FED not registering pellet being removed

[Claire Foldi]

Hi Lex, 

We are having an issue with a couple of our devices... annoyingly slightly different issues but, based on the thread below from a few days ago (FED3 display/program freezes) we might be looking at a main circuit board problem. Just wanted to check in before reaching out to OEPS. 

The first device says there is a pellet in the magazine when there is not - we have tried cleaning the magazine sensors and pellet well with ethanol and/or compressed air, reflashing, changing SD cards to a "known functional" card and switching out the Adalogger boards with a known functional board. No love whatsoever. 

The next device will deliver a pellet, but won't register the pellet being removed!! We didn't know what was going on for a while but with the new code that counts pellet retrieval time on screen in milliseconds it became clear that the pellet removal was not registering. This one surely can't be an issue with the magazine sensor because it detects a pellet being dispense but not retrieved. HOW?? We have tried all the above solutions with this bad boy too... And neither seem to be a sleep mode issue as discussed in the other thread here because it happens from the first poke/pellet when the FED is turned on (i.e. before sleep mode could have been enabled).

Any suggestions? 

@LauraMilton do you have any more details to add to this thread about the specific issues?

/C

[Lex Kravitz]

Hi Claire,

Do you have a multi-meter?  It doesn't have to be fancy, there are lots for ~$20US online that work just fine.  If you have one you can do some more advanced troubleshooting to track down the specific issue with each FED3.  When I find time I'll record a video on how to do this kind of troubleshooting with a multimeter.

But some ideas for you - The first device sounds like it could be a bad photo-interrupter sensor (this part).  The photo-interrupter uses an IR LED and IR detector to sense the pellet.  Relative to ground, this sensor will read 3.3V when the IR light from the LED hits the IR detector (ie: when there is no pellet) and 0V when something (the pellet) blocks the beam.  So if this sensor malfunctioned (either by the LED burning out and not giving off IR light anymore, or the IR detector malfunctioning and not returning 3.3V when open) the FED3 will act like there is always a pellet in the well.  So that could be the explanation for what you're seeing.  If you have a multimeter you can read the voltage between the output of the IR detector (red circle) and ground (green circle) like in the photo below.  It should read 3.3V when the photo-interrupter is open and 0V when blocked.  I suspect yours will always read 0V, which would explain the behavior you're seeing. Unfortunately it's not that easy to remove and replace this sensor from the PCB, but it is possible.  Confirming the issue is the first step!

The second device sounds really mysterious!  I would make sure the battery is fully charged, sometimes odd things can happen with a low battery voltage.  I would also try to explicitly turn sleep mode off (put fed3.disableSleep(); right after fed3.begin(); in the void setup() loop). FED3 will start using sleep modes right when it boots up even before the first poke, so it may make sense to just turn this off to be sure it's not somehow entering sleep and not leaving (although if the retrieval counter on the screen is working that's unlikely to be the issue).  Did you also try swapping over a known good Adalogger board to localize the issue to the PCB vs. Adalogger?  I'm stumped on how this sensor could detect the pellet entering, but not when it's removed! 

Hope this helps somewhat,

-Lex