DCX2496 I/O mods, need some advice

I have completed the printed circuit board design for trying out Gary Pimm's input/output mods to the DCX2496, http://home.pacifier.com/~gpimm/behringer.htm

I would surely be willing to either share the design, or offer a group buy if anyone else is interested in some circuit boards. The boards are currently set up for Lundahl LL1676 on the input, and LL1690 on the output.

Before I order the prototype board, I have a question regarding the grounding layout. Perhaps someone (even Gary ?) has had experience with 2496 I/O mods, and can offer a suggestion.

The following link shows the IDC connectors from the ADC/DAC board to the I/O board: http://www.dcx2496.fr/tweakanalog8.php

As can be seen, there are multiple AGND (analog ground) pins on BOTH connectors. I checked the resistance between the pins from one connector to the other, and there is continuity. On my board, I am taking the +/-15V and AGND pins from the 20 pin connector, as this would seem most appropriate for powering the opamps. Wouldn't seem prudent to take 15V off one connector, and return current off another.

With the 26 pin connector, there is the signal input and output, along with more AGND pins. So it would seem most appropriate that the bias voltage of 2.5V be referenced to these AGND returns, as they are probably tied into the ADC circuit.

Then the question becomes do I tie the two AGND inputs together on my circuit board? This would technically be a ground loop, of course, which is why I bring up the question. If I don't, I fail to see the engineering prudence in using an opamp which is powered by one AGND reference, and generating a bias voltage referenced to another ?

Dunno, anyone have ideas ?

Thanks.
 
zigzagflux said:
Perhaps a simpler question:

Does anyone have a schematic of the ADC/DAC board for the DCX2496 ? I could then answer my original question.

Thanks !


There is a complete set of schematics in the files section of the yahoo DCX2496 user group. There is also *some* info in my article for AudioXpress about the passive output board, as well on my website on the remote control / active output board.

Jan Didden
 
So, kind of what I expected. They are tying the AGND signals together at both ends, and at least as far as the schematic shows, there is no attention paid to avoiding ground loops.

O well, I'll just proceed as my original intention. Most likely with the xfmr coupling on input and output, the affect will be minor. My printed circuit board is set up with a jumper, capable of fitting a 1/4W resistor, to tie the two AGND signals together. I can experiment with various resistor values to see if there is any improvement.

At least the ADC is a differential input.
 
zigzagflux said:
So, kind of what I expected. They are tying the AGND signals together at both ends, and at least as far as the schematic shows, there is no attention paid to avoiding ground loops.

O well, I'll just proceed as my original intention. Most likely with the xfmr coupling on input and output, the affect will be minor. My printed circuit board is set up with a jumper, capable of fitting a 1/4W resistor, to tie the two AGND signals together. I can experiment with various resistor values to see if there is any improvement.

At least the ADC is a differential input.


I use 1 ohms in my remote control I/O board...
Works OK for me. I had some noise (probably smps) before that.

Jan Didden