joakimsoderberg

There is no wait, it only locks if it finds some prey. Of course there are occasional false positives. But those are very rare (haven't compiled the exact statistics for that though).

The chase thing has happened, but since the door to get in is very small other cats don't tend to follow.

Since it's really cold here in the winter I also chose to design it so that the default mode is for the door to be open. So if there's a power outage it will always be open.

Regarding the weight thing. I don't think it's feasible unless your cat always brings in rabbits or something, the weight difference is so tiny. Too many variables come into play as well, temperature differences being one, a wet cat compared to a dry one and so on. Might work, might not :)

Yup she's done exactly that at times, not the "prop door open", but she dropped the prey right in front of the door, went away for a while, came back, picked up the prey and went inside... But I wouldn't say it was because of detailed scheming on her part, but luck rather :)

Another thing she has done is to learn to back out of the corridor as the door locks, and then wait a while, and try going back in all while having the prey in her mouth.

EDIT: So all small things like that has led me to tweak the algorithms, and add some extra settings to be tweaked. I mean the backing out move she did several times like 6 times in a row... Some of those times she managed to get in.

So to remedy that I added some settings for extra timeouts so she cannot try again as quickly and so on :)

To be specific. I only took the general idea from that project. That is, to match the cat head against a lit background to make it easy to get a good profile image.

None of the implementation has anything to do with that project however.

As I mention in the project I bought one of these: http://www.priority1design.com.au/shopfront/index.php?main_p...

Turns out the antenna that comes with it is way too small for anything other than manually brushing it against the cats neck knowing exactly where the chip is.

However I got some mails from another guy that built a cat feeder thingie. He ended up doing his own, much bigger antenna so that the cats head fit through it (he made a structure in cardboard wrapping the coil around that).

Anyway, he had some contact with the company, and they recommend using the RS232 version instead of the USB one for such applications... They had some better spec in some way.

EDIT: I looked up the mail, and to summarize he used 0.6mm enamled copper wire and a 120mm computer fan (minus the fan in the middle, that he wrapped the coil around. The reader then outputs the MOF value so you can know what frequency you're getting.

I've bought some copper wire myself and I'm planning to make a better antenna myself, but haven't yet.

Yes I'm living the dream of 90% test coverage! ... Well I used this project as an excuse to write testable C code, so that's why.

It's common I'd say. And I live in a small village with lots of room.

Here are some real images (including hard cases): https://github.com/JoakimSoderberg/catcierge-examples/tree/m...

I got lots more, just haven't uploaded them

Author of the project. Yes this happened, not with birds but with dead mice... And I did not realise it until like a week or so, not a pleasant cleaning operation.