minimal / lightweight as possible.

Especially if this is your road gear you Should always have transformer based BalUn and isolation with you to interface with unknown external gear.
its simply the most reliable and universal solution.

Neutrik NA2 F-D0B-TX
Would be small and light if this is your concern.

 

[Clint Goss]

Since this topic is so confusing (to me, anyway), I decided to do a Full deep dive into the subject from a musician (not EE) perspective. Hope it can be useful to some …

Not sure I got it all correct … so … if anyone has the cycles to look through this draft PDF …

Balanced to Unbalanced Audio Connections

… I would love to hear any and all feedback.

And, BTW, we’re still on the road, so I can’t actually test any of this yet … hence it will stay DRAFT for a while …

Also, I have not addressed the situation where, say, the shields of several RCA connectors on the unbalanced destination unit are tied together and I am sending a balanced signal using adapter cables to several of those RCA destinations. If the sources are, say, cross-coupled, I think things will get “messed up” …

 

[sergeauckland]

Firstly, a Good Effort, you probably understand a lot more about it now. I've always thought that if you can't explain it simply, you don't understand it!

However, a few errors or incomplete things I've noticed:-

Page 1
AES-EBU is not bi-directional. It is unidirectional, source to destination, but carries both channels in one stream, plus various status bits.

A Balanced signal only requires two conductors. The screen is optional at line level, generally necessary on microphone circuits, or very long line level cables but not necessary at line level for shorter circuits or for digital signals.

Page 2

??? Is This Valid Generally yes, whether transformer balanced or electronic.

Page 3

Transformer Balanced Outputs. Don't forget centre tapped transformer balanced outputs.

Each Signal isn't 'half the voltage', it's the whole voltage but of opposite polarity. It's only half the voltage if ground-referenced such as electronically balanced but not cross-coupled, or centre-tapped transformer outputs.

Page 4

Do It Right

Using headphones as you suggest won't tell you much as there's no way to distinguish between centre-tapped transformer and electronically balanced (non-servoed). Also, with fully floating outputs whether transformer or electronic, headphones will produce sound out of phase. You do mention this later on.

I've always found this whole issue of balanced outputs into unbalanced inputs largely misunderstood, so have always preferred physical transformers. I don't like cables or adapters for all the reasons you mention. I do use such cables at home, but then I know what's inside the equipment I'm using, so the cables are wired (and labelled!) correctly. You can't do that in the field with unknown random equipment so transformers are the best by far.

Finally, the Words of Caution should be engraved on the heart of every sound engineer! Don't EVER disconnect the safety earth except in the comfort of your own home, when you're totally sober and put it back as soon as you've confirmed the problem. NEVER EVER walk away and leave the safety earth disconnected even at home.

Hope this helps.

S.

 

[somebodyelse]

It's a good start. Another point to add to the list above:

All transformer-balanced and electronically-balanced (cross-coupled and inverter-
balanced) outputs are Symmetrically-balanced. These designs send the audio
signal on two of the three balanced output connectors. Each signal carries half of
the signal voltage, and the two signals are mirrors of each other (“polarity-
inverted” with “equal and opposite waveforms”). This mirroring is the basis for the
substantial noise-cancelling characteristics of balanced connections.


• Impedance-balanced outputs are Asymmetrically-balanced. They send the entire
signal voltage on only one of the balanced output connectors. The connector that
would normally carry the mirror-image (“symmetrical”) signal is held at 0 volts.
Interestingly, the noise-cancelling system still works, and some practitioners
actually consider this simpler and less costly design to be preferable.

My bold/italic for the bits that represent a common misunderstanding of how the noise cancelling works. Re-read Whitlock, or The G Word - probably more than once if my experience is anything to go by. If it was easy people wouldn't still be making the same mistakes after all these years.

 

[sergeauckland]

It's a good start. Another point to add to the list above:

My bold/italic for the bits that represent a common misunderstanding of how the noise cancelling works. Re-read Whitlock, or The G Word - probably more than once if my experience is anything to go by. If it was easy people wouldn't still be making the same mistakes after all these years.

The noise cancelling properties of a balanced circuit is because the receiver responds ONLY to the difference between the two legs, and the interfering signal is deemed to be equally in both legs so cancels out in the receiver. For this to happen, the source and the receiver have to be balanced, as well as differential. By 'balanced' this means that the impedances and capacitances to signal and to ground have to be identical so that interference genuinely induces equally into both legs, to be cancelled in the receiver.
Now in reality, nothing is ever perfectly balanced, capacitances and impedances will vary, so the cancellation of interference will be limited. Nevertheless, this 'Common-Mode rejection' will typically be at least 60dB especially at audio frequencies, so that represents a pretty impressive reduction in induced interference. There is also a limit to how big the common-mode interfering signal is, as semiconductor input stages won't handle high amounts of interference. Fortunately, there's rarely need for very high common-mode voltage capability, but one can occasionally get caught out. It's another benefit of transformer balancing as these can take hundreds of volts of common-mode without damage, if designed to do so.