19 replies to this topic

[Pete Brown]

Any recommendations on a waterproof temperature sensor I can use with Netduino? I'm fine with writing the driver code assuming the protcol is reasonable. This is for an aquarium and will need to support being submerged 100% of the time. Pete

[Pete Brown]

I never really considered taking a regular temp sensor and putting it in silicone, but this seems somewhat promising

http://www.practical...perature-sensor

There's another one out there which showed putting one in a bit of copper pipe, but copper is a no-no in aquariums if you have any shrimp or snails.

Pete

[Nevyn]

I never really considered taking a regular temp sensor and putting it in silicone, but this seems somewhat promising

If you are willing to look at encasing the sensor then have you heard of Polymorph? Melts at 60 C but then I think your fish would have problems if you start getting near this temperature.

BTW - not tried this, just wondering if it would help.

Regards,
Mark

Edited by Nevyn, 08 February 2012 - 07:35 AM.

[Pete Brown]

@Nevyn Thanks, yes, I have. I haven't checked to see if it's safe for aquariums, but it's an idea. An unknown to me is thermal conductivity of these various options. The copper pipe, of course, is pretty good. These others seem like they'll add a fair bit of insulation and therefore delay. Hmm. Maybe I can use aluminum instead of copper. I'll have to see how reactive it is with this, and if I can get thinwall aluminum tubing. (I don't think aluminum foil will hold up) Pete

[Mario Vernari]

Here's my suggestion.
I don't know how much is your budget, but I'd suggest a PT-100 probe, stainless-steel shielded, along with a good analog adapter (ADC biult-in as plus). I guess that stainless-steel should haven't any problem with any kind of fish/plant. (Let me ask to my wife...)

We're using two kinds of temperature probes for heavy dusty, and hard climate environment systems: KTY81 and PT-100. The first is much cheaper, less precise (-/+ 1'C), and not so reliable. Many customers switched briefly from KTY81 to PT-100: pricey, but you forget them.

We have some solution, but it's not tailored for home/hobbistic purpose, thus maybe will be not so cheap. Anyway, just PM me in case.

The tip about PT-100 is also about the precision: it's 0.1 'C, which is pretty important for some kind of fish.

Reptiles: I know that the turtle eggs kept at 26-27 'C yield mostly males, instead 28-29 'C yield mostly females.

Also, when using a so precise probe, the Netduino ADC is not recommended. Instead, should be used a calibrated analog converter, tailored for PT-100, and maybe exchanging data via SPI/I2C/UART.

Hope it helps.
Cheers

EDIT: Just asked my wife. Stainless-steel is surely better other than other common metals, especially when the water is salted. However, no material is "perfect": even plastic is toxic!

Edited by Mario Vernari, 08 February 2012 - 08:57 AM.

[CW2]

Any recommendations on a waterproof temperature sensor I can use with Netduino?

How about waterproof DS18B20 (Adafruit, Sparkfun)? I guess TO-92 package is not that hard to seal, perhaps it would be enough to wrap the leads with heat shrinks or cover with silicone...

[MDS]

You might try using type K thermicouple. Both chromel and alumel should be water resistant. You will need to put an opamp in the circuit but you should have no trouble getting good values to read with the adc and a table look up for the temp.

[Mario Vernari]

Both chromel and alumel should be water resistant.

The main problem would not be the life time of the probe, but the fishes!
Chrome and Nickel are highly polluting, Aluminum less, but depends of the acidity of the water (e.g. salted or not)
Cheers

[Paul Newton]

Hi Pete,

Many years ago at school I used an LM35 and potted it in two part adhesive to the end of a glass tube. The tube was a similar diameter to the LM35 package, so the wires slipped inside and were kept away from any liquid, but the plastic body of the probe was outside the glass. I don't now how inert the cured adhesive is - it might not be good for fish.

I should mention that I suggested using the LM35 to < cobolstinks > in this thread http://forums.netdui...rature-project/ but that he had to abandon the LM35 in favour of a one wire device. The issue was too much noise on the analogue signal probably due to long length cables back to the Netduino.

< Coding Smackdown > was also working on a temperature monitoring project, in this case for a home brewery http://forums.netdui...erature-logger. He must have sealed his sensors in some way.

Perhaps you could combine the fish tank and brewery - fishy beer!

Hope this helps.- Paul

[ErikN]

Is glass not the best option? I would think it would have low-to-no reaction to the water and - assuming the temperature probe were in direct contract with the glass - it should react fairly quickly to temperature change. Maybe you could just get a long test-tube, affix the sensor to the side near the bottom and submerge. Keep the open end above the water line and cork it. Isn't this how most aquarium heaters are built? For reaction time, I'd try to have the smallest amount of air volume possible - so maybe after affixing the sensor you could use a silicon seal just above the sensor to insulate it from the remaining air column in the tube/vial? This approach should keep everything except the glass completely away from the water for both electronic and fishy health. -Erik

[Paul Newton]

Simple ideas are always the best Erik! I'll bet you could use some heat sink compound at the bottom of the test tube to thermally bond the probe to the glass before you backfill with silicon. Paul

[ErikN]

Maybe it'd be a good idea to heat the tube to expel excess moisture in the air and possibly draw a bit of a vacuum when you seal it. That will further reduce the air volume and help eliminate condensation from affecting your board and causing corrosion. Alternatively maybe you could drop in a desiccant packet from a pill bottle if you don't want to heat blast the tube. Either way, I think it's a good idea to remove as much moisture from the sensor area before you seal it.

[H07R0D]

Maybe it'd be a good idea to heat the tube to expel excess moisture in the air and possibly draw a bit of a vacuum when you seal it. That will further reduce the air volume and help eliminate condensation from affecting your board and causing corrosion.

Alternatively maybe you could drop in a desiccant packet from a pill bottle if you don't want to heat blast the tube. Either way, I think it's a good idea to remove as much moisture from the sensor area before you seal it.

I ended up making one using a LM35DZ and a papermate flexgrip pen. I connected everything up, slid the sensor into the pen and filled the whole pen up with silicone. I made sure the sensor was exposed to air, but nothing else, and it works like a charm. I'm using it to monitor my aquarium temperature.

[Pete Brown]

Tons of great ideas. Thanks all. I like the idea of a stainless probe in the tank, but that method also sounds expensive. Using a glass tube is probably the best low-cost idea as mentioned by several folks here. Aquarium heaters are 100% submerged, but yes, they are generally glass for the heating element. The temperature sensing part is often part of the plastic head as opposed to the glass body. In general, their ability to hold a stable temperature increases with the price paid. I've blown through some mid-range ones in the past (luckily they just died without boiling the tank). The best one I had was a titanium one; no glass, and great heat transfer. The tank is fresh water, but kept somewhat acidic (how rams like it). I want to monitor and plot first, with the possibility of using the Netduino to control a number of things in the tank later. Just because I can We have so many sensors at our disposal, but the ones that work in aquaria are often way too expensive, having been designed for industrial chemical monitoring. pH, hardness and others are examples of that. I'm still learning what's out there; the temp one is likely the easiest. Pete

[Magpie]

I made a few submersible temperature sensors using a LM135, heatshrink and silicon (neutral cure). One of them was in my solar hot water cylinder for a few months and it was ok. I took it out because the tanks seal where the sensor cable entered, had busted, but the sensor itself was fine. The sensor was ok at boiling point. I dont know how they would effect the fish but they are definitely cheap and easy to make. Using the analogue inputs with this type of sensor you need a really good analogue power supply into Aref and in to the top of the LM135. So I went through a linear regulator, to chop out the noise. eg.7805 and a resistor divider. to bring it down to 3.3. Which improved things greatly.

[Pete Brown]

Stumbled across these. I had never considered looking for brewer info. These are 1-wire Dallas-sensor based temperatore sensors

http://www.brewersha...rature-Sensors/

Edit: They're stainless steel. They even offer the probe ends separately. They're made for the Dallas ds18b20.

Pete

[Mario Vernari]

Stumbled across these. I had never considered looking for brewer info. These are 1-wire Dallas-sensor based temperatore sensors

http://www.brewersha...rature-Sensors/

Edit: They're stainless steel. They even offer the probe ends separately. They're made for the Dallas ds18b20.

Pete

I'm wondering whether my infrared library could solve even the one-wire protocol interfacing. Unfortunately I don't have any 1-wire chip now.
Cheers

[H07R0D]

Stumbled across these. I had never considered looking for brewer info. These are 1-wire Dallas-sensor based temperatore sensors

http://www.brewersha...rature-Sensors/

Edit: They're stainless steel. They even offer the probe ends separately. They're made for the Dallas ds18b20.

Pete

Is Stainless ok in Fresh Water? I can't do it in my reef tank, which is why the Silicone DIY worked so well. That and I'd rather not spend an arm and a leg for a temp probe (that's for the pH probes.....)

[Pete Brown]

Stainless is fine in fresh water. Also, that same site had a comment where they'll use a different SS alloy if you want the probe for saltwater tanks. Glass is generally the most non-reactive, but it's also very inefficient for temperature transfer. Pete

[mcinnes01]

I tested out the stainless steel ones in an empty tank for a few weeks, you need to be sure about the quality of the probe (mine was quite cheap probably like the adafruit ones). The probe I have lost its stainless steel coating revealing the copper underneath it, obviously not good putting copper in a fish tank. The first link for pratical maker that was posted seems a good route I too looked at that, although you could get a glass test tube and seal a probe in with aquarium safe silicone. As suggested further up glass seems a sensible option, and as it will be fully submerged and temperature of a large volume of water is not likely to change very quickly, the glass should be a good enough conductor. Also if you think most thermometers are glass. The ds18b20's seem very accurate in my testing and I believe there are some other versions that allow you to assign an address, in case you have multiple probes e.g. sump, filler tank, main tank etc this would make it easy to identify the probes.