Balancing an unbalanced circuit

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
And I'm still haunted by one, perhaps rhetorical, question about it. The superiority of THAT1646 over other ICs is that the dumping error current in the ground is very small or absent in the case of a short circuit of one of the balanced outputs. Why is this so important when it is also proclaimed that in balanced systems currents and voltages in ground connections are not part of the useful signal anyway (pin1 - problem and solution drama)? What do we care if there are any currents flowing through the shield?:cool:
 
For 1646, LTspice says:
64.7dB CMRR with zero-ohm source, and 13.2dB with 2.5k Zs.

Regarding 2142, I get 185(!)dB with zero source and 18dB with 2.5k. Of course these figures should be seen in the context of an idealized circuit, with zero-tolerance on resistors, but I think the degraded figure is about correct.

I'll look for a Spice model for DRV134. Actually the diagram in the datasheet is absolutely identical to SSM242.
 
And I'm still haunted by one, perhaps rhetorical, question about it. The superiority of THAT1646 over other ICs is that the dumping error current in the ground is very small or absent in the case of a short circuit of one of the balanced outputs.
My understanding is that it minimizes current spikes that happen with other designs.
Why is this so important when it is also proclaimed that in balanced systems currents and voltages in ground connections are not part of the useful signal anyway (pin1 - problem and solution drama)?
Indeed, but not being useful does not mean they are harmless.
What do we care if there are any currents flowing through the shield?:cool:
If there are "currents flowing through the shield", they have an origin and a destination. When the origin and destination are not strictly coincident, they pollute other circuits.
 
I have always been of the opinion that these ICs (SSM2412, DRV134, THAT1646) were invented not to get a solution that works superior to transformers, but to get a simple solution that would be good enough to replace a transformer and much cheaper and much less weight and much less space to take up.
I tend to agree. There was a definite jump in perforlmance when PA equipment went from unbalanced to half or fully balanced (half-balanced meaning the outputs are unbalenced and the inputs differential). It resulted in an instantaneous hum reduction of about 30dB.
When additional hum reduction became necessary (going from 1kW rigs to 10+kW rigs), transformers became almost mandatory. That's when clever designers tackled the problem with cross-coupled circuits. However, the improvement was not that significant because of the tolerances in resistors in discrete designs. Only when monolithic solutions appeared, the level of performance became compatible with 100+kW rigs.
DIY'ers may not be concerned, but one must not neglect the ease of implementation of these IC's, compared to the number of resistors needed in discrete solutions.
 
These dedicated I/O chips were developed to sell silicon...I have told this story before, I was working for a large target customer late last century when these chips were new. I had a short list of criteria for these guys to fufill. Inside Peavey they were not competing with audio transformers but common multiple IC op amp circuits that were already in wide use in house and worked adequately. Some of these medium performance dual op amps cost around $0.12 ea, resistors were often less than a penny. We had automatic insertion equipment so not much labor saving available. Peavey used lots of audio transformers for customers who don't accept substitutes (like the instal market).

The only reason to change to these parts would be a demonstrable benefit. On my short list was the ability to run from unregulated rails with enough built in voltage gain to clip those higher rails. Further the output drivers need to suppress power on/off transients (perhaps with a mute pin). Ultimately they must be cost effective compared to existing solutions.

The last dog and pony show I attended in Meridian last century with several TI factory and FAEs in attendance barely acknowledged my request list. Instead they pushed their already finished designs. I understand the concept of selling what you got, but why promote the meeting as an engineering fact finding visit. ;)

At some point these may get cheap enough, and junior engineers inexperienced enough that these canned solutions make sense. It didn't last century when I was still managing an engineering group.

JR
 
Using TINA with the DRV124 model, I get 80dB with zero-Z and 20dB with 2.5k.
Thank you very much for your effort. Obviously all three ICs are roughly equal in terms of CMRR when not optimally driven.
If there are "currents flowing through the shield", they have an origin and a destination. When the origin and destination are not strictly coincident, they pollute other circuits.
True. And I'm not very happy that all three ICs are highly referenced to the signal ground. In this sense, transformers are still superior, IMO.
 
And I'm not very happy that all three ICs are highly referenced to the signal ground. In this sense, transformers are still superior, IMO.
I may disagree. The CM impedance of a fully functional IC is about a few kiloohms.
The CM mode impeadnce of a xfmr is dominated by the capacitance between windings.
With a typical value of 5-6 nF for a quadfilar xfmr, the CM impedance falls below 4K beyond 7 kHz, and falls at 250 ohms at 100kHz.
This becomes a cause for unstability. It's a perfect case for a parallel LR between the stage's output and the primary. I discovered that when I made two versions of a 500 mic pre.
 
Last edited:
As with everything in life you have to look at all the benefits or issues surrounding what you are trying to achieve. For a manufacturer making lots of 'product' the ability to buy a load of ICs and solder them in then forget is very attractive. Like evolution in general, you stop trying to 'improve' when you get to a certain level. Transformers used to be the universal 'high point' when mains hum was the largest threat to interconnection integrity, and arguably they could return as long as you accept the very high frequency and low frequencies are compromised if looked at in an 'audio purist' sort of way, but they are an effective way of avoiding modulated microwaves (mobile phones) getting into your gear. The search for 'perfection' is all well and good in an accademic sense but it then can leave a sense of 'disappointment' if you buy a hundred dollar piece ofgear and you read that the balanced input or output doesn't have a CMR exceeding 140dB measured from 'DC to light'.
Like a long running thread on a 'studio gear discussion site', noting which gear uses certain A/D D/A chips it is about as relevant as noting the 'serial numbers' on railroad cars/locomotives.
 
The CM mode impeadnce of a xfmr is dominated by the capacitance between windings.
With a typical value of 5-6 nF for a quadfilar xfmr, the CM impedance falls below 4K beyond 7 kHz, and falls at 250 ohms at 100kHz.
This becomes a cause for unstability. It's a perfect case for a parallel LR between the stage's output and the primary. I discovered that when I made two versions of a 500 mic pre.

When I said that IMO transformers are still superior in that post, I meant a situation where one of the IC outputs is shorted to remote ground (unbalanced receiver ground). If that remote ground is not perfectly at the same potential as the ground of the trasmitter (IC) ground will the output voltage of the IC be increased exactly 6dB and without any interference (voltage difference between the two ground)? The transformer is in that sense IMO superior for sure.
Your example of a transformer with a quadrifilar winding is not suitable because such transformers that have a huge capacity between the windings should not be used for connection over long distances or for signal transfer to different places (for example mic splitter to FOH - Monitor MIX - Recording room). Such transformers are used IMO for local signal transfer where it is of the utmost importance that the signals are equal in value and where a different ratios is desired.
Take a look at the datasheet from Lundahl LL1517, with the optimal way of use, no IC can match it IMO, especially in extreme conditions.


1656013409229.png

And Lundahl transformers are not the best IMO that can be find, if for a member Newmarket THAT ICs are state of the art, for me they are some WSW Siemens transformers.
 
When I said that IMO transformers are still superior in that post, I meant a situation where one of the IC outputs is shorted to remote ground (unbalanced receiver ground). If that remote ground is not perfectly at the same potential as the ground of the trasmitter (IC) ground will the output voltage of the IC be increased exactly 6dB and without any interference (voltage difference between the two ground)?
I'm not sure I understand your question. There's plenty of evidence that cross-coupled circuits and remote ground sense circuits adequately remove the ground differential from the recovered signal. Adequately is not perfectly, but neither transformers do that.
The transformer is in that sense IMO superior for sure.
Your example of a transformer with a quadrifilar winding is not suitable because such transformers that have a huge capacity between the windings should not be used for connection over long distances or for signal transfer to different places (for example mic splitter to FOH - Monitor MIX - Recording room).
Maybe they shouldn't, but I have sold many channels of mic splitters for OB vans and remote recording, using multifilar windings and nobody ever complainesd
Such transformers are used IMO for local signal transfer where it is of the utmost importance that the signals are equal in value and where a different ratios is desired.
Take a look at the datasheet from Lundahl LL1517, with the optimal way of use, no IC can match it IMO, especially in extreme conditions.
Just like one could conclude that the THAT, SSM & DRV IC's are a solution in search for a problem, the same could probably be said for this type of transformers.
Extreme conditions are the sign of lack of preparation. It should not be assumed from the start.
 
Maybe they shouldn't, but I have sold many channels of mic splitters for OB vans and remote recording, using multifilar windings and nobody ever complainesd...

I don’t think further discussion on this makes sense, everyone has put forward their point, the audience will come to their own conclusion anyway.
 
Back
Top