XLR to RCA cable - how can it work?

[jan.didden]

In this case, I fail to understand what benefit might the OP receive by grounding an active output which sends nothing to the other device

You're very good with colors and bold and italic, I give you that ;-)

I have explained this at least twice in this thread. Does anyone actually read other people's posts??

Third time:
At the output the signal is the difference between the two wires. Each wire can and often has hum and noise with respect to ground (that's the reason we use balanced).
So if you use only one wire, you introduce the hum and noise into the se input. Really very simple.

Now if you connect the other wire to the se ground reference - presto! hum and noise are gone, because the se input now again sees the difference between the two wires.....

Jan

 

[HiFiNutNut]

I finally got a reply from RME this morning. They said there is no problem shorting pin 3 to ground at the XLR outputs of the RME Babyface Pro.

So in what way they would have implemented the XLR outputs do you think?

 

[HiFiNutNut]

I guess I have now two options.

1) Use a XLR - RCA cable. Pin 3 is grounded at the RCA end.

2) Use a XLR - RCA Adaptor and connect with a standard RCA cable. Pin 3 is grounded at the XLR end.

Which one has lower noise?

 

[jan.didden]

In theory 1) should have lower noise. Not sure how big the difference is In practice - depends on cable type and length, environment, etc. Maybe just give it a try.

Jan

 

[torgeirs]

This is the RME partner breakout cable for the original babyface.
ALVA: Products
I don't think RME is picky regarding cables And as a old Babyfase user for some years I agree!

Maybe they have changed output circuits from original babyface, but this is what the old babyface manual states:
The short circuit protected, low impedance 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 will cause a decreased THD (higher distortion) and increased
power consumption!

The unbalanced headphone output is also an option if it is free (from old babyface manual):
The analog outputs Phones (channels 3/4) are available at the breakout cable and on the right
side of the unit. The connectors are not separated electrically. When connecting two headphones
the volume might be reduced. The low impedance and unbalanced output signal is – in
terms of quality – identical to the ones at the line outputs, but limited to +7 dBu.
In case the phones output is to be used as line output, usually an adapter TRS plug to RCA
phono plugs, or TRS plug to TS plugs is required. More on cable codes and pinouts can be
found in chapter 27.2. 8 User's Guide Babyface © RME

Mor info on old babyface in http://www.rme-audio.de/download/bface_e.pdf
page 80

 

[danielwritesbac]

Which poison has the better flavor? Option 1 is orthodox and has slightly lower noise than Option 2.

Both options are easy ways to defeat the high end output feature of that source, straightaway turning it into a mundane source with less audible resolution. For computer-based source that is not an adequate fix, even though it costs less than differential receivers.

For computer sound upgrade to actually be an upgrade, you have to avoid all contact with the really noisy computer V+ and 0V; and, therefore, valid cabling options (if an actual upgrade is desired) does Not include using computer ground for audio signal.

So, if you defeat the noise reduction tech, then it is possible you could get really ordinary computer sound, no better or maybe slightly worse than the built in sound chip. That is possible. Just sayin. If that's what happens, then you'll need some THAT1200 differential receivers at the preamp. And, that's because there's no way to find out what the high end source is supposed to sound like until, after, you hook it up correctly with an XLR to XLR cable.

 

[torgeirs]

This is RME. Far away from internal computer sound specs. No noise at all in everyday use. (That is noise at -110 dB, and is not paperfigures at is real measured noisefigures)
The only comlaint can be that it does not colour the sound in any way.

This is just my experience over several years. And i think a lot of other RME users agree.

 

[JMFahey]

You're very good with colors and bold and italic, I give you that ;-)

Thanks.
The point is to introduce my comments at the approppriate point, but avoid confusion between original text and later comments.

I have explained this at least twice in this thread. Does anyone actually read other people's posts??

I have read what you wrote, but don't agree, hence the comments.

Third time:
At the output the signal is the difference between the two wires.

Well, sort of.
It depends .
And I hope you are meaning "the difference between the two hot wires" (pins 2 and 3), because if not you are giving up before starting.

* If the output came from a floating winding yes, that's definitely the case, because from any end there is NO signal relative to ground (what the RCA input needs, because itself is referred to ground) unless you ground the other end.

* If the output is active, which I find way more plausible, unless somebody states and shows a transformer there (so far nobody did) the standard way to do it is to have two gain stages supplying 2 equal signals, out of phase, each of them referred to ground (big difference with the other possibility) and each equally capable of driving an unbalanced input, referred to ground.

The standard way to do it (this particular image coming from Douglas Self site none the less):

Pin 2 is Hot, Pin 3 is Cold, Pin 1 is Ground.

In that case, each and any of those out of phase but referred to ground outputs, found on pins 2 and 3 respectively, can drive the unbalanced referred to ground RCA connected input.
Sorry for using more words than strictly needed to say so, but I want to be as clear as possible.

In that case, I do not see any advantage in shorting any of those outputs (the unused one) to ground, uselessly, we end up feeding current to ground for no useful purpose.
Just look at the schematic above, it would mean shorting the output of one Op Amp to ground.
What for?
In what way can that help?

Each wire can and often has hum and noise with respect to ground (that's the reason we use balanced).

I have a problem with your use of "has" , if any please say "gets - captures - etc." meaning that it acts as an antenna and gets hum-noise from "outside", from a "dirty" electromagnetic environment.

So if you use only one wire, you introduce the hum and noise into the se input. Really very simple.

Supposing you mean: "a wire carrying an audio signal gets-captures noise-hum from the environment", yes, it's so, no news here.

In fact, both wires, if travelling all the way, will do so, again no news.
But let's see what happens when they reach the receiving end.

Now if you connect the other wire to the se ground reference - presto! hum and noise are gone, because the se input now again sees the difference between the two wires.....

Absolutely not.

The unbalanced RCA receiving end will be connected between one of the wires (pin 2 or 3, pick one) and ground, will receive the audio voltage difference between the used wire and ground, which will be comprised of the desired audio signal plus any hum-noise picked along the way, and amplify it.

It will NOT receive audio voltage from the unused wire, period, since such wire is unconnected to the input and does not develop any voltage from input to ground, which is the audio signal being transmitted.

Now suppose you ground the unused wire (what you are suggesting), either end, be it the transmitting or receiving end, I don't care, it will develop no voltage relative to ground, so it can not compensate for any hum-noise picked along the way.

To make it clear whar are we talking about: the receiving end will receive and eventually amplify any audio voltage difference between its input and ground.

Either XLR pins 2 or 3 can be connected to RCA input; ground is ground of course and by definition it's 0V away from itself.

IF you connect the unused XLR pin to ground (what you are suggesting), it will also be 0V away from ground, there is no way that it can affect the signal actually reaching RCA so there is no way it can affect hum-noise picked along the path, one way or the other.

Similar to: "any number + 0 is the exact same number".

 

[HiFiNutNut]

Thanks, Guys.

I have just downloaded the RME Babyface Pro owners manual and read it. It has the same content as the old version RME Babyface. It does not recommend grounding pin 3, on the contrary to the email advice I received from them this morning. So I have to give up the idea of using a XLR to RCA cable.

However, it says its TRS Phones output jacks , powered by 2 independent driver circuits, can be used as line outputs. It has:

1) a TRS 1/4" 6.35mm output with impedance of 10 ohm, output level at 0 dBFS, 1 kOhm load: +13 dBu, max power @ 0.1% THD: 50 mW to drive a high impedance headphone; and

2) a TRS 1/8" 3.5mm output with impedance of 2 ohm, output level at 0 dBFS, 1 kOhm load: +7 dBu, max power @ 0.1% THD: 70 mW to drive a low impedance headphone.

Which TRS output jack is better suited to drive a preamp?

 

[Davey]

You can still use the XLR outputs......just don't ground pin 3.
The language in the babyface user manual quoted by torgeirs seems perfectly correct.

If you're going to build a cable then just leave out the pin 3 connection, or if you're using XLR/RCA output adaptors simply perform some surgery to cut the wire from pin 1 to pin 3.

I'm not sure why some posters still feel there's some sort of balanced interface advantage that can be gleaned here by grounding pin 3. There isn't. This is an unbalanced connection scheme and there's no improving it by grounding pin 3.
And in the case of outputs that aren't cross-coupled, you can't even recover the 6db gain lost with the single-ended connection.

Cheers,

Dave.

 

[Davey]

Elaborating a bit further:

One of the issues I have noticed with this scheme is that some amount of channel-matching is lost on units that have multiple outputs. More components and their tolerances now factor in where they wouldn't have if using a conventional balanced drive scheme. In that case, only the tolerances of the (using the example of Fahey's schematic above) components of the inverting amplifier would factor in. This requires a "calibration" of the output levels of a multi-channel unit in some cases where it wouldn't been required previously.

FWIW,

Dave.

 

[rdf]

This is an unbalanced connection scheme and there's no improving it by grounding pin 3.

All correct answers, agree with everything you wrote. Typically - transformer and instrumentation stages excepted - all grounding pin 3 does is shunt an unused opamp output to ground through an internal isolation resistor. With low output impedance devices this can be worse than letting it float.

 

[torgeirs]

Thanks, Guys.

I have just downloaded the RME Babyface Pro owners manual and read it.

1) a TRS 1/4" 6.35mm output with impedance of 10 ohm, output level at 0 dBFS, 1 kOhm load: +13 dBu, max power @ 0.1% THD: 50 mW to drive a high impedance headphone; and

2) a TRS 1/8" 3.5mm output with impedance of 2 ohm, output level at 0 dBFS, 1 kOhm load: +7 dBu, max power @ 0.1% THD: 70 mW to drive a low impedance headphone.

Which TRS output jack is better suited to drive a preamp?

Best gong to the correct manual. Could not find it yesterday.
1) is about 4 volt rms Max
2) is about 2 volts rms Max
Think 2) is most common for ubalansed Line in.
If you can't play high enough try 1)

If you play dynamically compressed music, I would attunate in rme mixer -6dB and use 2) to avoid intersample overload. (Or maybe RME handles it without attunation?)

 

[jan.didden]

Thanks.
The point is to introduce my comments at the approppriate point, but avoid confusion between original text and later comments.


I have read what you wrote, but don't agree, hence the comments.


Well, sort of.
It depends .
And I hope you are meaning "the difference between the two hot wires" (pins 2 and 3), because if not you are giving up before starting.

* If the output came from a floating winding yes, that's definitely the case, because from any end there is NO signal relative to ground (what the RCA input needs, because itself is referred to ground) unless you ground the other end.

* If the output is active, which I find way more plausible, unless somebody states and shows a transformer there (so far nobody did) the standard way to do it is to have two gain stages supplying 2 equal signals, out of phase, each of them referred to ground (big difference with the other possibility) and each equally capable of driving an unbalanced input, referred to ground.

The standard way to do it (this particular image coming from Douglas Self site none the less):

Pin 2 is Hot, Pin 3 is Cold, Pin 1 is Ground.

In that case, each and any of those out of phase but referred to ground outputs, found on pins 2 and 3 respectively, can drive the unbalanced referred to ground RCA connected input.
Sorry for using more words than strictly needed to say so, but I want to be as clear as possible.

In that case, I do not see any advantage in shorting any of those outputs (the unused one) to ground, uselessly, we end up feeding current to ground for no useful purpose.
Just look at the schematic above, it would mean shorting the output of one Op Amp to ground.
What for?
In what way can that help?


I have a problem with your use of "has" , if any please say "gets - captures - etc." meaning that it acts as an antenna and gets hum-noise from "outside", from a "dirty" electromagnetic environment.


Supposing you mean: "a wire carrying an audio signal gets-captures noise-hum from the environment", yes, it's so, no news here.

In fact, both wires, if travelling all the way, will do so, again no news.
But let's see what happens when they reach the receiving end.


Absolutely not.

The unbalanced RCA receiving end will be connected between one of the wires (pin 2 or 3, pick one) and ground, will receive the audio voltage difference between the used wire and ground, which will be comprised of the desired audio signal plus any hum-noise picked along the way, and amplify it.

It will NOT receive audio voltage from the unused wire, period, since such wire is unconnected to the input and does not develop any voltage from input to ground, which is the audio signal being transmitted.

Now suppose you ground the unused wire (what you are suggesting), either end, be it the transmitting or receiving end, I don't care, it will develop no voltage relative to ground, so it can not compensate for any hum-noise picked along the way.

To make it clear whar are we talking about: the receiving end will receive and eventually amplify any audio voltage difference between its input and ground.

Either XLR pins 2 or 3 can be connected to RCA input; ground is ground of course and by definition it's 0V away from itself.

IF you connect the unused XLR pin to ground (what you are suggesting), it will also be 0V away from ground, there is no way that it can affect the signal actually reaching RCA so there is no way it can affect hum-noise picked along the path, one way or the other.

Similar to: "any number + 0 is the exact same number".

You are wrong on a basic level, You worry to much about circuitry. Balanced connection is about the difference between two wires. The relation of each to ground is irelevant. Each may have hum and noise, but the point of balanced is that the hum and noise on each is equal and herefor cancels at the balanced receiver. If you use only one wire you get the hum and noise again, for free....

I also have the impression that you feel that balanced means two signals of equal amplitude and opposite phase which of course is not the case.
But you will get the last word - either you get it or not, not much more I can do.

Jan

 

[benb]

Generally the output resistance of the source driving XLR pin 3 is high enough that grounding it won't damage the circuitry. On some XLR outputs pin 3 isn't driven actively and is grounded via a resistor. No issue in that case.
In any case, I don't think it's good practice to pull the pin 3 output to ground in this way and I modify XLR/RCA adaptors to remove it or when building cables leave it unconnected.

Some professional equipment senses this type of pin 1/3 short on XLR outputs and reprograms gain to increase automatically.

I'm not sure I'd call it "reprograms" (unless it's something strange I'm not familiar with). This circuit labeled "Active Floating Source" in Fig. 12 simulates the action of a floating transformer in regard to different impedances on the + and - lines (if one output is grounded the other carries the whole differential voltage):
Balanced Interfaces

Jan,

I certainly understand where you're coming from on this.
There are two (maybe more) schools of thought on this particular bal/unbal interconnect scheme. I've seen it recommended in a few different ways.

Here's one that might be of interest:
Sound System Interconnection

I was going to post that.

You are wrong on a basic level, You worry to much about circuitry. Balanced connection is about the difference between two wires. The relation of each to ground is irelevant.

Balanced is about the two wires having the same IMPEDANCE to ground, both at the sending end (the two Rs resistors in the schematic) and the receiving end (though that's mostly irrelevant in the OP's case of connecting to an unbalanced input).

Consumer devices? This [the sending end grounding pin 3 as opposed to driving it] is the rare exception in the pro and semi-pro world in my experience, usually reserved for lowest end 'prosumer' gear.

You mean like the M-Audio Delta 44? I don't know of any others offhand, but that's one example of where it's done.

 

[Davey]

I'm not sure I'd call it "reprograms" (unless it's something strange I'm not familiar with). This circuit labeled "Active Floating Source" in Fig. 12 simulates the action of a floating transformer in regard to different impedances on the + and - lines (if one output is grounded the other carries the whole differential voltage):
Balanced Interfaces

Well, call it whatever you want. I'm talking about a cross-coupled scheme where if one output of an active driver pair is grounded it increases the gain on the other side by 6db. It's common in much of the professional gear I've seen. (In all of the Behringer gear I've seen.)

It's certainly not the best scheme and the source now becomes unbalanced, but it does allow to recover the 6db gain lost when not using a proper balanced interface. Whether that's a good thing...gain structure wise....is another topic. When using "professional" gear in a "consumer"/domestic setting, losing that 6db is probably preferable.

Cheers,

Dave.

 

[tomchr]

Are the THAT chips as good as or better than some good transformers?

In some aspects, they're better.

A fair comparison transformer is the Jensen JT-11-P1. I'll compare it against the THAT1200.

Cost: JT-11-P1: $70. THAT1200: $6.10
CMRR (60 Hz): JT-11-P1: 107 dB. THAT1200: 90 dB
CMRR (3 kHz): JT-11-P1: 73 dB. THAT1200: 75 dB
THD+N (+4 dBu, 20 Hz): JT-11-P1: 0.025 %. THAT1200: 0.00065 %

Basically, for an additional $64, you get higher distortion and higher CMRR at 60 Hz. Whether the additional CMRR of the transformer justifies the >10x cost riser is up to you.

The Jensen transformers are nice transformers. I use a pair in my DG300B. In tube circuits, input transformers can make life easier as they allow for the elimination of the input capacitor.

Tom

 

[Davey]

Well, I don't know if the cost comparison is exactly fair. You obviously need some support circuitry and power supply to make that $6.10 THAT1200 chip work.
Maybe something like this?? THAT Receiver – Mono

All-in-all, I think the transformers are the cheaper overall option.
Jensen has a variety of configurations of the ISO-MAX units at decent prices. (Although they have raised the price on those somewhat.)

Dave.

 

[JMFahey]

You worry to much about circuitry.

Maybe because it all boils down to it?

herefor cancels at the balanced receiver.

Please show me a balanced receiver circuit (which by definition has two out of phase same gain same impedance inputs) built around a single RCA connector which by definition has one input and ground.

 

[Pano]

The simplest questions are the hardest, it seems.