A fascinating fault with my automatic soap dispenser

[Andrew Lorimer]

I have a Dettol automatic soap dispenser (see fig. 1) which uses a little propellor-type pump to let out a small amount of soap when you put your hand in front of the passive IR sensor.

It's been working fine until just before, when I turned the lights off in the kitchen (night time, so moderate brightness became pitch black). And out of the blue, the bloody thing wakes up and ejaculates soap all over my kitchen sink. I wasn't even remotely close to the device.

I considered reflections off the stainless steel sink, but it's never done this before, and sure enough when I switch the light from elsewhere, the dispenser still triggers. So this rules out any IR interference getting to the device.

Turns out the device is also triggered by by phone torch, and the nearby dining room lights. But it only triggers when any of these lights are turned off, not on - in other words, it is triggered by drop in brightness, not increase.

My second thought was that there might be an IR-pass lens just before the sensor, which might be degrading, causing changes in visible light to be picked up by the sensor. But this seems like a bit of a stretch.

Another idea is that there is some sensitivity-control logic which automatically adjusts tendency to release soap. Perhaps the microcontroller hit a bug. I'm yet to try resetting the device (battery powered), but I'm not sure how this would be explained since it's at least a year since the batteries were changed, and I've never experienced this before.

Anyway, ideas would be appreciated so that I don't have nightmares about my house flooded with soap.

Fig. 1: Beware the innocent-looking soap dispenser

 

[anorlunda]

That sounds like a fun problem. Others here at PF are much better than me at troubleshooting, but I found the following.

They suggest mounting the PIRs in such a way that the PIR cannot "see" out of a window.

It is also recommended that the PIR not be placed in such a position that an HVAC vent would blow hot or cold air onto the surface of the plastic which covers the housing's window.

Sensors are also often designed to "ignore" domestic pets, such as dogs or cats, by setting a higher sensitivity threshold, or by ensuring that the floor of the room remains out of focus.

The also use lenses and mirrors. A mirror out of position might "aim" it in an abnormal direction.

 

[berkeman]

LOL

My second thought was that there might be an IR-pass lens just before the sensor, which might be degrading, causing changes in visible light to be picked up by the sensor. But this seems like a bit of a stretch.

I'd suggest adding another IR lens in front of the existing opening. Just tape one on to see if it solves the problem. If it does, you can make it look nicer somehow. If it doesn't, it might be an EMI problem. Are these lights that you are switching on and off incandescent? LED? Compact fluorescent? Fluorescent?

 

[JBA]

I have a battery controlled atomic digital wall clock that goes berserk when the batteries get low. You might try new batteries on your unit to see if that helps the problem.

PS On a general note about battery replacements. On my field GPS I use rechargeable batteries and always test them before recharging; and, through this process I have found that consistently the battery with its + pole connected to the device in any series battery set will lose its charge while the remaining battery(s) will still be good. I have since found this to be true for both standard and rechargeable ++A and +++A batteries. In fact, switching the position of my two GPS batteries at the end of each period of usage extends the field life of the GPS between charges by balancing the amount of charge loss in both batteries. I have no idea why this is the case.

 

[berkeman]

PS On a general note about battery replacements. On my field GPS I use rechargeable batteries and always test them before recharging; and, through this process I have found that consistently the battery with its + pole connected to the device in any series battery set will lose its charge while the remaining battery(s) will still be good. I have since found this to be true for both standard and rechargeable ++A and +++A batteries. In fact, switching the position of my two GPS batteries at the end of each period of usage extends the field life of the GPS between charges by balancing the amount of charge loss in both batteries. I have no idea why this is the case.

Very interesting! Can you start a new thread in the EE forum about this and post a link to it here? Offhand I have no idea why that would be true, but it does sound like you've done careful measurements. That will be an interesting thread! Thanks.

 

[Andrew Lorimer]

Update the following day...

Interestingly, the soap dispenser was triggered this morning when my mum opened the blind just above the dispenser, which was closed last night when the problem started. Opening the blind let in a lot more natural light. So it can be triggered by increase in light, not just decrease as I thought before. Unless it was just the decrease in light from my mum's shadow or something.

My mum turned the dispenser off to put a stop to its nonsense, and I turned it back on soon after to see if the problem was fixed. It's hard to tell because it's now daytime so there is much more ambient light, but as far as I can tell the problem is now fixed. (I shielded the sensor with some paper and used a bright torch to test it). This suggests a microcontroller bug, probably a stack overflow or flipped bit due to EMI or something.

I'll update if the problem persists. Thanks internet!

 

[Keith_McClary]

the battery with its + pole connected to the device in any series battery set will lose its charge while the remaining battery(s) will still be good

I have seen a device (can't remember what it was) that taps between series connected batteries.

 

[JBA]

As suggested, I have started a thread under the title "Observed battery set discharge unbalance" on the Electrical Engineering forum for discussion about the battery discharge issue. Please post any further feedback on this subject on that thread.

https://www.physicsforums.com/threads/observed-battery-set-discharge-unbalance.948903/

 

[DaveC426913]

Huh. I always assumed these sensors were triggered by IR ranging i.e. detecting the distance of something physically within range.

 

[berkeman]

Huh. I always assumed these sensors were triggered by IR ranging i.e. detecting the distance of something physically within range.

The Passive Infrared (PIR) sensor is very inexpensive, so it's used for most of these IR motion sensor applications.