Automatic Litterbox Central

http://www.youtube.com/watch?v=XImzPzCiAyM

Video shows it in operation.

What litter are you using???

Nice...so you have an electronic sensor? Won't it be affected by the urine, sani-solution, or the poop slurry?

To explain further, this is the second evolution to my electronic solid-state water sensor mod (see previous post on first generation).

The Cat Genie uses an IR emitted and sensor LED which are looking down the two sides of the long plastic water sensor that is notorious for getting dirty and causing the dreaded 3 beep error. If you take your plastic sensor out and shine a small flashlight down one side, you will see that it reflects a large portion of the light back to you on the other side. This is becuase of the two 45 degree angle cuts in the bottom of the sensor, the light reflects off their surfaces and it comes straight back up at you. If you then dip this end into some water, you will see that the water disrupts this reflection and the light that is reflected back to you goes out. There is only a few millimeters of effective transistion area on this plastic water sensor, so the Cat Genie doesn't really sense 'level', just whether it is full or less than full. Simply shining your own IR light into the CG's sensor allows you to replace the plastic sensor with some other solution.

My first solution used a Darlington pair of transistors driving an inverting output for an IR LED that I mounted directly under the CG's sensor. MY water sensor was a pair of small copper wires stuck down into the hopper in the same location that the plastic sensors goes. The unit was powered by a 9v battery. This worked fine for a day until the copper water sensor corroded over and the IR LED never went off. I came home to a box REALLY full of water, because the CG was constantly told that the box was empty.

This solution has the following improvements: AC power supply, red/green LED on circuit board for status, stainless steel sensor electrodes, and sensor placement under litter bowl. The circuit still uses a Darlington pair, but then drives two other transistors, one for the green and IR LED and one for the red LED, one inverting the other of course. I noticed that the green LED stays on barely when the red one is lit, it didn't do that on my breadboard but not a big deal. The placement of the sensor under the litter bowl should remove it from the seriously poopy water for the most part, and the stainless steel hopefully won't corrode.

Only time will tell if this solution will be viable.

I'm using the regular CG pellets, but probably will get some of the ebay ones.

I put the sensor under the bowl this time, instead of in the hopper where all the poopy goodness is at. Should really only be urine under the bowl. Hopefully the stainless steel screws don't corrode over. If they do I will try to see what metal they use in bathroom stalls, those have to be able to stand up to airborne urine (ammonia, etc.)

Still working good today. It has run about 4 cycles since I have installed it. BIG BIG difference in the amount of smell generated by the CG during drying cycle. If you are experiencing the three beep problem, then VERY likely the water sensor is dirty'ing in the first or second fillup, which means that the CG skips the second and/or third fillup, or sometimes I have seen it skip the first fillup and run the whole cycle dry. The stench created is just.... indescribable. With this, not only is there no constant BEEP BEEP BEEP... BEEP BEEP BEEP... but it doesn't stink because it is giving all three fillups every time. Plus if you wanted to change how far it fills up, in case you need more water, you can just move the probes up a bit higher to make it fill up more, just can't move them up too high.

You could probably go with some platinum coated probes...just don't be trying to steal my "cat" converter!

I can't imagine how much platinum connections would be. I tried to find stainless steel nails, but apparently Lowe's doesn't have those, though I'm pretty sure they make them for untreated wood fences and such, they don't stain the wood over time.

Would anyone be interested in these if I made it into a more professional solution? I am already working on a pcb layout to make a better board.

Here's some wire...

Any hardware store should have stainless steel bolts...maybe even some rods or wire. It's best that you ask them so they can git you what you need.

Just so you know...there are different grades of stainless...

I'm always interested in advancing the technology...so I'd be interested in what you come up with. What are the costs so far?

That platinum wire is pretty high. I think that a good stainless should work pretty well. I just got back from a weekend away and *gulp* left the CG running, whereas I normally put down a regular litter box and unplug the CG when we are gone. We have, before installing this, had no confidence in the CG running properly while being gone, even for a single night. Came back with no beeps, no errors, no odor, no mess. It is a CG utopia at the moment.

So far, have probably $10 worth of parts, but that's RadioShack prices. The IR LED is $2 (ouch I know, but gosh darn it, its so convenient to just run in there and get parts), 4 2n2222 transistors at about $2 (ouch again...), handful of resistors from a variety pack (negligible in price), two colored LEDs (probably $1), some wire that I have piles of, and two stainless screws (pack of 5 was $.84). That doesn't include the power supply that I have lots of, dunno what they cost standalone.

I could probably build these with etched PCBs (instead of the predrilled proto board I have) and sell them for $30 + shipping. If there is interest then I can make some.

ku4zs wrote:Would anyone be interested in these if I made it into a more professional solution? I am already working on a pcb layout to make a better board.

I'm a big proponent of open source/open hardware so a schematic for fellow tinkerers would be nice- maybe post something on Instructables? There seem to be a few CG mods there..

I would be interested in purchasing one...let me know how much

zeropanic wrote:

ku4zs wrote:Would anyone be interested in these if I made it into a more professional solution? I am already working on a pcb layout to make a better board.

I'm a big proponent of open source/open hardware so a schematic for fellow tinkerers would be nice- maybe post something on Instructables? There seem to be a few CG mods there..

I will be making a schematic to post, right now I just have my paper scribble. I will look into a detailed posting on instructables as well. I am already preparing to make some etched pcbs so that I can produce complete kits for those who just want to buy a solution. I am not rushing into that quickly for the simple reason that I want to make sure that what I have continues to work well. So far there has not been a single hiccup and the unit has worked flawlessly.

In the meantime, instead of a schematic, if you want a wordy description of the circuit...
12vdc wall wort supply (though it is giving 19.5v ?). The two probe connections under the bowl are connected (each) to an 8.2k resistor, one resistor to +19.5 and the other to the base of a Darlington Pair transistor set. This Darlington Pair is pulled high on its collector by 8.2k to 19.5v, and its emitter connected to ground. The two other transistors are connected through 82k resistors to the collector of the Darlington Pair, connecting to the base of the other two transistors. One of the transistors has the following wiring... 19.5v ---(1.0k resistor) --- (Green LED) --- (IR LED) --- (Transistor collector).(Transistor emitter) --- gnd. The other transistor is connected as... 19.5v --- (3.0k resistor) --- (Red LED && Transistor collector). (Red LED) --- (Transistor emitter && gnd).

I don't think I missed anything there, that was from memory. There should be 4 transistors and 7 resistors.

The 12v dc wall-wort is giving the high voltage beace it is not filtered and the load isn't high or not there at all when you measure it. Plaing a big capacitor across it to smooth out the voltage will bring it down.

2cats2dogs wrote:The 12v dc wall-wort is giving the high voltage beace it is not filtered and the load isn't high or not there at all when you measure it. Plaing a big capacitor across it to smooth out the voltage will bring it down.

Ok, that explains it. I tested two different ones both saying 12v and both put out 19.5v.