diyAudio (
-   Solid State (
-   -   Balanced Driver / Receiver (

arigy 24th August 2009 10:13 AM

Balanced Driver / Receiver
Yet again question about best balanced driver / balanced receiver for, let's name it, high-end studio equipment. We have five options here:

1. Transformers
2. Burr-Brown DRV134/INA134
3. Analog Devices SSM2142/SSM2141
4. THAT Corporation THAT1646/THAT1246
5. High speed opamp based driver/High speed opamp receiver with input differential stage with something like 2SA970.

Let's exclude first option (transformers) from this list. It's quite hard to obtain high quality audio transformers here, and they are quite sensitive to overload. What should I use then?

Keith Taylor 24th August 2009 05:25 PM

The THAT Corp devices are the most advanced designs when it comes to real-world applications rather than test bench measurements. I am a little puzzled as to why you would choose the 1246 line receiver rather than the 1206 (InGenius series). While the 1246 may be a direct replacement for other manufacturers chips it does not have the high common-mode input impedance (via bootstrapping) that the 1200,1203 and 1206 have.

arigy 24th August 2009 05:41 PM

Well, I was thought actually that 1246/1646 complements each other ;) My mistake.

So, am i right, THAT devices actually performs better than AD/TI ones?

Keith Taylor 25th August 2009 05:59 PM

Under ideal conditions there would not be a lot of difference between THAT and AD/TI. The advantage of the THAT line driver becomes apparent when, (whether by design, or accident) one wire of the balanced pair is grounded. When this happens all the earlier designs have very unfortunate behaviour in that a lot of current with distortion components flows from the power supply into the grounded length of cable. This spiky current could crosstalk into adjacent pairs in the cable. This is something a transformer will not do because it has infinite common-mode impedance.
To truly emulate transformer behaviour a line driver would need to be powered from "floating" power supplies. Battery power is an example. The thing is you would need separate floating supplies for each line driver, which gets a bit expensive. The THAT 1646 is a good compromise in that it goes some way towards emulating a transformer by steering the input signal away from the half of the output that is being grounded thus getting around the heavy supply current problem. There are links to this subject on THAT's Web site. Similarly the high common-mode input impedance of the 1200 series InGenius line receivers confers advantages you can read about. When choosing which of 1200,1203 or 1206 to use, proffessional users would go for the 1206 which has a 6dB loss (differential to single ended) giving it the highest overload margin. Noise performance is not a consideration at line level. At the driver end of the bal' line the 1646 raises the level 6dB giving us a single ended to single ended gain of unity. Feel free to ask more questions or have me supply more detail.

Steve Eddy 25th August 2009 06:40 PM


Originally Posted by arigy (
Let's exclude first option (transformers) from this list.




arigy 26th August 2009 12:44 PM


Originally Posted by Keith Taylor (
To truly emulate transformer behaviour a line driver would need to be powered from "floating" power supplies. Battery power is an example. The thing is you would need separate floating supplies for each line driver, which gets a bit expensive.

So, for best effect I should use separate linear power supply for powering THAT 1646?

Keith, I see the example schematics in THAT1206/THAT1646 data-sheets, how are they? Is there anything special, like special decoupling technique or advanced filtering? I found ferrite beads in driver schematics in signal path, won't them affect sound in some unpleasant manner?

Keith Taylor 26th August 2009 04:14 PM

Arigy, what I was trying to say is that transformers are the "gold standard" for interferance imunity in a balanced line system. The reason for this is that the balanced pair of wires are not connected to anything and are free to assume any voltages that may be generated by magnetic fields along the length of the cable. These unwanted signals are called longitudinal or common mode signals and may even be higher voltage than the wanted (differential) signals. The magic part of balanced systems is that these signals get rejected because they do not appear[B]across[B]the transformer windings.
While transformers are good at rejecting unwanted signals they do have their own problems such as being expensive, bulky, cause distortion at high levels of low frequency etc.

I went on to mention that to emulate a transformer electronically would require floating power supplies. This remark does not apply to the THAT chips or any other chips. It is only a remark about what would be needed to get transformer behaviour without a transformer! I believe this has been done but I am not aware of any products in the market place.
You can power as many THAT chips from a single supply as you like and they are not particularly critical of layout/bypassing etc if you follow the application data. One critical factor is that the 1646 should be fed from a low impedance source which does not include the moving arm of a gain control. The ferrite beads would be for radio frequency filtering and have no effect at audio.

The Outsmarts and InGenius topology was invented by Bill Whitlock of Jensen transformers. At see applications/white papers/AES papers (last one, Bill Whitlock) for a description of the line receiver and many other articles on balanced lines. Also has plenty of discussion about THAT products. If you go to the "Build" forum, 2nd page you will see a thread called "1646 in a distribution amp" started by a DIY Audio member who migrated there at my suggestion.

snellwilco 30th January 2010 02:12 PM

problem with THAT 1206 versus INA134
I'm bumping a thread again rather than start a new one.
In my previous posts I talked about using an INA134, but, discovered on this forum that the THAT 1206 is thought to be a better chip.
(frame of reference...building a ESP P101, but want to add balanced inputs to it, because I have a noisy environment...)
I'm relatively new to building circuits - but have successfully completed a PIC based timing circuit kit and the Velleman K4700 without having to do any rework, not to mention the power supply which is ready to go too - but I've hit my first major roadblock. If you look at the THAT 1206 datasheet you will see the circuit is slightly different than a "standard" balanced line receiver like the INA134. I wired a breadboard for the THAT 1206 and could not get it to work at all. No sound. Double checked everything. I figured I'd test the INA134, so I made the slight mod to the breadboard config - remove the "bootstrap" capacitor between pins 5 & 8 and jump pins 5 & 6 - and the INA134 works! NB I followed all the usual precautions of testing voltages & polarities before installing the chip into the chip I don't think I hosed the first 1206 I bought. I figured, though, that maybe it got zapped by static and so I tried the 2nd one. Nope. Doesn't work either. (supply voltage is coming from the 12V wall wart that came with a Soundcraft it's not a complete POS)

Is it even conceivable that mouser sent out 2 busted versions of the chip? Has anybody else had a hard time getting that one wired up?

snellwilco 30th January 2010 03:34 PM

Quick update - I've just realized that jumping pins 5&6 would tend to produce output no matter what because of the internal topology of the INA134. it even works without a power supply. There's no noticeable difference in sound with or without. Does this mean my breadboard was wired totally wrong? There's a very very slight pop when connecting or disconnecting the power...thanks all. (I've also tried a 12V battery in case the power supply was at fault. It makes no difference.)
This is taking up way more time than I expected so I think I'm just going to use the 134s. But to a complete newbie, I'm trying to understand the situation with pins 5&6. Does 6 put out a signal that represents the "correction" to any noise-influence that comes through pin 5? (which I suppose is the XLR "hot" signal) Is that how the INA134 works? One signal path just looks like pin 2 (the input/"hot" pin) followed by some resistance (50k total). All that would do is lower the signal level slightly right? Obviously I have a lot to learn hahaha...thx in advance.

InfiniteGain 8th March 2010 05:59 PM

Sorry to jump in here, but the THAT1646 balanced line driver is what I am looking to build at the moment. Figure 8 of the spec sheet to be exact.

My issue is that although I know what a ferrite bead does, I have no idea how to choose one and I am not going to buy a plethora of them to tinker for an eternity.

There must be some way to calculate which ferrite bead is apt here. I read up on the "pin 1 problem" that might have given some insight, but the papers all just mention that a ferrite bead helps without being specific as to how to go about choosing one.

If anybody knows how to choose a ferrite bead that would sit on the -ve/+ve pins of an input XLR, let's have it! I guess we want to start attenuating frequencies downward pretty soon, but most devices seem targetted at 1Mhz and up!

I know I'm building equivalent-use circuitry to Arigy, but the ferrite bead topic should have application to anybody reading and may help both of Arigy and myself in the near term. Certainly, it would be good to know what beads to use for the input/output lines in both balanced and unbalanced circuitry.

(Oh... one caveat... no commentators saying it shouldn't matter if I'm not running huge runs of cable - I get Radio Moscow through my phono stage in the early hours of the morning and I'm in the UK).

All times are GMT. The time now is 05:43 AM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 18.75%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2017 DragonByte Technologies Ltd.
Copyright ©1999-2017 diyAudio