DS18B20 Failure

rq3

Senior Member
Basically just a heads up. Appliances always seem to fail when one is on vacation. Sure enough, on my return from the Caribbean yesterday, the fridge was a bit warmer than usual. One of the three picaxe controlled DS18B20 temperature sensors on my controller board had failed! No apparent reason. A new one fixed it right away, and the board has otherwise been performing flawlessly for many months. The failed unit displayed a constant zero degrees celsius.

I suspect moisture induced failure of the sensor, and have potted the new sensor in a small brass tube with technical urethane. The sensor gets thermally cycled from -17C to +7C, and sees quite a bit of condensation. Epoxy is too rigid, and may stress the leads, while the urethane is designed for potting thermally cycled devices from -72C to +120C, and has very low moisture transmission. Hope this helps someone!
 
Last edited:

PhilHornby

Senior Member
DS18B20's in the fridge

My fridge has been controlled by a pair of DS18B20's, since June 2014. I used the 'waterproof' ones, like this: https://www.adafruit.com/product/381 (though I paid a lot less).

I only used these because I couldn't think how to mount a bare DS18B20 ... but it seems I accidentally made the right choice :)
 
Last edited:

inglewoodpete

Senior Member
I wonder if there is a suitable flexible compound that could be used (like the so-called 'silicone' sealants). I imagine they could make the DS18B20 respond quite slowly but this would not normally matter for refrigeration.
 

Jeremy Harris

Senior Member
I've a loads of these running for years, potted in short lengths of B&Q 6mm OD tube dotted around outside, underground, in water tanks etc, and never (yet) had a failure. I've just put very thin heat shrink around the soldered joints, soldered a very short (~3mm) length of 5mm diameter brass bar in the end of the tube as an end cap, then filled the tube part way full of thin epoxy (I've used West epoxy, but only because I have it around for boat building). Warming the tube slightly makes it easier to pour the epoxy in, as it gets very runny when warm (but does go off a lot faster, so you need to work quickly). A DS18B20 is a nice tight fit inside a bit this brass tube, and when pushed down to the bottom it displaces some of the epoxy and also drives out any air bubbles. The waste epoxy is wiped off from around the top and it's left standing vertically for the epoxy to cure. Once the epoxy has cured, I finish the job off by fitting a short length of adhesive heat shrink around the top, just as strain relief around the cable and as an added seal. The cable I use is the 3 core screened stuff from Rapid Electronics, the Def Stan stuff here: https://www.rapidonline.com/unistrand-defence-standard-3-core-screened-signal-cable-7-2-3c-100m-02-0272 It's a good fit inside the top of the bit of brass tube, and leaves just a big enough gap to let the bubbles and excess resin out.

I've made some very long sensors like this for water tanks (around 400mm long) as well as lots of shorter ones (around 30mm long). The hardest part I find is fitting the very thin heat shrink around he wire to DS18B20 soldered joints and dressing the joints so they slide into the tubes easily.

The response time doesn't seem to be affected much, if at all, by adding these brass sleeves, probably because the thin (0.5mm) wall thickness of the brass doesn't have a lot of thermal resistance and the epoxy provides a reasonably good heat transfer path. That particular screened cable is also very tough and mechanically up to being pulled through ducts or buried in the ground.

I've never had a problem with the epoxy seal failing, but that may well be because there's no chance of movement with the whole lot inside the brass tube.
 

PhilHornby

Senior Member
Funny smells

Cheap 'Waterproof' ones off eBay ?
Where else? :)


If making your own, there is something to note - especially in the O.P.'s case, of use in a fridge...

...Years ago, I fixed a leaking car window seal, using bathroom sealant. The guide I was following, strongly advised against this - pointing out that many sealants give off a pungent, vinegary smell, while curing. Oh boy, did I live to regret my choice! It's probably well worth reading the sealant/epoxy instructions, before mixing it with your food supply!

(I just found the car article (from 2004) - it recommended this: http://www.everbuild.co.uk/450-builders - for that particular job)
 
Last edited:

Buzby

Senior Member
... many sealants give off a pungent, vinegary smell, while curing.
That smell is acetic acid. Acetic acid and copper react, in the right circumstances, to form copper/cupric acetate.

You really don't want any kind of chemically formed corrosion on your nice shiny copper PCB, especially if its not plated.

Steer clear of bathroom sealants for electronic work.

Cheers,

Buzby
 

Pongo

Senior Member
I've a loads of these running for years, potted in short lengths of B&Q 6mm OD tube dotted around outside, underground, in water tanks etc, and never (yet) had a failure. I've just put very thin heat shrink around the soldered joints, soldered a very short (~3mm) length of 5mm diameter brass bar in the end of the tube as an end cap, then filled the tube part way full of thin epoxy (I've used West epoxy, but only because I have it around for boat building). Warming the tube slightly makes it easier to pour the epoxy in, as it gets very runny when warm (but does go off a lot faster, so you need to work quickly). A DS18B20 is a nice tight fit inside a bit this brass tube, and when pushed down to the bottom it displaces some of the epoxy and also drives out any air bubbles. The waste epoxy is wiped off from around the top and it's left standing vertically for the epoxy to cure. Once the epoxy has cured, I finish the job off by fitting a short length of adhesive heat shrink around the top, just as strain relief around the cable and as an added seal. The cable I use is the 3 core screened stuff from Rapid Electronics, the Def Stan stuff here: https://www.rapidonline.com/unistrand-defence-standard-3-core-screened-signal-cable-7-2-3c-100m-02-0272 It's a good fit inside the top of the bit of brass tube, and leaves just a big enough gap to let the bubbles and excess resin out.

I've made some very long sensors like this for water tanks (around 400mm long) as well as lots of shorter ones (around 30mm long). The hardest part I find is fitting the very thin heat shrink around he wire to DS18B20 soldered joints and dressing the joints so they slide into the tubes easily.

The response time doesn't seem to be affected much, if at all, by adding these brass sleeves, probably because the thin (0.5mm) wall thickness of the brass doesn't have a lot of thermal resistance and the epoxy provides a reasonably good heat transfer path. That particular screened cable is also very tough and mechanically up to being pulled through ducts or buried in the ground.

I've never had a problem with the epoxy seal failing, but that may well be because there's no chance of movement with the whole lot inside the brass tube.
I've used a very similar construction for aquarium sensors except using thin wall epoxy/fiberglass tubing (sold for kite stays). The epoxy bonds very well to the tube and I've never done any other sealing at the sensor end. They have survived years in warm salt water with no issues.
 

cravenhaven

Senior Member
That smell is acetic acid. Acetic acid and copper react, in the right circumstances, to form copper/cupric acetate.

You really don't want any kind of chemically formed corrosion on your nice shiny copper PCB, especially if its not plated.

Steer clear of bathroom sealants for electronic work.

Cheers,

Buzby
In OZ you can get bathroom sealant that doesnt have that smell. Its generally referred to as 'Neutral Cure' silicon.
 

premelec

Senior Member
Once i was caulking some tiles and used two different RTV silicones - one put out acetic acid and the other ammonia - wonderful needle crystals appeared apparently of ammonium acetate.. they wiped off easily. There are many special silicones; I tried the usual acetic acid type on a mirror back and it ruined the mirror... glass shop people had the right kind... I even have some that touts 'water wash up' it's not as robust as other kinds but does cure ok. Recent acquisition and use of Rescue Silicone self sealing tape has shown this product can make a good water tight seal if your can stretch it enough while winding on... I'll try some on an 18B20 as it takes a lot of heat ok too.
 
Top