Ok I found my answer to my stupid question on PSW so you can belay that request.

My next request is: Can you post the schematic for what you built?

That's if your taking requests today.

otisroom wrote:Here's a stupid question:

Why?

Why do you need something to unbalance Digidesign's 192HD converter outputs?

Because Digidesign can't design versatile output amps. The 192s are built with non-ground sensing balanced output amps. It is one of many new products out there that have regressed in technology, and flexibility.

Most transformerless balanced outputs use ground sensing circuits to allow it to connect to unbalanced inputs without compromising level, or fidelity (like a transformer). If either polarity of the balanced signal is grounded (like when you plug it into an unbalanced input), its opposite should step up 6dB in level, and provide an solid unbalanced signal to the input.
Non ground sensing outputs, like the 192s will not do this. The grounded polarity's amplifier will try to drive the ground with as much current as it can, instead of shutting off. This leads to a distorted signal, and crosstalk on your board. Very annoying.

A typical workaround for this problem is to make snakes that lift the negative polarity. This provides the shield and positive leg to the unbalanced input (-6dB though). You can't do that with the HDs, because the balanced signal is floating from Earth. In other words, the shield is not the ground reference for the audio. It is connected to the chassis, and ground pin of the power connector like it should be, but it seems that the "0 Volt" reference for the audio power supply is separate from Earth ground (which is not uncommon).

Certain "smart" bipolar power supplies will have a servo that centers the circuit ground symmetrically between the two DC rails so that the amps see the same +- voltage value. This provides the best overall headroom potential (imagine a power supply that gives you +19v DC and -17 v DC, the servo will make "ground" +1v DC so that the audio has the ability to maximize its headroom (in this example, having the ground at 0v DC would limit the negative excursion of the audio signal, by 1 volt). I am making an assumption that this is what is going on with the HD. I have no other explanation for the issues.

Anywho,
The shield coming from the HD is an improper ground reference for either pole of the balanced signal. When I tried lifting the (-) side, which is something they sheepishly recommend, I was left with an unusable level of clock/processor noise on the audio. The crosstalk was gone though!

For a while, only solution was to send the output through an i/o package that was professionally designed and manufactured to be flexible and great, our Studer A820. The irony was not lost on our customers when they saw our 540 lb., $80,000 machine being used as a through put for the Digidesign Tivo192HDs. One ridiculous day, we had to move a 24 track from studio A to B for this purpose, the other machine being used as the front end, tape sound-effect for PTools.
I built this so we wouldn't have to do that again.

Oh..........


Crap

Nice explanation Greg! Thanks!!
Digidesign are shit.

nice work too!

Thanks for that explanation.

Now I'm wondering, how do you get a 820 from studio A to studio B? Secret elevator? or intern hernia day?

otisroom wrote:Now I'm wondering, how do you get a 820 from studio A to studio B? Secret elevator? or intern hernia day?

The Heart of Gold

Thanks for that explanantion, Greg. I think I learned a few things there.

Out of interest, what would happen if instead of grounding the -ve signal to convert to unbalanced we tied it to ground via say a 10K resistor? Would this not stop the distortion? I suppose the dirty ground might still be a problem though.

I'm not sure I get where the crosstalk comes from. I'll need to think about that.

Rodabod wrote:Thanks for that explanantion, Greg. I think I learned a few things there.

Out of interest, what would happen if instead of grounding the -ve signal to convert to unbalanced we tied it to ground via say a 10K resistor? Would this not stop the distortion? I suppose the dirty ground might still be a problem though.

I'm not sure I get where the crosstalk comes from. I'll need to think about that.

The crosstalk comes from the ground plane of the console being driven by line level audio. As good as our Neotek's ground is, there is still a little bit of resistance to Earth from channel to channel. That allows there to be a tiny bit of audio on the console's ground. Imagine an extreme example, connecting the output of a power amp to the shield of an unbalanced console. You will hear the output of that amp on every channel, and monitor selection. The HD has 24 little power amps doing exactly that when the -ve is terminated to console ground. The crosstalk is not very high in level (maybe -50 or -60 dBu) but enough to ruin things. The distortion is coming from those tiny amps, which do have to-92 transistors for added current drive, trying to drive a <1 ohm load.

Adding that resistor might keep the -ve amps from working hard, and the voltage drop created by the low resistance to ground when it hits the console may improve the crosstalk issue. That still leaves you with the floating +ve and the improper reference of the shield as was the case when I lifted the -ve.

I wrote this fast so let me know if something sounds fishy.

otisroom wrote:Now I'm wondering, how do you get a 820 from studio A to studio B? Secret elevator? or intern hernia day?

Two ways: The chain hoist above the loading dock and a lifting strap, or I put on my kilt and girdle and huck it up there with my super-human caber-tossing skills.