Driving transformers to play nice with audio interface line outs & line ins

[z11111]

I frequently mix using “Hardware Inserts” in Pro Tools through my audio interface / converters, so the chain is:

• AD/DA converter balanced line output > outboard gear (balanced) input > outboard gear (balanced) output > AD/DA converter balanced line input

However, I have some outboard equipment that doesn’t play nice or sound right when interfaced this way (say something like a Pultec -- tube stuff and/or other pieces with transformers...), so I’m thinking of a simple solution / "utility box" for interfacing outboard hardware between audio interface / converter AD/DA to avoid potential impedance issues.

My first thought was to make a line driver to go between the line outputs on the D/A and inputs on gear, particularly w./ transformers on the inputs. But am I overthinking this WAY too much?

1. 1st option I thought of would be to lay out a little board based on the data sheet for the THAT1606 line driver. Little tiny QSOP-16 package, what a blast to hand-solder. My brain went to the 1606 (rather than1646) because it already offers a differential input (e.g. balanced line coming out of audio interface), but does that even matter? Would a THAT1646 be fine?
   
   Found this old thread on some ideas for using the THAT1646 as a transformer driver: THAT1646 Current Booster and Transformer Driver - Pro Audio Design Forum
   
2. 2nd option was to use an API application note from the 2520 datsheet for a “balanced transformer-less amplifier” to drive a transformer (not unlike a 325, I guess), and/or just use something w./ transistor buffers like the Aphex/B&B “hybrid DOA” scheme to keep things simple and still be able to drive low impedance loads.
   
   
3. 3rd option, to piggyback off of the API idea, would be to lay out a differential driver into "a high-speed, high output-current buffer inside the op amp feedback loop", like this scheme using the LME49724 which "drives 600R loads with full output signal swing" but I'd use LM4562 (datasheet) because I already have those around, and a pair of LME49600 (datasheet) which are headphone buffers: Source / original link to the scheme: https://www.eetimes.com/product-how-to-differential-line-driver-with-excellent-load-drive/

Final option: just get some 600:600 transformers (or something) & wire to XLR pigtails to interface between the converters and the outboard gear & call it a day?



This thread was also helpful: Driving step-up transformer directly from opamp?



For background regarding the I/O:
I’m using an RME UFX+ and a Ferrofish Pulse MX (over MADI to the RME). Input impedance for the RME A/D line ins are: 8 kOhm unbalanced, 12 kOhm balanced. Output impedance for the RME D/A line outs are 75 Ohm unbalanced, 150 Ohm balanced. Ferrofish data sheet doesn’t show impedance specs but says it is using RC4580s on both the AD and DA sides. Output level is switchable between +4 dBu and -10 dbV (or -8 dBu, ish) on both units. Likewise, I'm over here assuming my issues are D/A, but could it actually be the A/D side of things?

We have Burl Mothership B80 as our main converters at our studio, which has transformers on both the input and output cards. I rarely run into any issues with "Hardware Insert" interfacing through the Burls.

 

[ruairioflaherty]

I likely won't have time to follow up here but before you engineer a solution I'd encourage you to really diagnose the issue.

Transformers are not a panacea for impedance issues, and if incorrectly deployed they can cause or exacerbate issues.

So the question to ask is why is certain gear giving you issues? There may not be one answer, it could be several distinct problems presenting (sometimes more than one at the same time).

What I haven't see you mention is drive capability of your interface D/A. A real pultec (or authentic clone) can be a very hard load to drive, from (fuzzy) memory I remember extreme settings presenting an input impedance of less than 100 ohms. This will draw more current than your interface can source.

 

[gyraf]

doesn’t play nice or sound right

We really need to know in detail what you mean by this

 

[z11111]

We really need to know in detail what you mean by this

Sorry, wrote so much other crap that I forgot about the most important details…

High frequency roll off, very noticeable discrepancies in volume and presumably frequency response — issues that don’t happen or alternatively, can be avoided depending on what order the problematic outboard equipment is patched (when using other hardware pieces in a chain) etc

 

[Script]

issues that don’t happen or alternatively, can be avoided depending on what order the problematic outboard equipment is patched

There's the clou. Now could get hold of and read some schematics.

 

[ccaudle]

Now could get hold of and read some schematics.

I would start with just the I/O specifications for each device, such as input and output impedance, and designed nominal and maximum signal levels.
That would go a long way to identifying problems.

 

[ruffrecords]

Most modern interfaces are designed to drive a 10K bridging load. Most are not really designed to be able to drive a 600 ohm load such as would be presented by the input of a Pultec EQ.

The drop in level it not surprising. With a pair of 75 ohm resistors in the outputs such as you describe, the level drop driving a 600 ohm load will be nearly 2dB. In addition, many common op amps are not designed to drive loads below 2K but the above load is 150+600 = 750 ohms so these op amps will be struggling, their headroom will be compromised and their distortion will increase. If they have insufficiently large output capacitors then the LF response will suffer as well.

As you correctly surmise, what you need is an output buffer that is genuinely capable of driving a 600 ohm load.

The THAT 1606 seems an ideal choice.

Cheers

Ian

 

[Brian Roth]

https://groupdiy.com/threads/600-ohm-inputs.47279/
Everything old is new again!

@z11111

To be bullet-proof, the THAT chips require a very low source driver impedance, so you really need a buffer stage as a driver. With that in mind, you could "lose" 6 dB with the driver and use a 1646 which is available in DIP-8.

Bri

 

[z11111]

https://groupdiy.com/threads/600-ohm-inputs.47279/
Everything old is new again!

@z11111

To be bullet-proof, the THAT chips require a very low source driver impedance, so you really need a buffer stage as a driver. With that in mind, you could "lose" 6 dB with the driver and use a 1646 which is available in DIP-8.

Bri

great thread, thanks for this

 

[ruffrecords]

https://groupdiy.com/threads/600-ohm-inputs.47279/
Everything old is new again!

@z11111

To be bullet-proof, the THAT chips require a very low source driver impedance, so you really need a buffer stage as a driver. With that in mind, you could "lose" 6 dB with the driver and use a 1646 which is available in DIP-8.

Bri

THAT1606 is 10K balanced in AFAIK so no need for special low Z drivers.

Cheers

Ian

 

[Newmarket]

THAT1606 is 10K balanced in AFAIK so no need for special low Z drivers.

Cheers

Ian

Some confusion ,?

From the datasheet

2. Both devices must be driven from a lowi mpedance source, preferably directly from opamp
outputs, to maintain the specified performance.

fwiw I DIYed four such channels to deal with unbalanced outputs from my converter box (grounded to earth PC) to an earthed mixer. With simple Z balanced outputs into diff inputs on the mixer. Not properly measured but ears and meters say big improvement.

 

[CJ]

Neve 8078 will handle all that gear

 

[kags]

We have a lot of young producers who come to our studio to run their track through the gear they’ve heard about- Pultec, 670s, StA Levels, 436s, etc…. Even one artist with Eurorack synths tries to go directly into our gear. They do t want to use the 9000J, just their gear patched into our outboard. I’m putting together some JML stuff to make an 8-channel unbal-bal and vice-verse using his tartigrade modules
https://www.jlmaudio.com/shop/tradigrade-kit.html?display_tax_prices=1
Should solve the 600ohm drive issues.

Anyone try these?

 

[Brian Roth]

I didn't see a schemo to verify the exact design. Many "unbal to bal" designs have the horrible Tascam Problem.

Bri

 

[ruffrecords]

Some confusion ,?

From the datasheet

2. Both devices must be driven from a low impedance source, preferably directly from opamp
outputs, to maintain the specified performance.

fwiw I DIYed four such channels to deal with unbalanced outputs from my converter box (grounded to earth PC) to an earthed mixer. With simple Z balanced outputs into diff inputs on the mixer. Not properly measured but ears and meters say big improvement.

Definitely some confusion. If you look at the 1606 equivalent circuit diagram from the data sheet it has 5K series resistors on each of the two inputs. That looks like a balanced input to me. But if you look at the pinout you see that what was labeled In- is now labeled Gnd so you can't actually use it as a balanced in/out device. So the OP is going to need a balanced in op amp with an low impedance output just to drive this thing. Hmmm, two chips where one would do really. SOme of these new fangled ships are too smart for their own good.

Cheers

Ian

 

[Brian Roth]

I just followed the data sheet for the THAT 1646 and have sold over a hundred of these:

http://recordingservicesandsupply.com/item/otari/output-driver-replacement-for-/lid=44716247
Works perfectly for the application. Never any returns or complaints.

The original Otari "hybrid" will die into an unbalanced destination when patched.

1646 is the "Bee's Knees" to do any sort of interfacing to 600 Ohm loads....balanced or otherwise.

Bri

 

[Script]

THAT124x into THAT1646i s almost a classic -- expensive though.

 

[Brian Roth]

Well....checked recent prices for line output transformers??? lol

Bri

 

[KA-Electonics.com]

THAT1606 is 10K balanced in AFAIK so no need for special low Z drivers.

Cheers

Ian

The 1606 still needs a "0Ω" driving source impedance. Any external source impedance is in series with the internal input resistors and becomes part of the OutSmarts® bridge network degrading noise and signal balance.

 

[abbey road d enfer]

Definitely some confusion. If you look at the 1606 equivalent circuit diagram from the data sheet it has 5K series resistors on each of the two inputs.

That's correct as far as the input is concerned, but any impedance unbalance, in particular resulting from one leg being driven by a different impedance, results in deteriorated CMRR, as Wayne mentioned.

That looks like a balanced input to me. But if you look at the pinout you see that what was labeled In- is now labeled Gnd so you can't actually use it as a balanced in/out device.

Actually, any input can be driven, even both of them, so it's really balanced.

So the OP is going to need a balanced in op amp with an low impedance output just to drive this thing. Hmmm, two chips where one would do really.

It's the same with SSM2141/42 and DRV135.

SOme of these new fangled ships are too smart for their own good.

The main market for these chips is OEM, where a zero-ohm source is existing in most cases. DIY'ers have no say in the matter.