Balancing output of a consumer EQ

[JohnRoberts]

If your interface is electronically balanced then it may be that you are losing 6dB at the input to the unbalanced EQ. To recover it you could fit a 10K:10K isolating transformer at the EQ input. Then you could simply impedance balance the output without losing any signal level.

Cheers

Ian

good idea....

Realistic IIRC is a radio shack brand so not even semi-pro. What kind of connectors are on the inputs and outputs?

If two circuit like RCA, it is surely not differential.

FWIW some balanced output drivers sense for  (single ended) grounded legs and adjust up the level on the other active leg.

Adding a differential input may be cheaper than adding a transformer, but a transformer is easier (and you can keep it after you replace the EQ).

JR 

 

[Rocinante]

good idea....

Realistic IIRC is a radio shack brand so not even semi-pro. What kind of connectors are on the inputs and outputs?

Adding a differential input may be cheaper than adding a transformer, but a transformer is easier (and you can keep it after you replace the EQ).

JR

Lol.  That was actually pretty funny.

Lots of good information on this thread.

 

[craigmorris74]

If your interface is electronically balanced then it may be that you are losing 6dB at the input to the unbalanced EQ. To recover it you could fit a 10K:10K isolating transformer at the EQ input. Then you could simply impedance balance the output without losing any signal level.

Cheers

Ian

Thanks Ian-this looks like a pretty easy solution.  Need to get a transformer from Edcor to try it out.

 

[craigmorris74]

What loss? This circuit is unity gain. If you experience loss, it means you have a wrong connection.
Can you detail how you connect it (input and output)?

I'm getting 6dB of loss coming from a UA2192 through the EQ and back to the UA2192 vs running out and in of the UA2192 with nothing patched in.

I  get less than 0.5dB of loss when inserting a piece of transformer balanced gear before the EQ.

 

[abbey road d enfer]

I'm getting 6dB of loss coming from a UA2192 through the EQ and back to the UA2192 vs running out and in of the UA2192 with nothing patched in.

I  get less than 0.5dB of loss when inserting a piece of transformer balanced gear before the EQ.

Is it so difficult to describe exactly how you connect things? You don't help us to help you.

 

[craigmorris74]

Is it so difficult to describe exactly how you connect things? You don't help us to help you.

Are you talking about the cable I used?  I used cables constructed according to the Rane recommendation for connecting balanced to unbalanced gear.

 

[abbey road d enfer]

Are you talking about the cable I used?  I used cables constructed according to the Rane recommendation for connecting balanced to unbalanced gear.

There are several ways of connecting bajanced to unbalanced, some are better than others, depending on the nature of the balanced connection. If you lose 6dB, then the connection is not right.

 

[moamps]

If you lose 6dB, then the connection is not right.

According the UA2192 manual:
(14) Analog Line Outputs - Analog signals are output from the left and right channels of the 2192 D/A
converter via these balanced line-level male XLR connectors. Pin 2 is hot. For unbalanced operation,
Pin 3 can be grounded, and the output level will be attenuated by 6db. The analog outputs are factory
calibrated so that a digital signal level of -18dBFS will output an analog level of +4dBu, for 18dB of
headroom. The line outputs can drive high or low (600ohm) impedance inputs with no changes in level.
( see page 30 for more information)
maybe isn't possible to avoid 6dB loss.
Btw, analog part  looks very interesting:

 

[ruffrecords]

According the UA2192 manual:
(14) Analog Line Outputs - Analog signals are output from the left and right channels of the 2192 D/A
converter via these balanced line-level male XLR connectors. Pin 2 is hot. For unbalanced operation,
Pin 3 can be grounded, and the output level will be attenuated by 6db.

Sounds like typical impedance balanced outputs.

Cheers

Ian

 

[Newmarket]

Sounds like typical impedance balanced outputs.

Not to me. I'm understanding 'impedance balanced' to be simply similar impedances on both 'L' and 'R' legs but no signal on 'R' (XLR Pin3  or R(ing) of TRS jack by current convention). So 'grounding' pin 3 simply bypasses the balancing impedance. There's no signal to lose ?

 

[moamps]

Symmetrical output (signal balanced) is usually impedance balanced in the same time.
Only impedance  balanced output doesn't have signal at cold output.
From the info I quoted, we don't know is the output impedance balanced (but we hope so),
but we know that there is a signal at the cold output which will be lost by bypassing this pin to ground.
IMO, this is setup similar to GSSL output topology (active, symmetrical, but no sensing grounded output in order to rise remaining active output for 6dB, as SSM and THAT output chips do). 

 

[JohnRoberts]

Not to me. I'm understanding 'impedance balanced' to be simply similar impedances on both 'L' and 'R' legs but no signal on 'R' (XLR Pin3  or R(ing) of TRS jack by current convention). So 'grounding' pin 3 simply bypasses the balancing impedance. There's no signal to lose ?

Agreed, sounds like simple differential output (two opposite polarity legs) but with outputs tolerant of persistent short circuit.

Improved output drivers use cross connected negative feedback that sense if if one leg is shorted and boosts the other to maintain output level, at least until it saturates, 6dB sooner than with both legs active.

JR 

 

[user 37518]

Agreed, sounds like simple differential output (two opposite polarity legs) but with outputs tolerant of persistent short circuit.

Improved output drivers use cross connected negative feedback that sense if if one leg is shorted and boosts the other to maintain output level, at least until it saturates, 6dB sooner than with both legs active.

JR

Thats known as a Cross-coupled Output Stage https://www.eetimes.com/cross-coupled-output-stages-for-balanced-audio-interfaces/#. It is attributed to both MCI and HP

 

[Bo Deadly]

According the UA2192 manual:
(14) Analog Line Outputs - Analog signals are output from the left and right channels of the 2192 D/A
converter via these balanced line-level male XLR connectors. Pin 2 is hot. For unbalanced operation,
Pin 3 can be grounded, and the output level will be attenuated by 6db.

That's somewhat atypical for pro gear. The 6dB loss occurs when the balanced output is implemented by independently generating the cold output but the input it's connected to is not differential. Imagine a signal that is 1VAC. Meaning when the hot leg swings to +0.5V, the cold leg is -0.5V. A differential input computes the "difference" and recovers 1V for 0dB loss. But if the cold input is grounded, that half is lost and only the hot 0.5V is used which represents a 6dB loss.

 

[moamps]

That's somewhat atypical for pro gear.

If you look little bit closer into picture I attached, you will see a little bit atypical analog part/front end. 

Imagine a signal that is 1VAC. Meaning when the hot leg swings to +0.5V, the cold leg is -0.5V. A differential input computes the "difference" and recovers 1V for 0dB loss. But if the cold input is grounded, that half is lost and only the hot 0.5V is used which represents a 6dB loss.

Unfortunately, I imagined that for the first time more than 35 years ago. 
BTW, you are expert for "pin1" problem here, what we will do with pin 3, is it correct to connect them  directly to pin 1?

 

[abbey road d enfer]

If you look little bit closer into picture I attached, you will see a little bit atypical analog part/front end. 
Unfortunately, I imagined that for the first time more than 35 years ago. 
BTW, you are expert for "pin1" problem here, what we will do with pin 3, is it correct to connect them  directly to pin 1?

Looks to me like the infamous "Tascam problem", with a twist, where the outputs are protected against permanent short.
If it is, unused pin 3 should be better left floating. But in absence of a proper schemo, we can only speculate.
However, what is certain is that this is not proper practice, whatever the mfgr says.

 

[moamps]

.... But in absence of a proper schemo, we can only speculate....

We have detailed and complete schematic of the EQ in first post, and we know that the output of the converter looks as in picture below. I think that brains trust here have enough input data to decide which type of connections (cables) should be used, and second, to agree that without modding 6dB loss can't be avoided.

 

[Bo Deadly]

We have detailed and complete schematic of the EQ in first post, and we know that the output of the converter looks as in picture below.

Sorry I haven't been following this thread diligently but how do you know that circuit is the output of what? That circuit does indeed suffer from the "Tascam problem" which is well known and not to be used. But I sincerely doubt that UA or just about any vaguely respectable manufacturer would put that in any new product.

In fact, I don't recall seeing a UA or Urei schematic that ever used that circuit at all. The last time I saw that output on a new item was a cheap "Livewire" DI from Garage Center (and it didn't even have buildout resistors!). I had to hack the spit out of it to be impedance balanced and +-9V.

 

[moamps]

Sorry I haven't been following this thread diligently but how do you know that circuit is the output of what?

You described it in your reply 33.

That circuit does indeed suffer from the "Tascam problem" which is well known and not to be used.

What's exactly "Tascam problem"?

But I sincerely doubt that UA or just about any vaguely respectable manufacturer would put that in any new product.

This schematic is part of a Rane product, I used it only to illustrate how it was done in principle in UA2192.

 

[ruffrecords]

Not to me. I'm understanding 'impedance balanced' to be simply similar impedances on both 'L' and 'R' legs but no signal on 'R' (XLR Pin3  or R(ing) of TRS jack by current convention). So 'grounding' pin 3 simply bypasses the balancing impedance. There's no signal to lose ?

I agree. I do not know why I thought it was impedance balanced. Brain fart probably.

Bottom line the simplest solution not to lose 6dB is to add a transformer to the EQ input.

Cheers

Ian