balanced to unbalacend converter

[Onvinyl]

Hi,
I want to build a balanced to unbalanced converter. Walt Jung has a few examples, but I prefer to use a discrete design with or without (preferred) transformers.

It should be capabale to take a huge input range (line level), shall exhibit at least good CMRR, should sum both phases (instead of losing one).

I don't necessarliy need a ready-to-work circuit diagram, a concept how this would be done e.g. in recording studios would be a great input as well!
Rüdiger

[AndrewT]

Hi,
Nelson Pass' site has pre-amps that can be used as bal to bal, bal to unbal, unbal to bal, all from the same circuit.

He goes to a lot of effort in some of his schematics to exactly match impedances on the inverting and non-inverting inputs.

Oh, and he specialises in FETs, but I think BJTs could be substituted (don't tell John Curl I said so).

Borbely (also FET) has a complex (=expensive) balanced input pre-amp/line driver that can convert.

[boraomega]

Standard solution....

[AndrewT]

Hi Bora,
why is the output ground shown connected to the power ground rather than the signal ground?

How well do the capacitance multipliers work?

How close to 22V can the supplies be taken for maximum output levels?

[Steve Eddy]

Originally posted by Onvinyl
I want to build a balanced to unbalanced converter. Walt Jung has a few examples, but I prefer to use a discrete design with or without (preferred) transformers.

You say preferred without transformers, but you can do the whole thing with just a transformer and no active circuitry at all.

It should be capabale to take a huge input range (line level)...

Can you better define "huge"? 6.5 volts RMS huge enough?

...shall exhibit at least good CMRR...

125dB good enough?

...should sum both phases (instead of losing one).

Ok.

Here's something to consider from CineMag:

CMOL-2x600T2

At about $97US it ain't cheap, but it's one hell of a transformer.

se

[Onvinyl]

Hi, thanks for the input so far.

Yes, I know (some of) the pass and borbely circuits.
I should be more specific:
I recall I saw once a diff input pair set up for balanced input, (providing a good cmrr), and the balanced signal was then converted in unbalanced with the aid of a current mirror. But I don't know where I saw it or how it might work. If Borbely or Pass have it on their respective sites, I apologize.

@Steve Eddy: that looks indeed promising. I have to do some calculating concerning input range. Price is a concern, since I have many parts right here including some fancy FETs

@boraomega: I have something similar in a mixing console I borrowed. It works ok, but I guess one can do better (and I have to give it back).

[boraomega]

-why is the output ground shown connected to the power ground rather than the signal ground?

As you see, this is just a SIM software drawing; hence, only one ground... that program "know" just for GROUND!

-How well do the capacitance multipliers work?

To my taste, that is always nice and recommendable addition for low current - low power - small signal stages. That work satisfactory for me... others don't have to agree....

-How close to 22V can the supplies be taken for maximum output levels?

I was never a follower of the "squeeze it to the last drop" approach when components are in question. Even if one IC is capable to operate at +/-22V, I would never forced it above say 18V. NE5534 is well capable to produce 10V RMS on 600 ohms load with 15V supply, and any unit following this stage must be really ill conceived if it is asking for more signal!

[janneman]

Originally posted by Onvinyl
[snip]@boraomega: I have something similar in a mixing console I borrowed. It works ok, but I guess one can do better (and I have to give it back).

Possibly, but not with a differential pair-to-single ended!

Jan Didden

[Steve Eddy]

Originally posted by Onvinyl
@Steve Eddy: that looks indeed promising. I have to do some calculating concerning input range. Price is a concern, since I have many parts right here including some fancy FETs

Understandable.

So what's your ultimate goal here? Are you looking to adapt a balanced output so it can be used for a number of unbalanced inputs, or are you basically only dealing with a particular unbalanced input? If the latter, another option would be to essentially balance the unbalanced input using something like their CMLI-15/15B which is a bit less than half the cost of the CMOL I mentioned.

However you proceed, good luck!

se

[Onvinyl]

Originally posted by janneman



Possibly, but not with a differential pair-to-single ended!

Jan Didden

Jan,
I'm thinking towards a complementary dual differential first stage (sk389/sj109) along the lines of curl/borbely/pass. I'm not sure about the input swing capability, though.

I will try those various implementations that have been discussed on this forum.

Do you think it's not worth it topologywise or or are you talking of a too simple diff-pair without CCS and cascode?
Rüdiger