Multiple pellets issue

[John Yun]

Hi everyone,

We are now in the pilot stage of using Fed3 in behavior experiments (very exciting!), but recently we noticed a peculiar problem in regards to multiple pellets. Usually, the motor would turn, wait a little bit, then turn again if the middle sensor did not detect any pellets. However, recently one of our systems went bonkers where the motor would turn, then turn again immediately without a small halt. Although this happens somewhat rarely (maybe once every 15-20 pokes) this would cause multiple pellets to emerge per nose poke in a FR program or even in the free-feeding program.

My initial thought was maybe the pellets were not detected by the middle sensor and thought of a way to maybe elevate the food well of the nose piece a little bit so it can be detected. However, while I was playing around with the system with this modified nosepiece, this one time the motor kept turning and turning without picking up anything. It ultimately dispensed ten pellets or so before it stopped (picture shown below), and I know it was not registered because the side screen said it was clearing a jam. At this point, could it be the photo-interrupter that's the issue? Hoping not because it's going to be tough to swap it out, if that is even possible. Would love to get some insights on this!

[Lex Kravitz]

Wow yes it seems like the pellet well did not detect the pellet.  A few ideas:

1) I would confirm that the center port is not detecting the pellet by manually poking something like a screwdriver into the well to block the photobeam during dispensing.  You should be able to see the little circle fill on the screen when the pellet well is blocked.  You can set it to free feeding for this test.

2) Is there any chance you're running this under IR or red lighting?  IR room lighting can interfere with the photobeam detection, as that relies on IR as well. 

3) You may want to update the FED to the latest software too, installation instructions here:

https://github.com/KravitzLabDevices/FED3_library/wiki

Let me know what happens with these, it is possible to swap out the photointerrupter but it isn't easy and will need to be done carefully if this is the problem.

-Lex

[John Yun]

Update:

So the center port is able to detect the pellet because when the pellet is in the well, the screen has the pellet circle shaded in. 

However, the multiple pellet per poke is still an issue. I tried different rooms and updated the library but there is no progress.

I was able to record a video of this issue this time. I originally thought the pellet in this video didn't fall in to break the beams which is why it turned again. However, if no pellet is dispensed, there should be a small hesitation (second video) before the motor turns again so now I'm starting to think maybe the beam isn't the issue.

[Lex Kravitz]

Thanks for the videos!  Good news that the center port can detect the pellet. 

Can you confirm which version of the code you're using?  The most recent version has a 1.5second delay between motor movements to give the pellet enough time to get down and break the beam to be detected.  In the video you attached showing two pellets being dispensed there seems to be almost no delay between movements?  I'm attaching a video of a FED flashed with the most up to date library so you can see how the dispensing timing is supposed to work.  This is version 1.3.31, flashed with the FixedRatio1 example.  Is this the version you have on yours?  

[John Yun]

Ok I think I see the problem here. So for the Fed3, I have been running the code from Fed3-1.1.48.ino because we have been playing around with the visual and the buzzer stimuli settings. After running from this code, going to FR1 setting I was getting these multiple pellet issues.

Going through the FR1 example from the recent library, I did not experience any of these multiple issues, which is good. 

Was I wrong in running the code from Fed3-1.1.48.ino to change the buzzer and lights? I am looking through the example codes of FR1 and classic Fed3 and I cannot see where I can change the stimuli setting there. 

[Lex Kravitz]

Glad that's figured out!  Does the double pellet stop happening when you flash the FR1 example program from the library? 

If you have your edits how you want them with v1.1.48 you can simply add a 1.5s delay after the dispensing code to slow it down and that may work for you.  But long-term you should use the library, it has several improvements to the code and is much easier to customize new behavioral programs.  You can see the functions you can call with the library here.

To keep this topic organized let's start a new thread if your want to discuss how to edit the lights and tones with the library?  If you can start a new message and explain what you're trying to get the FED3 to do I can coach you through it.  Thanks!

[Ross Rudolph]

Hi all!

We've recently been experiencing this same issue of multiple pellets dispensing at once during free feeding - on some FED3s, the pellet counter goes up by one even after several pellets dispense, while on others the pellet counter will accurately increase by the number of pellets dropped (however it is only supposed to dispense 1 at a time). There doesn't appear to be any rhyme or reason to it. One of them will begin dispensing pellets nonstop until it fills the chute, clogs, compacts with disintegrated pellet dust, and puts itself in a "clearing jam" loop. I have tried cleaning the sensors, updating the FED3 software, and testing out other programs, but nothing has worked. The FED3s will dispense several pellets even on other programs such as FR1 and FR3 as well. Any ideas why this might be happening/troubleshooting suggestions? Thanks very much!

- Kate Callahan

Ross Lab

Albert Einstein College of Medicine

[Lex Kravitz]

Hi Kate,

Is this happening with all of your devices or only some?

From what you describe, it sounds like the FED3 is not detecting the pellet.  That can be caused by the sensor malfunctioning or the pellet well (ie: the metal nose) not being positioned well to detect it.  To check if it's the sensor malfunctioning, if you put a screwdriver or something in the pellet well does that get registered as a pellet and stop if from dispensing?  If so it would suggest the sensor is OK, and it's a problem with the positioning of the pellet well such that the pellet is not landing within the beam.  Here's a photo of how the pellet well should be seated, as close as possible along the bottom:

One other idea - are you doing this under red/IR light?  Or under very bright lights?  I have seen FEDs malfunction in this case, as the pellet sensor uses IR and can get tricked by strong external IR.

-Lex

[Ross Rudolph]

Hi Lex!

Thanks so much for your response. This is happening on all 4 of our FED3s. The sensors seem to have no problem detecting a screwdriver - when I put it in the pellet well, the circle on the screen fills up and no pellets dispense, and when I take it out, the circle becomes empty and a pellet dispenses. We have our pellet wells glued on as close as possible, but the edges of the sensors do not appear as close to it as those in your photo do. I've tried tightening the arduino board so that it is closer to the FED3 body and pellet well, but it appears as close as it can possibly go. I have also tried the FED3s in a dim/dark room and still experienced these same issues.

I can try to take out the board and put it back in more evenly - perhaps it will set in more flush to the pellet well - since it doesn't seem to be an issue with the pellet sensor itself. Please let me know if you have any other suggestions. Thanks very much!!

- Kate

 

[Lex Kravitz]

Hi Kate! It sounds like the issue is that the metal pellet well is too far away, so it is not catching the pellet in the beam.  While it does not have to be completely flush with the photo-interrupter it should be very close. .  Here is a photo of this distance on a FED3 I have with me now that works, I'd estimate the gap is ~0.5mm. 

I would try to see if you can re-mount the metal nose piece closer, or even put some a small piece of rolled up masking tape or something in there to make the pellet stop within the beam and see if that solves it.  Please post back if that solves it - I suspect the issue here is that your prints are a slightly different dimension every 3D printer is different and it is critical that this part is aligned so the pellet stops in the IR beam.

[Ross Rudolph]

Hi Lex,

Thanks so much for all the advice. I got 3 of the 4 to pass the 200 pellet test using a little tape rolled up in the pellet holder! I am currently working on the last one, but I'm sure it just needs a little more adjusting. Can't thank you enough for all your help!

Best,

Kate

[Lex Kravitz]

You're welcome Kate!  If you are printing these housings yourself you may want to edit the model so the metal piece sits in there perfectly without tape - that will likely not be a good long term solution.  FED3 was designed in TInkerCAD and the editable design files are here:

https://www.tinkercad.com/things/0QaiVw7KR3Y

Let me know if you want a quick tutorial on how to modify the design, it might take a couple rounds of tweak-print-tweak to get it dialed in perfectly for your printer.  Best, -Lex