Thermal data system

For PC type software that runs under some PC opsys.

Thermal data system

Postby solar_dave » Tue Jan 06, 2015 7:19 am

Doug I want to do and inexpensive data acquisition for my solar hot water system, I know you did some alarm stuff around your place. I think 4 channels of input would be plenty. I think I prefer WiFi for transmission medium.
Dave Shiels

My TED 5000 power monitoring
http://phx-solar.no-ip.info:8081/Footprints.html
solar_dave
 
Posts: 108
Joined: Tue Aug 21, 2012 12:33 pm
Location: Phoenix, AZ

Re: Thermal data system

Postby Doug Coulter » Wed Jan 07, 2015 12:32 pm

Well, I've had some practice at that for an earlier system I had. Hot water is harder, in particular if the sensors have any contact with the water/steam. For one thing, there are zero reliable solid state sensors for the temperatures you'll be seeing, they will all die quite young (sometimes the first day). You therefore must use things like thermocouples and interface to them in some way (tiny millivolt signals). That said, it's definitely do-able, something off the shelf like an Arduino Uno can easily control some sensors and possibly some valves/motors etc with the right code. It will do a lousy job of logging (they have a SD shield and a kind-of file system but very limited) so in general you interface the thing with either a PC or something inexpensive like a raspberry pi for logging and any heavy lifting math-wise. The strategy here is to use the Arduino (or other small uP without a fancy opsys) to do anything time-critical, and buffer a little data for a larger machine that isn't real-time good (due to a pre-emptive multitasking opsys) but is good in other ways, while in our case...not using a lot of power. I'm testing a system right now that uses an uno and some solid state sensors for the other extreme - is my plumbing going to freeze...how's room temperature in a remote building on campus where the water system lives? I'll put up some plots as soon as I write the plotting software...

Right now, I'm using a raspberry pi with a wireless dongle to log data, and using adafruit's 10.1" LCD/LED display, and an arduino interfaced to that which actually collects the data, which is put in a file in a shared directory on the pi. The whole thing draws perhaps <10w, haven't measured it in detail yet, as I plan to run it off my 13.5v ups (separate from the house batteries) eventually, with tiny switching supplies which will cut the waste from the way I'm doing it right now - wall warts and bricks from the inverter power off the main system. I built a little UPS that keeps a deep cycle 12v battery up for just this kind of thing - it will run stuff like this for days even if the main house power goes out (it's a big trolling/deep-cycle battery with a temperature compensated charger that runs off a big 15v power supply used to run my water pump already in existence). Obviously, your details will vary.

Don't believe the high end of the temperature ratings on these solid state sensors, they are assuming a very short life at those numbers, and I've even had thermistors just plain fail. Depending on your system, your collectors might get VERY hot if there's no water circulating through them as well. If there's any steam blow-off anywhere it will eat wires, much less anything "delicate" so you even wind up encapsulating the thermocouples....Thank heavens I learned all that back in the day, nowadays it'd be a lot easier going in - uP's are cheap, simple, sensors are better, and so on - and I know what to not use, most important of all.

In your case you might not need/want super high time resolution and get by with just a raspi and some interfaces and of course, the usual UI stuff (keyboard, display etc) even if those in parens are only used initially to get it working. Right now, mine isn't fancy enough to do things like roll over logs, I'm forced to do that manually over the network and in person at the pi, but I plan to add code to do that sort of thing once I no longer care how it worked last year etc - else the sheer amount of data one collects simply won't fit on things like a small single board computer. Not sure where I'm heading vs details, but in general I'm trying to set up whole-campus data aq and storage on a lot of parameters around campus, solar stuff, weather stuff, temps and humidities in various rooms and crawl spaces, woodstove monitoring (we have actual !Winter! here - 3f + 25 mph winds tonight...), water collection system details and control...and so on. Some of that needs "right now" control, as in "the cistern is going to overflow, or the next flush is going to empty it" kinds of issues, or the inverters are going to cut out due to low battery any second now, and quicker if you cook a glass of tea in the microwave instead of on the stove top, so I'll be using all the tools in the box. And at least one always-on display and a beeper to alert me of any troubles...

The point being, I'm not getting any younger, and the more I can automate before I get too old and weak to do it by hand, the better, and the more time/effort I can save myself running off-grid, the more time I have for playing with physics and generally having fun.

I can give more focused input if I know more about the system you want to monitor and control, of course. Mine was a little off the wall as those things go and had it's own unique set of issues. It was pretty much useless in winter, and had high maintenance issues, so I ditched it in the end, but you are unlikely to have the same issues and situation.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Re: Thermal data system

Postby Doug Coulter » Wed Jan 07, 2015 3:14 pm

Here's some stuff I've been working up for here, but it's all for "cold" rather than hot water, and can therefore use the cheap DHT-22 sensors for temp and humidity, and the BMP-180 for temp and barometer.
There's a rather large discrepancy between my baro readings (converted from hecto-pascals to inches Hg by my code). Could it be that the local weather corrects for our altitude? Water never reaches 100c here on the stove, more like 98 tops.
(I'm at something over 2k feet). Still, relative pressure is really what one mainly wants, and it's going up as the real cold moves in.
WeatherJan7.png
Click for bigger pic - this is since yesterday evening IIRC. You can see water temp go down whenever I use some...evidently I have a leak as the "honest" use was at 1271 seconds or so into this run when I went over to use plumbing.


This all got pushed way up on the priority queue as you can see from the right half - they are predicting 3F temperatures tonight, and that's a record low for here, along with winds. Worst we've had, ever, since around 1979 was sub 10f and 20 mph winds for two days, but not very sub 10 - 7-9 at worst. And yes, everything that could froze, and it was a lot of work to replace. I didn't get as many sensors on there as I wanted yet - just had to move the stuff over there or have nothing. This beats nothing.

What's cool about this is I'm monitoring another building on campus without having to go over there in person, with wireless...yes, I'm not keeping it real warm there, it's impossible so far from the woodstove and so close to the floor where I'm measuring indoor temps. You can of course see the bump in indoor temp where I went to use the plumbing, and for that time, turned on the propane heaters full blast...comfort. Didn't want to freeze to the seat! You can also see about when I quit going over there last night to push more wood into the fire...and other interesting "metadata" that an "internet of things" might collect. I'll settle for a tight security "LAN of things" myself.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA

Re: Thermal data system

Postby solar_dave » Thu Jan 08, 2015 5:35 pm

The current system has a controller on it, a Goldline GL30 and it works correctly I think. It has temp sensors, one in the collector and one near the bottom of the tank on an outlet pipe. None are wet. I would like to get the tank temp near the top where it draws off the "preheat" water for the gas hot water heater, the location after the mixing valve going to the house, along with the return from roof temp and the two locations currently on the controller (denoted on the attachment with dotted lines) . Really in the summer the preheat is way over the setting on the gas heater so it fires rarely.

Here is a pretty good description of the controller:
http://shop.solardirect.com/product_inf ... cts_id=150
Attachments
Hot Water Solar.jpg
Dave Shiels

My TED 5000 power monitoring
http://phx-solar.no-ip.info:8081/Footprints.html
solar_dave
 
Posts: 108
Joined: Tue Aug 21, 2012 12:33 pm
Location: Phoenix, AZ

Re: Thermal data system

Postby Doug Coulter » Fri Jan 09, 2015 11:56 am

Well, doesn't look like that controller is going to do you much good. Where I live it would result in freezing my hot water tank in winter by trying the keep the collector from freezing...

At any rate, if you can get access to place a sensor where you want it, you can probably put together some lashup based on a raspberry pi..I'll have to look and see what they have that will work direct-connect to that w/o needing an arduino, which I think is out there, and with not much programming you could have the pi either directly display what you want, log it, make a web page showing it and so on. I don't see any obvious way to do it with that controller, and in fact, that controller is one of the things you might want to monitor-log with the pi as well. It would take a little coding to adapt one of the sample projects for the sensors to do all that, but probably not much, and that at a fairly high level...the pi easily runs apache (and probably nginx if you're a real tweaker), and does do CGI's with little effort for things like that, it's how I'm using one of mine to operate a camera as well as collect the output of an arduino with a ton of sensors on it and log it all.

Anything that thermal-cycles and hits over 100 something F degrees on the hot side is bad news for a semiconductor sensor of any kind. Encapsulation fails, and the doping of the silicon migrates much more quickly at high temperatures, rendering even fairly linert things useless in fairly short order. Even thermistors fail due to encapsulation issues. So I'd still go with a thermocouple for the hot stuff, they live really well under harsh conditions, and you can even seal them up in glass tubing if you're the worried type (and use high temp silicone gasket seal if you're not a glass blowing type). Type K couples with some amplifier with cold junction compensation should be ideal for that.

This might work for you. Type K thermocouples are cheap to make if you have the wire (I have a lot). They're also sold already welded at the junction.
http://www.adafruit.com/products/269

Edit - this isn't the only option, but probably the best one. With SPI, timing isn't critical, and you can hang a bunch on the same wires, keeping only /CS (chip select) separate for each device. I plan to use a couple here for monitoring things that have wild temperature swings, since they'll take it (and also to include my bullet casting pot). Note that their statement of possible inaccuracy of type K thermocouples stems from two separate effects - the purity of the alloys (nothing you can do there except buy the good stuff) and the fact that NO thermocouple is linear, there's actually a document describing a polynomial you can apply to the output to correct for that part, though, and it's generally most of the error.
And here it is. Took some looking, I think it got lost in the latest round of computer upgrades.
ThermocouplePoly.pdf
Apply this math to raw thermocouple output for better accuracy.
(48.69 KiB) Downloaded 276 times


Handy link for TC data: http://www.omega.com/thermocouples.html
One of those rare companies that realizes making the customer smarter is good for business.
Posting as just me, not as the forum owner. Everything I say is "in my opinion" and YMMV -- which should go for everyone without saying.
User avatar
Doug Coulter
 
Posts: 3515
Joined: Wed Jul 14, 2010 7:05 pm
Location: Floyd county, VA, USA


Return to PC

Who is online

Users browsing this forum: No registered users and 1 guest

cron