Design of a high temperature (50-100C) wireless thermometer

[mux]

I've got an interesting design question. How would one go about designing a long-lasting, high-temperature capable wireless thermometer?

I've got quite some years of experience designing all kinds of stuff, from embedded to my main focus power electronics, but at first glance this doesn't seem to have a clear-cut answer. Anyway, the use case is:

I will be installing solar panels on my roof in about a month. These are mounted on Consoles, which are big, black 'buckets' with the panels mounted on the slanted 'opening'. The temperature inside this bucket can go up pretty significantly, and I wish to know how much exactly. Not just once; I want to track it through the years. This thing has to be IP67 (submergeable and still functional) and it should last for years, not weeks. Typical temperatures in these buckets are sub-50C 99% of the time. Only in the perfect storm of no wind and maximum insolation the temperature may creep up to near the boiling point.

The main issue I'm worried about is the power supply. Obviously it's not acceptable to have a battery, at least I don't know of any easily available battery that will survive these temperatures for so long. Ultracapacitors: same deal. None of them are specced above 55C, and even then they only last a thousand hours or so. Running a wire to a battery 'outside' the hostile environment: not much better, because that's a black bituminous roof with the occasional bird that thinks batteries are edible.

so... what's reasonably left? Power it directly from the solar cell with a buck converter? LDO even? It's microamps. What would you do?

As for the rest of the circuit: I'm probably just not going to put any effort into that and use my trusty nRF24L01's with a random Atmel microcontroller on a massive sleep timer. Maybe something like once every 15 minutes or once every hour it wakes up and reports temperature, then goes to sleep. I don't see much reason to put significant engineering into that.

[LaserSteve]

Any STRONG reason the whole package needs to be at 100'C?

  There are some techniques using RF for high temperature electronics with a passive re-radiator  at twice the excitation frequency.  The receiving coil pumps a harmonic multipler which is modulated by the temperature data.

Problem is the diode leakage in the multiplier goes way up at 100'C.  Usually these mechanisms are based on a ceramic module, not a PCB. 

Look at Down well logging electronics for dealing with high temperatures.

My advice is you have a fixed installation. Forget the floating installation and just use a thermocouple. Just run it to cooler electronics elsewhere.  Ceramic wool or fiberglass wire for high temperature sensing is not expensive. In fact, 100'C is cool by my standards. Type J or type K thermocouples are cheap.

Steve

[mux]

Well, if I could just run a thermocouple to my roof, I can hook up the logging multimeter and do zero engineering. Possible sure, but fun? No. I want fun

It's impractical to run wires all the way around my roof and into the house. The whole idea behind the logging being wireless is convenience, nothing more, nothing less. I'd like to at least *try* to come up with a way to make this work. Also, there is a really big problem with birds and branches attacking wires. I'm not spending $40 on a 10 meter silicone sheathed wire (per thermometer!) just to fix that problem.

Anyone else willing to give this a bash?

[dfmischler]

Why wireless?  It is attached to a building and it won't move.  Put the cable in conduit if you have to.

100 degrees C is not really that hard.  A DS18B20 would survive your environment (rated for -55C to 125C).  You can find them pre-potted and cabled for cheap, or do your own packaging.

[Kalvin]

There are some cheap / quite inexpensive wireless WiFi modules which has been used for similar applications: ESP8266 and WR703n. Google will show some completed projects, so you may take a look whether these modules are suitable for your needs.

[tonyarkles]

As far as power goes... why not solar?

For the temperature measurement, you've mentioned these consoles/buckets. Would it be possible to run a sensor into the bucket without having the whole circuit submerged? I assume these are expensive, but a low-cost version would probably be makeable: http://www.ronan.com/products/rtd-probes-and-thermocouples. Alternatively, using something mechanically similar to that but using a DS18B20 (as mentioned above) inside it.

Once you've got those figured out, strapping a little wireless transmitter onto the side of the bucket would be pretty easy.

Another reason to not submerge the whole thing is that you're probably going to have a hard time transmitting from inside a bucket of water. I haven't done any math on it, but my gut feeling is that you're going to have a PILE of attenuation trying to do that. There's a reason underwater communication is really high powered and low frequency.

[mux]

Why wireless?  It is attached to a building and it won't move.  Put the cable in conduit if you have to.

100 degrees C is not really that hard.  A DS18B20 would survive your environment (rated for -55C to 125C).  You can find them pre-potted and cabled for cheap, or do your own packaging.

Well, again, it's a convenience thing. I don't *want* wires. Or at least not to the inside.

The temperature is obviously no problem for any solid-state electronics; I've designed stuff for 125C continuous with zero issues. The problem is that the power source should be able to survive this.

As far as power goes... why not solar?

Stupid as this sounds...  hadn't thought of giving it its own tiny solar panel. That is probably my best bet. I'm going to do some calculations on that.

There's a reason underwater communication is really high powered and low frequency.

It's not typically submerged, but the assembly should be IP67. The 'buckets' have drain holes, but still with the right wind a lot of water can come in and ruin exposed electronics.

So no, no attenuation issues in particular.