Openpandora has no sound


Glenn Doiron

Still Fresh
Joined
Apr 19, 2019
Messages
4
Symptoms: Openpandora audio stopped working a few weeks ago. Occasionally, produces static. Just found the time to look at it. First checked the audio jack, no sound from it either.

Checked U12.1, digital signal present.
Checked U12.2, no signal present -> Checked U21.19, no signal present (as expected) -> Checked U21.5, digital signal present (1.8V peak-to-peak with abundant over/undershoot).

Ordered a new SN74AVC8T245 (likely culprit) and PCM1773PW (just in case the bus buffer isn't the problem.)
 
Inb4: You take it appart and its a Software problem.
Make sure its not because starting with a first post claiming to take it appart leaves me with the impression that you have no idea what you are doing.
Flashing/SDbooting is the first step for every problem.
 
sound_issue.png

Here's the circuit in question, along with my observances on my oscilloscope. The SN74AVC8T245 is not under software control. It is being used as a buffer, and OE* is hardwired to ground , and DIR is hardwired to VCC. Therefore, I expect all signals at port A to be output to port B. There is also the possibility that there is a short to ground inside U12, which is why I also ordered a PCM1773PW.

sound_issue_2.png


The datasheet for the PCM1773PW indicates that this particular line is an input only, and cannot be misconfigured by software as an output. Therefore, software CANNOT be responsible for the attenuated signal. In fact, the signal could be completely flat, with what I was seeing being crosstalk from an adjacent signal line.
 
I didn't even mention the misspelling or the errant capitalisation OR the missing apostrophe; I am no grammar nazi. I do withhold the right to be pedantic though.
 
To attenuate the tension in this thread, I present you a video with dense information and high yield.
 
Just FYI, my cheapass Chinese heat gun did not get the IC hot enough to extract it, and preheating the board was not an option (nubs are on the bottom side and would probably be melted with a board preheater). This is less the fault of the heat gun, more the problem with having copper plane layers to soak up the heat and dissipate it. I ended up carefully sawing through all the legs which did not have a trace exiting from them or connected to an obvious visible copper pour (with a hobby knife); this allowed me to remove the IC with the heat gun. Wick up the pads with braid, clean them up with alcohol, and install the new IC with no-clean flux solder.
 
Back
Top