DS18B20 Salt Proof

[mginster]

So I mistakenly thought the basic DS18B20s were waterproof. Both of mine corroded and have been leeching for probably a few weeks. No adverse affects and running carbon now…but what does everyone do to actually make these useful?

Just saw people recommend the real ones from spark fun. $10 is fine but what the heck $12 shipping for probably 1 ounce…

 

[Jason_MrFrags]

Probably waterproof, but not ment for saltwater. You can seal it in heat shrink and silicon.

 

[Sral]

common problem indeed, some people have taken off everything and “repotted” the chip.

Not sure what that means exactly, but I think it’s similar to what @Jason_MrFrags said: apply your own silicone and make sure it’s very (!) well sealed. I think the biggest problem is always the gap between silicone and the cable, so the suggestion of using your own heatshrink there to reinforce is great !

Maybe apply a bit of silicone first and then let the heatshrink settle on top, so the silicone is well pressed and seals all the crevices left by the heatshrink ? Not sure though, never tried that personally.

 

[mginster]

common problem indeed, some people have taken off everything and “repotted” the chip.

Not sure what that means exactly, but I think it’s similar to what @Jason_MrFrags said: apply your own silicone and make sure it’s very (!) well sealed. I think the biggest problem is always the gap between silicone and the cable, so the suggestion of using your own heatshrink there to reinforce is great !

Maybe apply a bit of silicone first and then let the heatshrink settle on top, so the silicone is well pressed and seals all the crevices left by the heatshrink ? Not sure though, never tried that personally.

Yeah I may work on an idea to do a first pass with silicon to seal the end and connection to cable. Then do a second layer combined with heatshrink immediately to compress the silicon to the first layer.

 

[Wolfw28]

So I mistakenly thought the basic DS18B20s were waterproof. Both of mine corroded and have been leeching for probably a few weeks. No adverse affects and running carbon now…but what does everyone do to actually make these useful?

Just saw people recommend the real ones from spark fun. $10 is fine but what the heck $12 shipping for probably 1 ounce…

Don’t buy the cheap ones from Amazon or AliExpress their garbage. Every one of them I have has corroded the pro itself nothing to do with leaking into the sensor. The material they use on the probe is junk was actually going to make a post on this recently, but you beat me to it.

 

[Wolfw28]

The other thing is with Reef pie, which is my biggest argument is one of the sensor fails it goes to 0°, which then will turn your heater on and boil your tank

 

[gbroadbridge]

So I mistakenly thought the basic DS18B20s were waterproof. Both of mine corroded and have been leeching for probably a few weeks. No adverse affects and running carbon now…but what does everyone do to actually make these useful?

Just saw people recommend the real ones from spark fun. $10 is fine but what the heck $12 shipping for probably 1 ounce…

I just put them into food safe plastic tubing.

 

[DanTheReefer]

+1 to heat shrink sleeving and some sort of adhesive / epoxy to seal the end, mine have worked for years

 

[Wolfw28]

I just put them into food safe plastic tubing.

 

[Wolfw28]

I just put them into food safe plastic tubing.

Insulating them reduces the the accuracy and sensitivity of the sensor.

 

[Wolfw28]

+1 to heat shrink sleeving and some sort of adhesive / epoxy to seal the end, mine have worked for years

Insulating them reduces the sensitivity and the accuracy of a sensor

 

[gbroadbridge]

Insulating them reduces the the accuracy and sensitivity of the sensor.

No it doesn't.

It reacts and reads exactly the same as an encapsulated marine safe device which I ran in parallel as a test.

A case of testing proving supposition false.

 

[Sral]

No it doesn't.

It reacts and reads exactly the same as an encapsulated marine safe device which I ran in parallel as a test.

A case of testing proving supposition false.

I agree, as long as the cable doesn’t run through massively different temperatures, insulating the tip this way will not change accuracy.

It might reduce sensitivity to fast temperature changes, although the water reservoir is probably so big that it can’t really change that quickly anyway.

 

[Wolfw28]

That all depends on the size of the tank and sump. That be said this is the temperature reading from my 20 gallon all in one Red Sea nano. You can see how quickly the temperature changes from the time my chiller is turned off to how quickly the water heats up again to turn the chiller back on. My tank only fluctuates 1°F. With increased sensitivity of any sensor in crease accuracy. That’s why you have cheap measuring devices and lab grade measuring devices. @gbroadbridge sensor in the plastic tube I would have to see to believe not only does the plastic insulate the probe the air in the tube is an insulator. I completely did away with the dallas test sensor and replaced it with a marine grade temperature that a thermistor. It a simple 2 wire sensor.

 

[Sral]

That all depends on the size of the tank and sump. That be said this is the temperature reading from my 20 gallon all in one Red Sea nano. You can see how quickly the temperature changes from the time my chiller is turned off to how quickly the water heats up again to turn the chiller back on. My tank only fluctuates 1°F. With increased sensitivity of any sensor in crease accuracy. That’s why you have cheap measuring devices and lab grade measuring devices. @gbroadbridge sensor in the plastic tube I would have to see to believe not only does the plastic insulate the probe the air in the tube is an insulator. I completely did away with the dallas test sensor and replaced it with a marine grade temperature that a thermistor. It a simple 2 wire sensor.

That period looks like it's about 10-15 minutes. For 20 gallons at 1050 kg/m³ and roughly 4160 J/kg/K that equates to about 400-800 watt of heating/cooling power peak. If that comes from the evaporation of water, I would assume an ATO consumption of something like about 1/2-1 gallon per day ?

Sounds a bit high, but then again, I don't know anything realy. Might be that only parts of the tank, e.g. the top, heats up and cools down, that would decrease the calculated amount of cooling power und evaporated water drastically.

Judging from my test in air a while back, I would guesstimate the sluggishness of a DS18B20 sensor in air to only 1-2 minutes, meaning that you will read full changes about 2-4 minutes later. Coming back to the tube insulation from before:
that should not be much slower, I would guesstimate maybe 5-6 minutes or so. So there will be delay and therefore inaccuracy on changes faster than 6 minutes, but that is probably rarely the case.

 

[Wolfw28]

That period looks like it's about 10-15 minutes. For 20 gallons at 1050 kg/m³ and roughly 4160 J/kg/K that equates to about 400-800 watt of heating/cooling power peak. If that comes from the evaporation of water, I would assume an ATO consumption of something like about 1/2-1 gallon per day ?

Sounds a bit high, but then again, I don't know anything realy. Might be that only parts of the tank, e.g. the top, heats up and cools down, that would decrease the calculated amount of cooling power und evaporated water drastically.

Judging from my test in air a while back, I would guesstimate the sluggishness of a DS18B20 sensor in air to only 1-2 minutes, meaning that you will read full changes about 2-4 minutes later. Coming back to the tube insulation from before:
that should not be much slower, I would guesstimate maybe 5-6 minutes or so. So there will be delay and therefore inaccuracy on changes faster than 6 minutes, but that is probably rarely the case.

So calculate that delay what would be the actual temperature

 

[DanTheReefer]

Insulating them reduces the sensitivity and the accuracy of a sensor

It’s negligible for our application, water temp doesn’t change fast enough for us to notice a difference. In 3 years I’ve had two temp related warnings that saved the tank (heater issue and clogged overflow - saved the house too-)

 

[Sral]

So calculate that delay what would be the actual temperature

Well, hard-ish to say. Let me think about it.

If you have a fast-ish ramp that's roughly like your case: a rise of 1°F in 6 minutes and your sensor has a very slow rise time of 6 minutes (that's a bit on the unrealistically slow side) your measurement will be the following:

So in the extreme case when the temperature ramp stops at 1°F (case 2) the maximum error peaks at about 0.6°F after 6 minutes.

Even if the ramp never stops (case 1) the maximum error stays at about 1.1°F.

Both of those cases are for a comparatively fast ramp for a very slow sensor. In all likelyhood the error will stay lower because:
- the water temperature might not increase that fast throughout your tank
(that's equivalent of 500W of heating/cooling) on your 20gallon tank)
- the sensor is probably not that slow

So I'd guesstimate the typical error to be below 0.5°F

 

[Wolfw28]

It’s negligible for our application, water temp doesn’t change fast enough for us to notice a difference. In 3 years I’ve had two temp related warnings that saved the tank (heater issue and clogged overflow - saved the house too-)

In my case Dallas temperature sensor failed. Lost communication between sensor and raspberry pi. Reef pi defaulted to 0°F which then turn on a 1000 water heater and never turned it off. Which then boiled my farm tank. Water temperature 101°F bleached over $1000 worth of sps. One heck of a price to pay for three dollars sensor.

 

[gbroadbridge]

Redundancy rules, but I'm sure you know that now.