The mouse gap squeezer teaser machine

Following my smallest hole for a mouse experiments last year, I still wanted to test how narrow gap (as opposed to round hole) a mouse could still get through. It's easy to adjust such a gap, and even better if the computer can adjust it. Lego is ideal for making this sort of mechanism, and if I use Lego, I might as well build the whole experiment out of Lego. I'm using my old (1st generation) Lego mindstorms kit, which I previously used for this pager rotating machine for my old job at Blackberry. And yes, work paid for the Lego.


The motor is controlled from my Raspberry Pi board, using a relay board that plugs into it. I added some limit switches to stop the motor when the sliding door reaches either end. There's no feedback for the door position to the computer.


A transparent acrylic cover keeps the top secure while still being able to see what's going on inside.


I set this up in a little shed next to the garage at our house. This isn't as good a location as the shed I used when we lived in the country, but there is at least one mouse that showed up after I left some cardboard with Nutella smeared onto it.

I made a short vlog video about setting this up.


A reader had sent me these point to point wireless antennas for getting internet to the big garage workshop in the country. I had set these up, but decided to keep using the old ethernet cable I strung to the shop at least until it failed, and it never did fail. But for this they turned out very handy because I didn't have a good place to string an ethernet wire and this is out of range for the house's wifi. I would have had to leave a window open a crack and string the cable across where the kids like to play otherwise.


I didn't figure out how my software should detect whether the mouse entered the box initially, so I first set it up without auto-adjustments.

Eventually I decided to detect "mouse in box" by just looking for a slight reduction in average brightness inside the box. This feature was easiest to hack into my imgcomp program that I used to run the experiments and record imagery.


I set it up so that the gate closed by about 0.3 mm after each time the mouse went into the box and left. The gap was never adjusted while the mouse was at all present, so as not to scare it off.

The smallest gap the mouse was able to get through was about 1.5 Lego studs. Lego studs are spaced 8 mm, so this worked out to 12 mm.


After it got too tight, the mouse tried to chew the gap open a bit, but those Lego bricks are much harder to chew than wood. I kind of figured I would be sacrificing a few Lego bricks to the experiment.


But that experiment wasn't as interesting as I'd hoped, so I made a more convoluted gap for the mouse to squeeze through.


The limit again was just under 12 mm, though 11 mm proved too tight. So at least for this particular mouse, I have a very accurate figure for the minimum gap that it will still go through.

Though shrews (another mouse sized creature) seem to be much more determined and more flexible based on previous experiments, but I didn't have one of those show up for this experiment. I strictly use wild mice outside who come and go as they please.


See also

My imgcomp program (motion triggered time-lapses)

To my Woodworking website.