Cheer up! Snap out of it!

Posted 2011.01.20 9.44 in Hobbies, Pointless Blather by Stephanie

You hear stuff like that now and then, when you’re depressed. Seems more likely in the first 6 or 12 months. Once you get into a really long dark groove, eventually people clue in that it’s not something you can just snap out of, and that if you could just cheer up you probably would have done that already.

About the best I’ve been able to manage is to keep myself distracted. Keep forcing myself on this electronics hobby stuff. When I’m actually plugging away at it, I’m not feeling depressed. So from that angle, it’s working.

The problem is that it’s not consistent or fool-proof, and now and then things don’t go well then it just adds to the frustration and negative thoughts, rather than being a distraction from them.

Like this thermostat project. It’s starting to feel like every time I make any progress, there’s another setback to go with it. Two steps forward, three steps back.

It’s been two weeks since I talked about the last big setback. Two weeks of re-thinking, re-organizing, and replacing dead components. I got a new enclosure. I scrapped the first prototype board. Slowly things were coming back together again.

Last night I got back to the point I was at 2 weeks ago, of incorporating the ethernet port. This time, nothing went in backwards and nothing was wired wrong. The magic smoke stayed in. The ethernet port is installed and working. Except…

The ethernet chip generates a lot of heat, even when it’s sitting idle. By a lot, I mean, it feels just slightly warm to the touch. No big deal, right? Except having a little heater in the thermostat completely fucks up the thermostat’s ability to know the temperature of the room. Gawddamnit.

I always knew the temperature sensor would have to be isolated from the rest of the electronics. In the first enclosure you can see that it’s mounted externally. I mounted it on the outside of the new enclosure too, but the ethernet adaptor warms the enclosure up enough to throw the readings off by as much as 10 degrees F.

So I removed the sensor and mounted it on little standoffs, so now it’s about 3/4″ away from the enclosure. That gives it some room for air flow all around. So far, this seems to have reduced the margin of error, but it hasn’t eliminated it. It’s off by about 2 degrees F. If that is a consistant variance, then I can live with it – I can code a 2 degree shift into the firmware. I guess.

It still pisses me off, though.

I even tested the alternative – those wireless XBee modules. Running basically at idle, they seem to warm up to about 86 to 90 degrees F. So having one of those in the thermostat would be the same as the ethernet port.

The last option, which I haven’t ruled out yet, is to remote the temperature sensor. Stick it somewhere else in the house, where it can be isolated from the electronics, where it doesn’t have to look anything marginally attractive, and then send that signal back to the thermostat.

Maybe do both – have a remote sensor, and keep the one where it is on the thermostat.

Either way, it’s little setbacks like this that leave me wondering why the hell I’m bothering. Am I ever going to finish this? Is it going to work? Or will I end up freezing to death because my home-made thermostat failed one night?

Should I just give up?


  1. justin says:

    Keep going!!! your almost done 🙂 if your worried about the where to put the temp sensor, why not use a system similar to the weather monitors that you can buy. and my suggestion is put it lower then the thermostat. as for it failing, use a failsafe temp, if a sensor or the unit fails it should just go to the failsafe temp.

    Just my 2 cents 🙂
    Keep up the good work

  2. Stephanie says:

    Thanks for the words of encouragement Justin!

    The idea of a fail-safe temperature is excellent, that hadn’t crossed my mind. I have fail-safe / sanity tests for the target temp, so if it goes wacky a safe temp will be picked. I’ll add a failsafe temp tonight.

    It’s been running almost 2 days now (on my desk) since I added the ethernet port and it’s almost consistantly off by 1 or 2 degrees. I might add a simple adjustment value, like a fixed modifier of -2deg to compensate.

    And the idea of remoting the temp is also on the table. I added the possibility of it accepting a temperature reading from an external source, which would override the internal / local reading, with a timeout so if the external stopped sending, it would fall back on its own sensor after a while.

    Cheers, & thanks again for the kind words!

  3. Bob Wareham says:

    Hi it sounds like a very good project but to overcome your problem of the 2c temperature deferential I would suggest you use a remote sensor of the value you need, Send me your schematic and I will try to help as we make temperature units for vent control.

    All the best Bob

    1. Stephanie says:

      Hi Bob, thanks for the comments! I haven’t had a chance to draw up a schematic yet (it’s on the things-to-do list!) but I did find when I put the thermostat to work, the temperature differential disappeared. I believe it’s a factor of two things:

      Primarily, moving the sensor further away from the heat source. That is, not only is it a small distance away from the thermostat case, it is also laterally separated from the ethernet module.

      Secondarily, when the thermostat is installed upright on the wall, heat rising from the electronics travels within the enclosure and out ventilation at the top — when it was on my desk being tested, it was laying flat and heat rising from the electronics came upwards into the sensor.

      So all and all, it’s actually quite accurate now that it’s installed and in the correct orientation.


  4. J-man says:

    What you’re better off doing is using a different type of box with a divider, if you disasemble the Honey-well digital thermostats, its got something similar to this. Then somehow moving the ethernet port somewhere away from the sensor. I’m saying isolate it somehow within the box.

    ALSO, you can probably most likely run this on a battery somehow, or the 24-v current that comes to those wires. That being said, I am an HVAC technician. Most are 24-volt standard, and sometimes unlikely 12-v. If you can make this feed off of that, then you would eliminate an extra wire coming from the box.

    1. Stephanie says:

      Hi J-man,

      Yeah, I have had quite a few ideas on how to improve it. I am toying with the idea of a complete rebuild to implement all the new thoughts, but for now it’s actually working quite well. I’ve embraced the ‘mad science’ look with the wires hanging off and the external circuitry.


Leave a Comment