DAC37J84EVM: Data converters forum

Artem B

Part Number: DAC37J84EVM

Hi everyone

We are using DAC37J84EVM with removed output transformers with TSW14J56.
On the first picture there's a digital signal that goes to DAC.
On the second picture there's IOUTAP output and on the third picture there's IOUTAN output.
"Compliment Output Data" is disabled for both DACs.
Why are the signals look inverted?

over 1 year ago

0 Chase W over 1 year ago

TI__Mastermind 21455 points

Hi Artem,

The output of this DAC is differential, so the IOUTAP and IOUTAN will be 180 out of phase with respect to each other at all times.

The complimentary output setting will swap the positive and negative halves of the differential output signal per that specified channel. This will allow for easier PCB layout and routing options.

Regards, Chase

0 Artem B over 1 year ago in reply to Chase W

Prodigy 40 points

Thanks Chase
But according to signal that I send from FPGA, it looks like P and N should be vice versa.
Please look carefully at the pictures that I attached.

0 Chase W over 1 year ago in reply to Artem B

TI__Mastermind 21455 points

Artem,

This is likely due to the layout on our EVM. I imagine the balun is what is causing the polarity inversion. Do the signals swap if you toggle the complimentary output bit for IOUTA?

Thanks, Chase

0 Artem B over 1 year ago in reply to Chase W

Prodigy 40 points

Yes, they do. Just wanted to know if it is the correct DAC behavior.

0 Chase W over 1 year ago in reply to Artem B

TI__Mastermind 21455 points

I imagine the DAC itself is functioning fine and the balun is causing a 180 phase shift. Can you probe on each side of the balun to confirm if the hardware is causing this inversion?

0 Artem B over 1 year ago in reply to Chase W

Prodigy 40 points

As I said, we removed all the transformers from the schematics, because we needed DC, so I don't have the baluns on the board. So that's the DAC outputs themselves. If you can measure the signal before and after baluns and send me the results, that would be great.

0 jim s over 1 year ago in reply to Artem B

TI__Guru* 91947 points

Artem,

I just checked this on our EVM and you are correct, the signals are inverted. I am checking with design to see if it was designed this way or maybe the ball assignment is incorrect in the data sheet. Are you seeing this on all four channels?

Regards,

Jim

0 Artem B over 1 year ago in reply to jim s

Prodigy 40 points

Hi Jim, thanks for your attention.
Yes, they look inverted on all four channels.

0 jim s over 1 year ago in reply to Artem B

TI__Guru* 91947 points

Artem,

The designers that worked on this part are no longer part of our group. Due to limited resources it may take awhile to find out if this is truly an issue with the data sheet. For now, if these signals need to be swapped on your application, I can only suggest to use register 0x01 bits 4:7.

Regards,

Jim