An interesting failure mode ...

[JoyfulNoise]

I had to replace my water temperature sensor (Pentair 10k thermistor) as the wire got chewed through by a rat (I also reorganized a lot of the control wiring around my pad). But when I went to pull the thermistor from the discharge pipe, I noticed it was very hard to take out. At first I thought maybe the PB had drilled the hole a little too small but when I finally finagled the thermistor out, this is what I found -

Pentair 10K Thermistor Failed View 2

• JoyfulNoise
• Mar 8, 2021
• pentair thermistor water temp sensor

Failed 10k water temperature sensor

Pentair 10K Thermistor Failed View 1

• JoyfulNoise
• Mar 8, 2021
• pentair thermistor water temp sensor

Failed 10k water temperature sensor

This sensor is over 8 years old but I thought that it was interesting how the body is swollen & cracked. It almost looks like freeze expansion damage but we have never really had much of a hard freeze event here and anytime temps have gotten low, the freeze protect kicks in and circulates pool water. I think the coldest my pool water has ever gotten is 38F. Seems a bit odd to me ... Any who, the probe has been replaced and the wiring is bundled up now all in flexible wiring wrap so as to deter critters, or at least make their target harder to get at. Looks like I'll be placing some snap traps around the pad to take out the critters .... chunky peanut butter in a snap trap, gets them EVERY TIME.

 

[JamesW]

Was it working up to the point when the wire was cut?

Can you cut it open to show the inside?

Looks like a lot of iron in the water from the orange color.

 

[JamesW]

Even if the temperature was below freezing, there should not be any water inside the probe that could freeze.

 

[CrystalRiver]

It looks to me like the core of the probe is iron. The covering got a small crack or pinhole, or perhaps the moisture entered from the other end. Either way, it rusted, and rust expands, cracking the casing and causing the swelling. The rust also discolored the outside of the casing.

 

[JamesW]

See if a magnet will stick to it to see if there is iron inside.

 

[JoyfulNoise]

Was it working up to the point when the wire was cut?

Can you cut it open to show the inside?

Looks like a lot of iron in the water from the orange color.

The probe was working even when I noticed the rat chewed wire. But, it seemed to be registering a higher temperature than I thought when I would put my hand in the water. Sure enough, the new sensor is telling me my water temp is in the low 60's (and I confirmed it with an external temp probe) whereas before I replaced it the probe was saying the water was in the low 70's.

The wire eventually failed (when I picked it up) and the easyTouch panel showed "ERR" on the water temp display

The expanding rust theory does seem plausible given that my water has no iron in it....

 

[Flying Tivo]

I have another theory. If you look closely at the base of the sensor it has conformed to the roundness of the PVC pice. I suspect the sensor was tighten way too much and deformed the plastic and made it crack. When putting in the new one, only tighten enough so the O-ring seals and does not leak. Do it with the pump running at usual working PSI. 8 years is not bad at all.

 

[Jimrahbe]

Matt,

I have seen exactly the same thing here...

Thanks,

Jim R.

 

[cowboycasey]

This is so funny... I just pulled mine out Saturday and it looks exactly the same after 7 years.... It was still registering temp, i will have to look when I get home as it may still be sending temp... I thought I drilled the hole too small as I had to pry it out... it never did leak 1 bit, I wonder because of the swelling..

 

[JoyfulNoise]

I have another theory. If you look closely at the base of the sensor it has conformed to the roundness of the PVC pice. I suspect the sensor was tighten way too much and deformed the plastic and made it crack. When putting in the new one, only tighten enough so the O-ring seals and does not leak. Do it with the pump running at usual working PSI. 8 years is not bad at all.

Yeah, the new one is in place and lightly torqued down. If I remember correctly, the base of the probes are intentionally molded that way with a curve mating surface to better match the PVC pipe surface.

And yes, 8 years in direct, blazing AZ sun. The wire itself was very brittle from exposure. I hope to get a good long time out of the replacement.

 

[JoyfulNoise]

Seems like we have a common failure mode here. Some kind of crack, water penetration into the body and then swelling failure.

 

[Killer95Stang]

I had a similar finding recently with my system. I initially thought I had a small leak around my sensor, but realized that the drip was coming from the a cracked coupling in the SWG right above it. When I took the sensor out of the pipe, mine was cracked and swollen exactly like yours. Since it was still working and i didn't want to tackle that project at the time, I cleaned it up, installed a new oring and finished it up with a brand new stainless hose clamp. My sensor was about 6 years old and was the second sensor after we had the first one fail under Pentair warranty.

 

[JoyfulNoise]

The air temp sensor is the same exact 10k thermistor just hanging outside the easytouch panel. That sensor looks totally normal, like new. No swelling.

I wonder if these sensor casing material just isn’t design to be submerged all the time or perhaps the water pressure at the discharge pipe is enough to “crush” the sensor body a little bit abs induce cracking. Hard to believe that but perhaps the high flow rate of water is enough to do damage. These probes sit right in the water stream.

 

[JoyfulNoise]

I wonder if installing the probe using a PVC tee and backing the sensor out of the water stream would be better . 2" pipe at 10psi to 15psi (roughly) with 40-60 GPM of water flowing around it might be enough stress to induce cracking.

 

[JamesW]

For NTC's cobalt, nickel, iron, copper or manganese are common oxides.

The inside probably does contain iron.

The casing probably failed and the iron rusted.

The pressure isn't high enough to damage the sensor.

The sensor is made for the purpose, so the immersion and pressure should not cause failure.

Overall, the longevity is probably acceptable, but it could probably be made to last longer.

 

[JamesW]

Due to the swelling, I would suspect that the pressure came from the inside.

Maybe due to a reaction in the sensor for some reason.

There's a continuous current running through the sensor, so maybe the metal oxides just eventually fail due to galvanic type oxidation?

 

[JoyfulNoise]

I'll try to split it open later and see what's inside.

The currents and voltages should be really small but maybe there's enough to cause some issues. We do know that the IC thermistors on the flow switch fail regularly and they are not nearly as robust as the pipe mounted ones

 

[JoyfulNoise]

Some plastics do swell over time from constant water immersion. So it could be that the combination of water absorptions, hydrolysis causing the plastic to weaken and the forces involved are enough to cause stress fractures that eventually open up with time. Once water seeps into the thermistor body, whatever is in there will corrode and fail.

 

[JamesW]

It's designed to be used for the purpose.

So, it would seem to be a bad choice if the plastic was susceptible to damage from water.

In any case, it seems that they could probably increase the quality of the sensors for better longevity.

Certainly, the sensors in the flow switches are below acceptable quality based on the high failure rate.

The Intellichlor temperature reading needs to be displayed on the automation, which should be easy to do with a simple software change.

 

[JoyfulNoise]

Potted in black wax -



Reading 10.60k in the garage which is about 72F right now -


I think the resistance is a bit off but not badly so.