Beta two results
DanITman
as beta 1. This time I did get an error message. 2 blinking lights
indicating water sensor error again. Decided to unplug and clean
sensor. Sensor had some water on it but not much. Plugged unit back
in but the CG did not do anything. I thought it would continue where
it left off but nothing happened. I decided to unplug and then hold
down the start button and plug it back in. That started a cleaning
cycle to get me back to dry. That is working right now and it looks
like the pump is running and blower is running as well.
I still would like to get this through one full cleaning cycle.
I'll work this weekend to get the serial interface up and running. I
should be able to give you a better idea of what is happening.
Thanks
Robert Deliën
My sensor is pretty dirty now, and my box still works. It's way beyond the amount of dirt the box used to fail with in earlier versions. I'm not cleaning it now because I want to see where it fails. Jaime's and Cassio's boxes work fine too, so I'm really curious to learn why your box fails so consequently.
CatGenius will not continue where it left after a power interruption. That would be really hard to implement because we would have to store the context of the program: The program state, all motor states and all timer states. This is basically a pause function. For power interrupt recovery we would need to store this context a couple of times per second, which is rather bad for EEPROM life expectations.
Instead CatGenius has 3 box states:
1.
Box tidy
2.
Box messy
3.
Box wet
This state is stored in EEPROM and checked when CatGenie starts up. In the first state no action is taken. In the second state the cleanup program is started in dry mode, which will just surface the granules. In the third state the box is drained, dried and surfaced.
When a program starts, the state 'Box messy' is written to EEPROM. When a program opens the water vale, this state is overwritten with 'Box wet'. If a program finishes, it is overwritten with 'Box tidy'.
I don't really see how your box failed to start the cleanup program. Debug information will help there. Fortunately holding the start button at power-up did succeed to start the cleaup program.
Yes, debug information would be very helpful there. If needed I can build special versions for you to diagnose the problem a bit deeper.
________________________________
DanITman
great. I can fill the box up and it will auto shut off once water is
detected. Red light appears. I then activate the pump red light will
go away and water turns back on. Everything seems to be working when
I use geniediag.
I'm going to work on getting the serial installed to start pulling
logs.
> 6KViewDownload
Robert Deliën
> sensor works great. I can fill the box up and it will auto
> shut off once water is detected. Red light appears. I then
> activate the pump red light will go away and water turns back
> on. Everything seems to be working when I use geniediag.
That's weird. Because GenieDiag uses the same code as CatGenius. Or better: They share the same files implementing support for the hardware. GenieDiag is basically CatGenius with a different user interface and without the LitterLanguage interpreter.
But there may be a clue in here. Your water sensor basically works fine, but at some moments it starts to report falsely detected water. I have inspected the water sensor with a mirror, when my box was in this state and I discovered that just touching it, with my hand in the hopper cured the problem. This lead me to believe that a residual drop of water must have been hanging from the detector. This usually didn't occur right after draining, but a little bit later. Either some water was still splashing around, or some water ended up on the ceiling of the hopper or even above it, dripping down to the water sensor over time.
I thought this was mainly caused by me, running washing programs without soap. Soap breaks the surface tension of water, reducing it's capacity to from drops. Since I have inserted an original cartridge I rarely saw the problem. The remaining rare cases were solved by the new led-muting algorithm in B02.
Are you using any kind of cleaning agent in the box?
I'm not suggesting this as a permanent solution, but could you try so rub the bottom of your light guide, especially the edged parts, with the side of a candle, to make it a water repellent? Make sure there are no lumpy wax residues left. I would like to know if this improves things. In that case I need to make an adjustment in how the water sensor is used.
But before that, perhaps you could do a couple of tests with GenieDiag to see how the problem behaves:
1. Fill the box with water
2. Drain for 20 seconds
3. wait for 20 seconds
4. Drain for 20 seconds
5. wait for 20 seconds
6. Drain for 20 seconds
7. wait for 20 seconds
8. Drain for 60 seconds
9. wait for 60 seconds
After the first 3 seconds of draining, the Error LED should turn off, indicating that now water has been detected, and it shouldn't turn on anymore. Especially in the longer periods, step 8 and 9, you might see it come on if hanging drops are the cause.
> I'm going to work on getting the serial installed to start pulling logs.
That's always a good idea: The software logs what the program is doing, including waiting for the water sensor. But it also logs changes is sensor state. So that could pinpoint the cause more accurately.
Cheers!
Robert.
DanITman
second one right now.
When I was installing the Serial Port I took the time to wipe the
LED's off on the bottom. I'm not sure if this helped at all but I
didn't change anything else. Maybe taking it apart and back together
helped out, not sure.
Robert Deliën
> Working on a second one right now.
Personally I'd even classify that as GREAT news ;-)
> When I was installing the Serial Port I took the time
> to wipe the LED's off on the bottom. I'm not sure if
> this helped at all but I didn't change anything else.
Perhaps. But it really shouldn't be that sensitive.
> Maybe taking it apart and back together helped out, not sure.
I'm thinking about building a special version, that logs analogue values as well. We'd have to temporarily sacrifice to LEDs for that, and your serial port needs to work, but it will teach us something about the mechanics behind this problem.
DanITman
I'll follow up with you on the Serial port stuff, ran out of time
yesterday.
DanITman
have not had time to do any CG testing.
I found something very interesting today with my box.
I tend to get the error LED more often when I have Genie Dome on the
box then I do with it off. When I have it off the cleaning cycle
never fails when I put it on it almost always fails. I'm not sure why
this would make a difference but it does.
Is anyone else having trouble when you have the Genie Dome on?
Thanks
>
> - Show quoted text -
Robert Deliën
Welcome back!
That's actaully not that far fetched. After all, the water sensor is an optical sensor. And unlike the cat sensor, it is not able to distinguish it's own light from foreign light. With the dome off, more environment light will be shed into the hopper, where the water sensor is, than with the dome off.
My cats refuse to use the box with the dome on. But my garage, where the boxes are, is usually dark so it should be quite similar to having the dome on. I'm trying to recall the light conditions when I'm running a washing program, but I don't think I leave the light on for that.
Perhaps your water sensor LED is close to the burned out problem and some environmental light will help it to work, but I think that's a bit far fetched.
I will do a couple of more measurement on the water sensor, to figure out if CatGenius uses it the same as the stock firmware does, but after I have migrated my SBS2003 server to SBS2008 today ;-)
Cheers,
Robert.
DanITman
got the error 2 light again. This time instead of the box being stuck
it just continued on like the error didn't exist. I'm not sure why
this is.
Robert Deliën
I'm really wondering what's causing your water sensor problem. I still think something is wrong with your LED of CPU pin. Is anybody else in this group experiencing problems with the water sensor?
One of my original boxes has a water sensor problem too. I will try to diagnose what's causing the problem and figure out a way to verify if your problem have the same cause.
The behaviour you've observed isn't strange. The washing program has several stages in which it waits for the water sensor to detect water (or not). When the waiting starts, a timeout is started as well. When the desired state is achieved, the timeout is destroyed and the program continues. However, if the timeout expires before that time, an error is reported. This means a blinking error LED on the panel. Besides that, the running washing program should be aborted and the appropriate cleanup program should be started. But this latter part isn't implemented yet. So instead the program keeps on waiting for the desired state, way past the timeout, until the sensor finally reports the desired state. On a false positive, this can take from 10..30 minutes, so eventually the program will recover.
Cheers,
Robert.
________________________________________
From: catg...@googlegroups.com [catg...@googlegroups.com] On Behalf Of DanITman [dtr...@gmail.com]
Sent: Tuesday, June 22, 2010 14:33
Dan Troup
Robert Deliën
Dan Troup
Robert Deliën
I will build you a special version. But I will use the Cartridge LED to indicate the water sensor state, instead of the Error LED.
On a side note: You're not using warm/hot water in your box, are you?
Cheers!
________________________________________
From: catg...@googlegroups.com [catg...@googlegroups.com] On Behalf Of Dan Troup [dtr...@gmail.com]
Sent: Wednesday, June 23, 2010 3:31
To: catg...@googlegroups.com
Subject: Re: Beta two results
I think it might be helpful for me to see the red light come on when water is detected and go off when it's not detected during the normal cycle procedure. The same way you have it in Geniediag.
On Tue, Jun 22, 2010 at 5:46 PM, Robert Deliën <Rob...@delien.nl<mailto:Rob...@delien.nl>> wrote:
Hi Dan,
That is correct: Even though the cause of the error may have ceased to exist, the error state indication is not cleared as a result. The box will continue to blink until it is power cycled. For now this is purposely made, so I can easily see if any errors have occurred.
I'm still in doubt on how to distinguish active errors from recovered errors in the final version. Users may not be interested in errors that were solved automatically, but this might be important information, especially if it occurs with every washing cycle.
On Jun 22, 2010, at 16:12, Dan Troup wrote:
This seems to be what is happening, however, would the LED continue to blink? So it blinks (error 2) after the the first drainage and then continues through 2 more cycles all while continuing to blink.
On Tue, Jun 22, 2010 at 9:35 AM, Robert Deliën <Rob...@delien.nl<mailto:Rob...@delien.nl>> wrote:
Hi Dan,
I'm really wondering what's causing your water sensor problem. I still think something is wrong with your LED of CPU pin. Is anybody else in this group experiencing problems with the water sensor?
One of my original boxes has a water sensor problem too. I will try to diagnose what's causing the problem and figure out a way to verify if your problem have the same cause.
The behaviour you've observed isn't strange. The washing program has several stages in which it waits for the water sensor to detect water (or not). When the waiting starts, a timeout is started as well. When the desired state is achieved, the timeout is destroyed and the program continues. However, if the timeout expires before that time, an error is reported. This means a blinking error LED on the panel. Besides that, the running washing program should be aborted and the appropriate cleanup program should be started. But this latter part isn't implemented yet. So instead the program keeps on waiting for the desired state, way past the timeout, until the sensor finally reports the desired state. On a false positive, this can take from 10..30 minutes, so eventually the program will recover.
Cheers,
Robert.
________________________________________
From: catg...@googlegroups.com<mailto:catg...@googlegroups.com> [catg...@googlegroups.com<mailto:catg...@googlegroups.com>] On Behalf Of DanITman [dtr...@gmail.com<mailto:dtr...@gmail.com>]
Sent: Tuesday, June 22, 2010 14:33
To: CatGenius
Subject: Re: Beta two results
I watched my box run a full cylce yesterday. Genie Dome was off. I
got the error 2 light again. This time instead of the box being stuck
it just continued on like the error didn't exist. I'm not sure why
this is.
On Jun 13, 8:07 am, Robert Deliën <rob...@delien.nl<mailto:rob...@delien.nl>> wrote:
> Hi Dan,
>
> Welcome back!
>
> That's actaully not that far fetched. After all, the water sensor is an optical sensor. And unlike the cat sensor, it is not able to distinguish it's own light from foreign light. With the dome off, more environment light will be shed into the hopper, where the water sensor is, than with the dome off.
>
> My cats refuse to use the box with the dome on. But my garage, where the boxes are, is usually dark so it should be quite similar to having the dome on. I'm trying to recall the light conditions when I'm running a washing program, but I don't think I leave the light on for that.
>
> Perhaps your water sensor LED is close to the burned out problem and some environmental light will help it to work, but I think that's a bit far fetched.
>
> I will do a couple of more measurement on the water sensor, to figure out if CatGenius uses it the same as the stock firmware does, but after I have migrated my SBS2003 server to SBS2008 today ;-)
>
> Cheers,
>
> Robert.
>
> -----Original Message-----
> From: catg...@googlegroups.com<mailto:catg...@googlegroups.com> [mailto:catg...@googlegroups.com<mailto:catg...@googlegroups.com>] On Behalf Of DanITman
> Sent: Saturday, June 12, 2010 20:36
> To: CatGenius
> Subject: Re: Beta two results
>
> Sorry for the dealy on this. I have been traveling on business and have not had time to do any CG testing.
>
> I found something very interesting today with my box.
>
> I tend to get the error LED more often when I have Genie Dome on the box then I do with it off. When I have it off the cleaning cycle never fails when I put it on it almost always fails. I'm not sure why this would make a difference but it does.
>
> Is anyone else having trouble when you have the Genie Dome on?
>
> Thanks
>
> On Jun 7, 3:05 pm, DanITman <dtr...@gmail.com<mailto:dtr...@gmail.com>> wrote:
> > Ran 4 complete cycles without issue. Things seem to be working well.
>
> > I'll follow up with you on the Serial port stuff, ran out of time
> > yesterday.
>
Dan Troup
Thanks
Robert Deliën
@Jaime: Could you explain the valve repair procedure a little more detailed? Photos would help, an item on the Google Group page would be perfect.
It's relatively easy to tell if your box suffers from this problem, all you need is a stop watch (and the box). Start a full washing program and start the stop watch when the water starts running. Stop the when either the water stops or the Error LED starts blinking, whichever comes first.
If the Error LED starts blinking before the water is stopped (and hence the box is full), you can be pretty sure you have a water flow problem. This problem can be caused by a molten valve washer as Jaime described, but it could also be your warm water system that doesn't produce enough flow to fill up the box in time. In the latter case mixing in cold water could help.
If the water stops (and hence the Error LED won't blink at all), we need to know how long it took for the box to fill. If this time is close the the maximum time (2 minutes and 15 seconds, from the top of my head), it could still be a flow problem. Typical filling time is way below the maximum, around 60 seconds in my case.
I took the maximum filling time from the original box. We have fairly high water pressure and opening the valve for this time still didn't overflow the box, so it is a safe value even if the water sensor is producing false-negatives.
cheers!
________________________________________
From: catg...@googlegroups.com [catg...@googlegroups.com] On Behalf Of Dan Troup [dtr...@gmail.com]
Sent: Wednesday, June 23, 2010 13:41
To: catg...@googlegroups.com
Subject: Re: Beta two results
Yes, I am still using hot water to wash my box. After seeing other failures with hot water it seems like this might be the issue. The crazy part is that I haven't had these issues under the stock firmware.
Thanks
On Wed, Jun 23, 2010 at 6:30 AM, Robert Deliën <Rob...@delien.nl<mailto:Rob...@delien.nl>> wrote:
Hi Dan,
I will build you a special version. But I will use the Cartridge LED to indicate the water sensor state, instead of the Error LED.
On a side note: You're not using warm/hot water in your box, are you?
Cheers!
________________________________________
From: catg...@googlegroups.com<mailto:catg...@googlegroups.com> [catg...@googlegroups.com<mailto:catg...@googlegroups.com>] On Behalf Of Dan Troup [dtr...@gmail.com<mailto:dtr...@gmail.com>]
Sent: Wednesday, June 23, 2010 3:31
To: catg...@googlegroups.com<mailto:catg...@googlegroups.com>
Subject: Re: Beta two results
I think it might be helpful for me to see the red light come on when water is detected and go off when it's not detected during the normal cycle procedure. The same way you have it in Geniediag.
On Tue, Jun 22, 2010 at 5:46 PM, Robert Deliën <Rob...@delien.nl<mailto:Rob...@delien.nl><mailto:Rob...@delien.nl<mailto:Rob...@delien.nl>>> wrote:
Hi Dan,
That is correct: Even though the cause of the error may have ceased to exist, the error state indication is not cleared as a result. The box will continue to blink until it is power cycled. For now this is purposely made, so I can easily see if any errors have occurred.
I'm still in doubt on how to distinguish active errors from recovered errors in the final version. Users may not be interested in errors that were solved automatically, but this might be important information, especially if it occurs with every washing cycle.
On Jun 22, 2010, at 16:12, Dan Troup wrote:
This seems to be what is happening, however, would the LED continue to blink? So it blinks (error 2) after the the first drainage and then continues through 2 more cycles all while continuing to blink.
On Tue, Jun 22, 2010 at 9:35 AM, Robert Deliën <Rob...@delien.nl<mailto:Rob...@delien.nl><mailto:Rob...@delien.nl<mailto:Rob...@delien.nl>>> wrote:
Hi Dan,
I'm really wondering what's causing your water sensor problem. I still think something is wrong with your LED of CPU pin. Is anybody else in this group experiencing problems with the water sensor?
One of my original boxes has a water sensor problem too. I will try to diagnose what's causing the problem and figure out a way to verify if your problem have the same cause.
The behaviour you've observed isn't strange. The washing program has several stages in which it waits for the water sensor to detect water (or not). When the waiting starts, a timeout is started as well. When the desired state is achieved, the timeout is destroyed and the program continues. However, if the timeout expires before that time, an error is reported. This means a blinking error LED on the panel. Besides that, the running washing program should be aborted and the appropriate cleanup program should be started. But this latter part isn't implemented yet. So instead the program keeps on waiting for the desired state, way past the timeout, until the sensor finally reports the desired state. On a false positive, this can take from 10..30 minutes, so eventually the program will recover.
Cheers,
Robert.
________________________________________
From: catg...@googlegroups.com<mailto:catg...@googlegroups.com><mailto:catg...@googlegroups.com<mailto:catg...@googlegroups.com>> [catg...@googlegroups.com<mailto:catg...@googlegroups.com><mailto:catg...@googlegroups.com<mailto:catg...@googlegroups.com>>] On Behalf Of DanITman [dtr...@gmail.com<mailto:dtr...@gmail.com><mailto:dtr...@gmail.com<mailto:dtr...@gmail.com>>]
Sent: Tuesday, June 22, 2010 14:33
To: CatGenius
Subject: Re: Beta two results
I watched my box run a full cylce yesterday. Genie Dome was off. I
got the error 2 light again. This time instead of the box being stuck
it just continued on like the error didn't exist. I'm not sure why
this is.
On Jun 13, 8:07 am, Robert Deliën <rob...@delien.nl<mailto:rob...@delien.nl><mailto:rob...@delien.nl<mailto:rob...@delien.nl>>> wrote:
> Hi Dan,
>
> Welcome back!
>
> That's actaully not that far fetched. After all, the water sensor is an optical sensor. And unlike the cat sensor, it is not able to distinguish it's own light from foreign light. With the dome off, more environment light will be shed into the hopper, where the water sensor is, than with the dome off.
>
> My cats refuse to use the box with the dome on. But my garage, where the boxes are, is usually dark so it should be quite similar to having the dome on. I'm trying to recall the light conditions when I'm running a washing program, but I don't think I leave the light on for that.
>
> Perhaps your water sensor LED is close to the burned out problem and some environmental light will help it to work, but I think that's a bit far fetched.
>
> I will do a couple of more measurement on the water sensor, to figure out if CatGenius uses it the same as the stock firmware does, but after I have migrated my SBS2003 server to SBS2008 today ;-)
>
> Cheers,
>
> Robert.
>
> -----Original Message-----
> From: catg...@googlegroups.com<mailto:catg...@googlegroups.com><mailto:catg...@googlegroups.com<mailto:catg...@googlegroups.com>> [mailto:catg...@googlegroups.com<mailto:catg...@googlegroups.com><mailto:catg...@googlegroups.com<mailto:catg...@googlegroups.com>>] On Behalf Of DanITman
> Sent: Saturday, June 12, 2010 20:36
> To: CatGenius
> Subject: Re: Beta two results
>
> Sorry for the dealy on this. I have been traveling on business and have not had time to do any CG testing.
>
> I found something very interesting today with my box.
>
> I tend to get the error LED more often when I have Genie Dome on the box then I do with it off. When I have it off the cleaning cycle never fails when I put it on it almost always fails. I'm not sure why this would make a difference but it does.
>
> Is anyone else having trouble when you have the Genie Dome on?
>
> Thanks
>
> On Jun 7, 3:05 pm, DanITman <dtr...@gmail.com<mailto:dtr...@gmail.com><mailto:dtr...@gmail.com<mailto:dtr...@gmail.com>>> wrote:
> > Ran 4 complete cycles without issue. Things seem to be working well.
>
> > I'll follow up with you on the Serial port stuff, ran out of time
> > yesterday.
>