Water sensor error and possible solutions?

[Sonagi]

I wanted to post this to an old thread because I am still using Beta
2, but sent it only to Robert. Hopefully someone might find it
hopeful.

1. Fast blinking water sensor red LED error:
I have experienced a similar error to what Dan for a few times these
past two weeks on Beta 2. The Cat genie fills up with water but fails
to drain but blinks red led. In the red error state, the water is
filled and
the bowl is rotating, but the program fails to drain.. so the cat
genie is stuck in this mode essentially. This state persists for more
than an hour or so.. or even more, until I manually run the dry/drain
mode by pulling out/plugging back the power while holding the clean
button.

2. Possible Solutions: I cleaned the water sensor module then the
error was less frequent but not gone. When I replaced the water sensor
with a spare one (also previously used but cleaned and some candle
wax), the
error is mostly gone.

My custom detergent mix (dish washing liquid + oxyclean + water) might
have caused the problem as the oxyclean detergent has solidified and
has settled into the bottom of the detergent tank (thereby dispensing
detergent with soap sediments).

Another possible cause would be lower water pressure (the apartment
did some recent plumbing work and water pressure has gone down
somewhat). Also plumbing work might have caused rust to build up in
the water inlet hose and this might have cleared naturally.

I will keep an eye on this, but for now, I have not encountered this
error for about two days.

Cheers,
Sonagi

[Sonagi]

A similar error was originally discussed in the later posts of this
thread:
https://groups.google.com/group/catgenius/browse_frm/thread/eaeb43b60541731?tvc=1&q=turn+off

[Sonagi]

I am having very little luck in solving the fail to drain/constant red
LED blinking error. I full disassembled and cleaned the catgenie, the
water sensor, and used the original unrefilled cartridge. Now I am
getting the error almost without fail during initial water filling of
the cat genie. Touching the water sensor with hand seems to break the
error momentarily until the next water refill. I am sure it's not my
water sensor as I tested my spare, tried using WD 40 to waterproof,
and also tried using vinegar to remove any mineral buildup. My water
pressure seems normal. So I guess there is something with my water
sensor LED.

I have flashed to Beta 3 and installed the serial converter (which
seems to work) when I connected to an old desktop that has a serial
port. Now, I will try to get a USB serial connector and use my laptop
to log the catgenie when the error happens (cannot really move the
desktop to the catgenie location while it's hooked up).

Is it possible to make the Catgenius work with a faulty water sensor
LED (without causing catgenie overheating or flooding water)?

Thanks :D

On Sep 13, 3:33 pm, Sonagi <ashuny...@gmail.com> wrote:
> A similar error was originally discussed in the later posts of this

> thread:https://groups.google.com/group/catgenius/browse_frm/thread/eaeb43b60...

[Robert Deliën]

> I am having very little luck in solving the fail to drain/constant red
> LED blinking error. I full disassembled and cleaned the catgenie, the
> water sensor, and used the original unrefilled cartridge. Now I am
> getting the error almost without fail during initial water filling of
> the cat genie.

So you have an almost 100% reproducable scenarion? I find that rather good news, because it gives us a chance do diagnose the problem.

> Touching the water sensor with hand seems to break the
> error momentarily until the next water refill. I am sure it's not my
> water sensor as I tested my spare, tried using WD 40 to waterproof,
> and also tried using vinegar to remove any mineral buildup. My water
> pressure seems normal. So I guess there is something with my water
> sensor LED.

That's a bit strange: Both a faulty IR LED and a residual drop of water cause a false positive detection of the water sensor. But both are unique causes: Removing the drop of water will not cure a problem cause by a faulty LED. So there must be something I'm missing in my diagnose.
Can you make a photo of the IR LED using a digital camera while it is switched on? You force the IR LED to be constantly on by loading GenieDiag, detach the water and use GenieDiag to turn on the water. You will need to trigger the safety switch at the bottom on the processing unit with a plastic (non conductive) object. Be careful if you close the tap instead of disconnecting it: Pressure in the hose may still produce a squirt.

> I have flashed to Beta 3 and installed the serial converter (which
> seems to work) when I connected to an old desktop that has a serial
> port. Now, I will try to get a USB serial connector and use my laptop
> to log the catgenie when the error happens (cannot really move the
> desktop to the catgenie location while it's hooked up).

Can you tell me how many time the error LED flashing in a single second if this error occurs? This will tell me what the washing program is waiting for.
If you have a serial cable attached, it is better to use B02 to diagnose this problem, because it will output every step the washing program makes.

> Is it possible to make the Catgenius work with a faulty water sensor
> LED (without causing catgenie overheating or flooding water)?

I'm afraid not. But replacing it isn't much of a problem.

[CatGenius]

Ok guys, I'm starting to have water sensor troubles as well. First in
a wet box, but now they're also appearing even before any water is in
the box.

I suspect a dim LED to be the cause. I briefly checked this yesterday
evening with a camera, but since this can only diagnosed properly with
GenieDiag installed I will check again. But the LED did appear very
weak, so I'm afraid you guys have to replace the LED first if you have
any water sensor problems.

I will report the results tonight and possibly replace the LED.

[patrick]

As I mentioned before, I have had absolutely no problems with my model
60 running beta 2 after I removed and replaced the LED. I suspect the
dim LED causes the trouble on majority of the problem machines. I
used an LED from radioshack that came with an receiver(I didn't use
the receiver). After the fact, I had someone with a spectrograph
measure the color of the old LED, and if I remember correctly, it was
nowhere near the same wavelength as the replacement, but it didn't
matter. It still worked.

If you decide to replace the LED, be sure to use a camera and mark the
LED so you don't accidentally remove the IR receiver instead. Then
you'll want to disassemble the processing unit and remove the LED.
The PU is a real pain to take apart. The heater duct really isn't
made to be removed, and I almost broke mine. It's glued in place and
also held with some plastic tabs. The LED itself is held in with a
soft glue or silicone. I pried out some of the glue with a pick and a
knife, and then pressed the LED up and out from the bottom with a torx
driver. I then cut it off, cut off the heat shrink, desoldered the
legs, slid on the new heat shrink, soldered in the new LED, and hot
glued it in place. It's worked fine since then.

Due to the different wavelengths of various IR LEDs, your new LED may
appear brighter, dimmer, or the same as the old one when you look at
it with a digital camera. Mine looked the same through my digital
still camera, but *much* brighter through my cell phone camera. If
your old LED is dim, the new one will be brighter somewhere in the IR
spectrum and should solve your issue.

[CatGenius]

> As I mentioned before, I have had absolutely no problems with my model
> 60 running beta 2 after I removed and replaced the LED.  I suspect the
> dim LED causes the trouble on majority of the problem machines.  I
> used an LED from radioshack that came with an receiver(I didn't use
> the receiver).  After the fact, I had someone with a spectrograph
> measure the color of the old LED, and if I remember correctly, it was
> nowhere near the same wavelength as the replacement, but it didn't
> matter.  It still worked.

Yes, and we haven't forgotten about that, but I wasn't able to relate
all problem so far to the faulty LED. Now I have collected a lot of
logging around this problem I have a better insight. When I checked
the LED yesterday, it turned out to be very dim indeed, more or less
proving the theories.

With the results of your measurements, I went to my local electronics
hero and got a couple of standard LEDs. They didn't have much of a
choice in wavelength: 950nm seems to be the standard. And because the
LED is closely positioned above the light guide it seemed save to go
for a 10 degree opening angle. So a 5mm 950nm 10degree LED is what you
need.

> If you decide to replace the LED, be sure to use a camera and mark the
> LED so you don't accidentally remove the IR receiver instead.  Then
> you'll want to disassemble the processing unit and remove the LED.
> The PU is a real pain to take apart.  The heater duct really isn't
> made to be removed, and I almost broke mine.  It's glued in place and
> also held with some plastic tabs.  The LED itself is held in with a
> soft glue or silicone.  I pried out some of the glue with a pick and a
> knife, and then pressed the LED up and out from the bottom with a torx
> driver.  I then cut it off, cut off the heat shrink, desoldered the
> legs, slid on the new heat shrink, soldered in the new LED, and hot
> glued it in place.  It's worked fine since then.

Ha, those good'ol torx; I read this only after I was done, but I used
a torx outside it's intended purpose too for this job: Because of the
flat end, sometimes even with a little dimple in it, it seemed very
suitable to remove the LED. I have put it's end on the LED lens and
carefully knocked it inside using a small screwdriver as a hammer. The
black goo holding it in place seems to be hot-melt glue, which rarely
forms a strong bond with any material surrounding it. I didn't notice
a plastic tab, not even in the remains of the hot-melt I removed from
the LED later. There's probably some small differences between boxes.
From there I did the same: Desoldering, resoldering, heatshrink and
hot-melt glue. Reassembling and all works fine.

I didn't remove the fan, though. I opened the processing unit by
removing the rear shell. I disconnected all connectors from the
controller and then I could remove the front shell as well. There's
little room to work with there, but it is possible.

Using a camera before removing the LED is definitely a must. Since it
doesn't matter which one is in which hole, they may differ from box to
box. My LED was connected to the blue lead on the cable. That's
probably identical on all boxes, but checking never hurts, while
replacing your sensor with another LED will ;-)

> Due to the different wavelengths of various IR LEDs, your new LED may
> appear brighter, dimmer, or the same as the old one when you look at
> it with a digital camera.  Mine looked the same through my digital
> still camera, but *much* brighter through my cell phone camera.  If
> your old LED is dim, the new one will be brighter somewhere in the IR
> spectrum and should solve your issue.

I have put both LEDs in series, connected them to a power supply in
current-control and drove them with 15mA. I have made photos of them
side-by side, which I will post later. On the first photo's there's
little difference, but in a matter of seconds the original LED becomes
very dim. Even when looking straight into it's front.

[Sonagi]

Hi guys, just a little update on my problem. After ordering another
usb-serial connector (after losingmine) and waiting for more than 2
weeks (as it was coming from China), I was able to fix my problem,
which happened to be entirely my error. I dreamt of removing and
reconnecting the water supply line (a Watt's dish washer hose with
safety shut-off) to my cat genie. I woke up and did it in real life,
and the water pressure was restored to fill the Cat genie to not
trigger the timer error. Somehow when the Watt's water safety valve
did not fully shut off, but only restricted the water flow to the Cat
genie (slowing the fill rate by half).

I should have checked the water supply line before (as I remember
reading about water timer in one of earlier threads).

On the other hand, what do you think about the $35 water sensor kit
sold by one of litterbox posters? Would this kit work with the Cat
genius firmware?

[CatGenius]

> Hi guys, just a little update on my problem. After ordering another
> usb-serial connector (after losingmine) and waiting for more than 2
> weeks (as it was coming from China), I was able to fix my problem,
> which happened to be entirely my error. I dreamt of removing and
> reconnecting the water supply line (a Watt's dish washer hose with
> safety shut-off) to my cat genie. I woke up and did it in real life,
> and the water pressure was restored to fill the Cat genie to not
> trigger the timer error. Somehow when the Watt's water safety valve
> did not fully shut off, but only restricted the water flow to the Cat
> genie (slowing the fill rate by half).

I'm glad you have found a solution and it's great to hear you took the
effort of understanding the problem.

> I should have checked the water supply line before (as I remember
> reading about water timer in one of earlier threads).

Yeah, well, I suppose we all should something.

> On the other hand, what do you think about the $35 water sensor kit
> sold by one of litterbox posters? Would this kit work with the Cat
> genius firmware?

I'm afraid the electronic water sensor won't work with CatGenius
because it uses the water sensor a littler differently from the
original firmware. I still need to update the water sensor background
page in the pages area of the group. But since Google will soon drop
the pages and files functionality of Google groups, I'll probably look
for an alternative first. (I wonder what the were thinking?)

Anyway, the alternative water sensor is a good idea. Though the
implementation in CatGenius should already improve the existing water
sensor accuracy a lot. It also extends the LED life time: Where the
original software leaves the LED on 24/7, hurling 15mA through it's
guts, CatGenius only the pulses the LED: Currently 10 times a second
for a period of somewhere between 75us and 3ms. This reduces heat
buildup and extends the LED life time.

I'm afraid the $35 electronic water sensor will produce problems of
it's own over time because the water directly gets into contact with
the electrodes. The voltages on the electrodes will cause an
electrolysis process, causing one electrode to be eaten away and
buildup of non-conducting dirt on the other one.
A capacitive sensor would be a lot better, or an optical sensor, but
we've already got one of those ;-)

[Robert Deliën]

> I have put both LEDs in series, connected them to a power supply in

> current-control and drove them with 20mA. I have made photos of them

> side-by side, which I will post later. On the first photo's there's
> little difference, but in a matter of seconds the original LED becomes
> very dim. Even when looking straight into it's front.

Here are the two photos. Because the opening angle of both LEDs is relatively small and the camera is very close, I couldn't make a photo looking straight into both LEDs at the same time.

The first one shows both LEDs side by side and a multi-meter reading the current through both LEDs in series, so both LEDs are operating on the same current. This photo is taken straight into the new (right) LED.

The second photo is taken straight into the original LED. The difference may look a bit worse on the photo. When I was fiddling with the camera, the difference seemed much smaller.