Audio Adapter + stereo mode performance debugging?

[Andy-205]

On initial use of the audio adapter I find that when the input is set to stereo mode (by the jumper) and the supplied TRS cord is inserted into the connector that considerable noise is generated on both channels. The noise is persistent with the cord driven, open, or shorted.

Each channel works fine separately in mono mode. 

Representative images are shown below.

The small signal on CH2 of the mono screen shot is "signal leakage" as only one cord was connected...

Any guidance on trouble shooting and repair or just return?

 

 

[attila]

Hi @Andy-205

Make sure to enable the power supplies, see:
https://digilent.com/reference/add-ons/audio-adapter-plus/getting-started
https://digilent.com/reference/add-ons/audio-adapter-plus/start

[Andy-205]

Thanks for the response.  I should have clarified that the distortion is on the input channels (displayed on the scope) when connected to an external source.  I've used the head phone jack on my phone and also a portable (battery powered > "floating") audio recorder/player.  In these cases no wavegen or output connection was active/used.

Power on/off does not change the behavior.

The schematics seem to indicate that the sleeve/shield connections of the input jacks (J1, J2) are connected to pins 28 and 30 and to the main ground through a 470 ohm ferrite bead.  The captured waveform "looks like" ground may be an issue.  Can proper grounding be verified by application of a "multi-meter"?  If so, please detail.

I've checked resistance between the shield contacts on J1 and J2 also to pins 28, 30, and 26, all are open.

(If this is a grounding issue perhaps the title can be changed to include "audio adapter grounding requirements" etc.)

Edited December 31, 2023 by Andy-205
typo

[Andy-205]

It turns out that the jack shield connectors are not connected to the board ground!

As noted previously, shield J1 to shield J2 measures open!  That "can't be right", can it?

Both are also open to pins 28 and 30.... again.. not correct is it?

(Pins 28 to 30 measure as "common" as expected)

If I jumper from the J1 shield to pin 28/30 the board performs properly!

This is a "crazy" failure.. a problem with the connection from the solder pads to inner layer on the board?  If so, I'd guess many boards might be effected. 

Has this been reported by others?

If I can verify that all functions are normal with the grounds connected I may try to repair by "exposing" the ground plane to allow direct soldering to the jacks w/o using jumper wires.  Is this a 3 layer board so I can access the main ground trace from either side or do other traces exist (on additional layers) near the J1/J2 shield pins that I need to avoid?

(If easy to fix I'd rather do that than return but you may have more bad boards to look for...)

Let me know what you think... I may post a clearer comment to better help others.

[attila]

Hi @Andy-205

The audio/scope inputs have separate ground (GNDA) for better isolation from the rest of the adapter.

 

The Wavegen output amplifier is powered by the V+/- power supplies channels and the ferrites are for filtering.

 

[Andy-205]

Thanks for your attention.  I had found those 2 schematics.  Unless some detail is not shown, the resistance between the sleeve/shield contacts of the scope input jacks (J1 and J2) should be low/zero.  Same for the GNDA header pins 28 and 30, all 4 of those points should be common.

I find no connection between the J1 and J2 shields/gndA and "anything".

I've made the jack connection 3 ways; via the shield/sleeve on an inserted cable, with a probe in the jack, via probes on the solder pin at the board.

(The shield of a plugged in cord is common to the solder pin of the same connector so know the cord and jacks are fine.)

Also no connection between either shield and either pins 28 or 30.

I do find a connection directly between pins 28 and 29... indicating that they are properly soldered to the common "GNDA plane/trace"

A bright light inspection of the board seems to reveal a sub-surface "GNDA" plane under the jacks. 

The solder connections (at the surface of the board) of the J1 and J2 GndA pins look fine... this suggests a "via" failure on both the jacks.

That seems unlikely to be a completely isolated occurrence. Has this been previously noted?

Do I have a new run? (DB9E1A0, REV C, PB200-418)

It might be possible to repair by heating and "reflowing" the solder on the shield pins... drawing it deeper into the board.

If not, I think perhaps the sub-surface GndA trace/plane might be exposed and the missing connections added.

Please comment on the best repair approach as I'd rather repair than return (if practical).

Too many words, see the image.

 

 

 

 

 

[attila]

Hi @Andy-205

The board has 4 layers: top, bottom, GNDA(left)|GND(right), V-(left)|V+(right)
Pins 28-30 should reliably connect to TP1-3 since these are on GNDA, a large internal plane on the left side.
I only have revA and B prototypes, so I'll have to ask my colleagues to check this next week.

[Ionut]

Hi @Andy-205,

Would it be possible for you to solder on the board a wire from J5 pin 30 to J1 GNDA pad (TP1), and another wire from J1 GNDA pad (TP1) to J2 GNDA pad (TP3)? This is indicated in the image below with yellow arrows. Could you then run your test, check if the board works fine and let us know the result?

Thank you!

Best Regards.

[attila]

Hi @Andy-205

The solder shapes looks good but you could try to resolder TP1/3, add a bit of new tin/flux to it.

[Andy-205]

I went ahead and added the jumper wire and it seems that everything is working properly now.

Certainly the input channels are working w/o excessive noise.

Let me if you have any specific tests that be used to "completely" verify performance.

Otherwise I think we can consider this case solved.

Thanks for you help in this matter!