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.

tedsorvino

Well-known member
Joined
Aug 1, 2018
Messages
74
Location
Athens, Greece
Hi everyone. I'm used on building more guitar pedal circuits but lately I've started working more on "xlr balanced " circuits.

So I would like to know a bit more on what modifications should be made on an unbalanced in-out 9v circuit (for example a simple ross compressor guitar pedal project) to be transformed to a balanced in-out one.

Is there a general rule on how to transform unbalanced circuits to balanced ones?

Can a differential line driver like DRV 134 or THAT 1646 (maybe it's not needed) be used at the output stage and maybe a couple of other opamps like NE5532 or NE5534 in the input stage along with a +9v -0- -9v power supply - what really defines the voltage supply needed for a project (most times something like a double rail 15vdc to 18vdc -with center tap- is used)?

Thanks in advance
 
If the voltage rails are sufficiently high for the circuit to operate properly then voltage (upto the maximum allowed) simply defines the maximum output voltage before clipping.
Depending on the circuit / ic and load the output can usually get to within a couple of volts of the voltage rails.
There may be minor differences in specification from the typical +/-15V or whatever on the data sheet but basically it defines the dynamic range.
 
Thanks Newmarket for the reply. It somehow clarifies (and also generates some more questions) the voltage rail amount. But what about the rest of my question? Any suggestion?
 
A THAT1646 can certainly be used to provide a balanced output and it has the advantage the the total voltage swing is nearly twice the rail voltage. But you so not need to do anything so complex to achieve a balanced output. A balanced output just needs to have equal source impedances in the hot and cold outputs. The voltages (referred to analogue 0V) do not need to be equal and/or opposite. In fact the signal exists only between the hot and the cold pins. The ground wire (pin 1 on an XLR) is not a ground but a screen. So one easy way to create a balanced output from an unbalanced one is to connect the amplifier output to the hot and connect the cold to analogue 0V via a resistor equal to the output impedance of the amplifier driving the hot pin. A very large number of interfaces use this technique to provide balanced outputs.

Cheers

Ian
 
Thanks Newmarket for the reply. It somehow clarifies (and also generates some more questions) the voltage rail amount. But what about the rest of my question? Any suggestion?

Simple and technically excellent method is to use THAT ics for input (Ingenius) and output (Outsmarts).
As Ian says for a balanced output you need only balance the output impedances of hot and cold signal legs.
Although you do, of course, have half the output signal amplitude compared to driving both legs. I don't know if that is an issue taking your voltage rails into account.
The cross coupled output circuits (eg the THAT solution) do offer an advantage by also providing CMRR into an unbalanced output.
I don't have figures to hand or know if it is a consideration for you.
One thing to ask - what signal sources are you thinking. I guess the input is expected to be a low impedance balanced output.
But I'd suggest also including a Hi-Z input (switchable or whatever) for electric guitar etc use as DI unit,
 
Is there a general rule on how to transform unbalanced circuits to balanced ones?

Usually you just add a balanced receiver circuit that has an unbalanced output to the front of an unbalanced circuit.
You need to understand the signal levels you want to handle on input, and the maximum signal level you want on the output.
The usual approach is to use +15V and -15V power supplies because that is a standard voltage range for op-amps, and can handle almost any signal level.

The ThatCorp receivers are very good quality.
This TI part is less expensive and with fewer external components at the cost of common mode rejection that is not as good.
TI INA1651 balanced receiver.

I have not used that specific TI part, but I built the same circuit out of separate components, a dual op-amp for input buffer, and an INA137 for the differential amp stage. The INA1651 combines the input buffers, the differential amp, and the input bias resistors all in one part.

One thing to consider with the '1651 device is that it is unity gain. With separate components you could use the '137 diff amp, which is -6dB gain, so in the case of symmetrical output swing feeding the input you can handle higher signal levels. That is a concern if you want to handle 24 dBu inputs, but if you only need to handle up to 20 dBu the '1651 can output that running on 15V rails.
You could just nudge up to 24 dBu if you use 18V rails, so it goes back to understanding what levels you need or want to handle.

For the output you can either use impedance balanced outputs as Ian mentioned, or use one of the That Corp drivers, or just buffer with a dual op-amp, one configured as a unity gain buffer and the other configured as an inverting unity gain buffer.
The dual op-amp approach really only has an advantage of increased output level swing from the symmetrical output signals. So again, back to understanding the maximum output level you need. If you only need up to +20 dBu then probably impedance balanced output is the best choice.

For best noise performance having all the circuitry run from the same power supply is the best approach. If for some reason the circuit cannot run from higher supply voltages you would have to use a voltage divider to reduce the signal level at the output of the balanced receiver, and amplify again at the output of the circuit (assuming you wanted professional line level in and out, as opposed to microphone levels for example).
There will sometimes be an advantage to padding down at the input anyway, e.g. if the circuit could generate a boost in signal level you may want to accommodate that up front, rather than making the user reduce the signal level to prevent clipping.

If some of the terminology doesn't make sense yet, the first thing is to get comfortable with dB conversions. That is a way of using logarithms to compare levels, and balanced interfaces are commonly referenced in dBu, which references to 0.775V.
For maximum input and output levels check the datasheets for the parts, they usually reference the maximum output level related to the power supply (e.g. Vs+ - 0.8V, Vs- + 1V would mean that the output can go within 0.8V of the positive supply, and within 1V of the negative supply).
 
Note that there are a few alternative solutions.
One is to use transformers. Good xfmrs are expensive and somewhat bulky, and the circuit that drives the output xfmr maut be capable of providing the low impedance and current requirements. For these reasons, they're not the preferred solution for pedals.
The previously cited monolithic balanced line drivers are excellent, but require bipolar power rails, and increase significantly the power draw.
An interesting alternative is the balanced-impedance arrangement, which can be easily retrofitted on many pieces of gear. However it depends on an output circuit that has a relatively low (about a few 100 ohms) and constant output impedance, which is not the case with many peadls, where the output is taken via a potentiometer.
 
Thank you everyone ( Newmarket, Ian- Ruffrecords, ccaudle, cpsmusic, script, abbey road d enfer) for the replies. I need some time to study and process all the suggestions and somehow put them in a "real world" practical level by comparing some schematics I've already worked.

Just to let you know that the THAT 1646 is currently out of stock in most of the EU market and i ordered from a cheap ebay chinese seller. The result was horrible (bad ics nearly destroyed my build and took me 2 days back just to find that the ics were burnt). Actually most times I order from aliexpress and ebay with no problem.
I replaced the THAT 1646 with the DRV 134 which is somehow easier to find in my country and the rest of the EU, with really good results. They are pin compatible and at least in my projects ( eq and compressor) they work as they should. Can't tell if there is any difference on sound quality compared with THAT.

Thank you once again.
 
I replaced the THAT 1646 with the DRV 134 which is somehow easier to find in my country and the rest of the EU, with really good results. They are pin compatible and at least in my projects ( eq and compressor) they work as they should. Can't tell if there is any difference on sound quality compared with THAT.

Much of the differences is around the CMRR etc performance with various permutations of noise / imperfection. Eg wrt source impedance imbalance and/or a short circuit on an output signal. So if you don't have a CM noise issue to start with any differences may be negligible.
You need to have a defined 'problem' to compare performance.
 
The ThatCorp receivers are very good quality.
This TI part is less expensive and with fewer external components at the cost of common mode rejection that is not as good.
TI INA1651 balanced receiver.
It's interesting to me that TI still won't use the bootstrapped input circuit, even though my patent on the InGenius (THAT 1200) circuit expired years ago. They've simply raised resistor values in a standard instrumentation amp, making CM input impedance 250 kΩ, which is still nowhere near the 10 MΩ of the InGenius. But it's admittedly better than the standard op-amp and four 25 kΩ resistors that pass for a balanced line receiver in the vast majority of the world's audio gear. The benefit of the extreme CM input impedance of InGenius, just as with a transformer, is that it keeps CMRR high even when used with sources having hundreds of ohms of imbalance (like an unbalanced output).

Interesting that, after my patent issued, I approached Analog Devices, TI, and others about licensing - but they all had a strong NIH (not invented here) attitude (as if to say "if we didn't invent it, it can't be any good). Maybe now they just don't realize that the circuit is public domain. As good as THAT's realization of the circuit is, I'd still like to see someone else make the IC ... perhaps at a lower price. A lot more noise would go away around the (analog) world!
 
I would guess that a balanced line receiver for audio is such a niche application in this decade that they don't think it's worth taking engineering time away from more profitable market segments to lay out a new IC and fab it.
 
It's interesting to me that TI still won't use the bootstrapped input circuit, even though my patent on the InGenius (THAT 1200) circuit expired years ago. They've simply raised resistor values in a standard instrumentation amp, making CM input impedance 250 kΩ, which is still nowhere near the 10 MΩ of the InGenius. But it's admittedly better than the standard op-amp and four 25 kΩ resistors that pass for a balanced line receiver in the vast majority of the world's audio gear. The benefit of the extreme CM input impedance of InGenius, just as with a transformer, is that it keeps CMRR high even when used with sources having hundreds of ohms of imbalance (like an unbalanced output).

Interesting that, after my patent issued, I approached Analog Devices, TI, and others about licensing - but they all had a strong NIH (not invented here) attitude (as if to say "if we didn't invent it, it can't be any good). Maybe now they just don't realize that the circuit is public domain. As good as THAT's realization of the circuit is, I'd still like to see someone else make the IC ... perhaps at a lower price. A lot more noise would go away around the (analog) world!
I had many friends at TI, they had this thing (and still do) where they hire engineers, create new products, then fire them with a 2 year paid non-comp, then sell for a while and do it all again. Of course when you call with very deep questions, no one know what the hell is going on. They are very "we didn't design it-go away".
 
I would guess that a balanced line receiver for audio is such a niche application in this decade that they don't think it's worth taking engineering time away from more profitable market segments to lay out a new IC and fab it.
I wonder just how easy it would be to implement it discreetly using a regular FET input op amp. I guess a significant element of the performance is due to the laser trimming of resistors plus their physical closeness so that they are at the same temperature.

Cheers

Ian
 
I wonder just how easy it would be to implement it discreetly using a regular FET input op amp. I guess a significant element of the performance is due to the laser trimming of resistors plus their physical closeness so that they are at the same temperature.
Doesn't even have to be FET input, the beauty of the bootstrap arrangement is that you can use a bipolar input device, use lower value input resistors, but the effective impedance is still high because of the bootstrap.
I am told SSL used the arrangement back in the mid-'80's, but the first time I saw it was in the AES preprint from the 1995 convention describing the converters used with the Sony Oxford large format digital console (which started out as an SSL design, which is why it had SSL input circuits).
There was a toss away line I didn't catch at first in the description for the input circuit that stated "Common mode bootstrapping...is provided. The purpose of this is to increase the dynamic input impedance presented to common mode and single ended inputs.
"This is necessary to provide a better simulation of a floating input, more similar to transformer isolation."

Pretty sure that would have invalidated the InGenius patent if anyone wanted to pursue that course. I asked the designer about that once, he said that the higher ups at Sony and SSL said that as long as they weren't getting sued over their designs they saw no need to rock that boat.
 
I wonder just how easy it would be to implement it discreetly using a regular FET input op amp. I guess a significant element of the performance is due to the laser trimming of resistors plus their physical closeness so that they are at the same temperature.

Cheers

Ian

Yes. But increased voltage noise cf bipolar input types is an issue with low impedance sources.
 
It's interesting to me that TI still won't use the bootstrapped input circuit, even though my patent on the InGenius (THAT 1200) circuit expired years ago. They've simply raised resistor values in a standard instrumentation amp, making CM input impedance 250 kΩ, which is still nowhere near the 10 MΩ of the InGenius. But it's admittedly better than the standard op-amp and four 25 kΩ resistors that pass for a balanced line receiver in the vast majority of the world's audio gear. The benefit of the extreme CM input impedance of InGenius, just as with a transformer, is that it keeps CMRR high even when used with sources having hundreds of ohms of imbalance (like an unbalanced output).

Interesting that, after my patent issued, I approached Analog Devices, TI, and others about licensing - but they all had a strong NIH (not invented here) attitude (as if to say "if we didn't invent it, it can't be any good). Maybe now they just don't realize that the circuit is public domain. As good as THAT's realization of the circuit is, I'd still like to see someone else make the IC ... perhaps at a lower price. A lot more noise would go away around the (analog) world!
NIH is alive and well. I have experienced NIH from both sides of the table. For several years while still working at Peavey I was the inside guy reviewing external product/technology submissions. Sadly I never saw one idea that I liked enough to pursue (maybe thats why I only did that gig a few years :unsure: ). I have approached other large companies with my ideas, including patented technology with zero success. Its just the natural order of how things are.
==
When I managed an engineering group one difficulty was getting my design engineers to use existing in-house solutions and not constantly reinvent the wheel. Design engineers never saw a design that they couldn't improve.;) I can relate to that, and you need to give the design engineers a win occasionally to keep them engaged.

JR
 
One word of caution using the THAT 1646 or the DRV 134, ALWAYS drive directly from the output of an op-amp. No resistor, no pot, nothing...
Is this why I see so many schems of instrumentatiion amp-based mic pres, where the IA drives an op amp, then that op amp dives the DRV? Could never understand why the IA couldn't drive the DRV directly.
 

Latest posts

Back
Top