Is a 4-20mA current loop transmitter OK for high quality audio?

[e100]

Something like the Texas XTR117 http://www.ti.com/lit/ds/symlink/xtr117.pdf has a flat frequency response over the audio range and is described in the datasheet as a 'precision device'.
Outside of the telephone phone world there doesn't seem to be much information about current loop powered devices being used in audio applications.
Is the phantom powered balanced type circuit a better technical solution for some reason? 

[David Hess]

There is no particular reason it cannot be done.  I have used current outputs to drive audio inputs before to avoid ground loops with good results.  Quality can certainly be better than necessary for most applications.

[SiliconWizard]

Interesting question.

There is no distortion figures that I could find though. And I'm tempted to think that they would not be that impressive in real-life audio applications. You'd have to define "high quality audio", but IMO, distortion could be one point where this approach could fall short.

Would be interesting to build a test jig and get some measurements. "High quality audio" unfortunately needs expensive equipment to get accurate figures, especially distortion figures.
Just my 2 cents.

[NiHaoMike]

Would be interesting to build a test jig and get some measurements. "High quality audio" unfortunately needs expensive equipment to get accurate figures, especially distortion figures.
Just my 2 cents.

A sound card with 110dB SNR input and output is like $100. Not even expensive by mainstream audio standards.
https://www.newegg.com/Product/Product.aspx?Item=N82E16829132100

[T3sl4co1l]

Probably.  It won't save you from induced noise, however.  You need to use coax, or fully differential signaling on twisted pair, for that -- a naive implementation might keep one side at ground and suffer from common mode interference.  It's common enough to implement current loops with isolation, but that introduces other opportunities for noise (DC-DC converter), plus has the difficulty of needing a linear, low noise analog isolator.

Tim

[NiHaoMike]

It's common enough to implement current loops with isolation, but that introduces other opportunities for noise (DC-DC converter)

20mA is a small current. A very low noise and very high isolation (but very low efficiency) DC/DC converter for that application could be some LEDs illuminating a small solar panel. But it would likely make a lot more sense to just go digital for sending audio over a distance.

[ebclr]

Fiber will be much easier and better

[Zero999]

Fiber will be much easier and better

How will it be easier?

Optical fibre will be subject to less distortion, but there's the tooling and hassle of having to terminate the connector. With wire it's easy using terminal blocks or solder.

[SiliconWizard]

Probably.  It won't save you from induced noise, however.  You need to use coax, or fully differential signaling on twisted pair, for that -- a naive implementation might keep one side at ground and suffer from common mode interference.  It's common enough to implement current loops with isolation, but that introduces other opportunities for noise (DC-DC converter), plus has the difficulty of needing a linear, low noise analog isolator.

Tim

Absolutely.
The sound card approach will get you some preliminary measurements, but there are many pitfalls as you mentioned. Besides, you wouldn't be able to really publish your results.
If all you intend to build is a home-made or in-house device, that's fine. But if you intend to sell a product flagged as "high-quality audio"... not so much.

This kind of equipment is relatively standard for audio measurements: https://www.rohde-schwarz.com/us/product/upv-productstartpage_63493-7558.html
May seem overkill but it's hard to beat with a $100 sound card. Of course in a preliminary prototyping phase, the sound card will probably be fine.

[Zero999]

But if you intend to sell a product flagged as "high-quality audio"... not so much.

I don't see why not. Many audio companies seem to pull figures out of their arses and they get away with it too.

The specification for the XTR117 is more than good enough for audio. I doubt any double blind listening tests would be able reliably distinguish it from a NE5532 or anything else half decent for that matter.

[LaserSteve]

Take a look at "Holm Impulse" software for the measurements.   Pretty Good...
Designed for speaker and room impulse testing, but remarkably good on the bench, too...

Steve

[mikerj]

The only thing I can't see in the datasheet that would be pertinent for high quality audio is noise performance.

[SiliconWizard]

But if you intend to sell a product flagged as "high-quality audio"... not so much.

I don't see why not. Many audio companies seem to pull figures out of their arses and they get away with it too.

Just because this is not a dependable and calibrated measurement equipment. A basic (IMO) metrology consideration. This would be the same, IMO, as selling power supplies that would be designed and controlled with cheap and non-calibrated $30 multimeters. If you're not in a regulated industry, nothing prevents you from selling that of course (it should still comply to basic safety regulations but that's another matter). But it's just my take on this. I don't see a reason why we should consider audio devices as not worthwhile of being treated as other kind of devices. We're talking "high-quality" here, as I understood it, and not cheap products. Companies that "pull figures out of their arses", especially in the audio field where those given figures are often not only misleading but wildly optimistic, are not dependable companies. They are hurting the business, the concept of quality and common sense. Just my 2 cents again.

The specification for the XTR117 is more than good enough for audio. I doubt any double blind listening tests would be able reliably distinguish it from a NE5532 or anything else half decent for that matter.

Well, maybe, but I don't know. Again, they don't seem to provide some key figures such as distortion figures, and given the nature of a typical voltage-to-current converter, I'd suspect potential issues with that.

Again, that's probably "good enough" as you said (which would be a pretty vague concept), but somehow, I have a hard time seeing "high quality" (proven with correct measurements) and "good enough" (being acceptable to the listening ear) in the same area. Even so, I'd be interested in listening to the result.

[Zero999]

The only thing I can't see in the datasheet that would be pertinent for high quality audio is noise performance.

The bandwidth and slew rate are ample for audio and the non-linearity is low enough that it should translate into respectable THD figures.

Of course it should be tested, especially as it's being used for an application it was never intended for, but there's no point in going over the top.