Solar and other Green Heating

Forum: Hardware Talk
Topic: Solar and other Green Heating
started by: spark-o-matic

Posted by spark-o-matic on Jan. 31 2008,04:12

Hi everyone,

This kinda crosses between hardware, Linux and Free Softaware, and Programming and Scripting.

I just moved, about a year ago to a desert region of Eastern Oregon USA. Although it can get as cold as 0 deg F there are very few cloudy days, so i think Solar heating could be very practical.

I remember some solar heating panels on a few houses out here in the 80's so I tryed asking around the heating contractors but got no information.

On the Net I found that Solar is a very simple consept and the parts used to build it can be very inexpencive, but the computer used to control it is the most expencive part.

So I started playing,

I live in an older moble home on just over an acre with an ancient electric central furnace. Studying the schmatic, I found it uses a 3 speed blower controled by a SPDT relay and 2 staging relays for 4 elements.

What I have come up with so far is a 2 stage computer controled furnace.

The hardware I am using is a Relay board discarded by Walmart that would have switched 8 car stereos into 6 sets of front speakers and 4 sets of rear speakers. It was capable of being controled by a touch screen, but that was never implimented. The computer is a 66Mhz 486 with 8megs of memory. I analized the signals on the board with my recording osiloscope and was able to duplicate them from the com port and convert that signal level to the TTL used by the board with just 1 resistor and 1 transistor. The temperature sensors are simple Thermistors which are read by the analog game inputs. Right now only 2, one for indoor and 1 for outdoor.

The software is written in Qbasic under MSDos 6.22 for now just for ease of experimenting without having to compile after each change.

Current features include:
Temerature raised or lowered at preset times

'Away Mode" Press A then enter the time of return. lowers the temperature then raises it back up sortly before i return

Logging of heating costs. Knowing how must electricity costs per KWH and what the furnace draws it calculates heating costs based on run time. It shows cost this hour, cost last hour, cost today and cost yesterday. It also predicts this hours heating costs based on last hours and run time so far this hour.

Graphing. A 24 hour graph is plotted with lines for each day since I first started using it and today in a different color with a thick 2 color bar at the current hour showing the cost so far this hour and predicted cost this hour. (If the graph is off the chart I know to turn the heat down).

Backup system. In the event of a failure of this system, there is an old mercury thermistat wired in parallel that turns the furnace on if it gets too cold. This has only happed once, but it saved me waking up to an ice cold house.

Things in the works:
Limiting the costs by automaticly turning the heat down.

Intigration with the NOAA Digital Tabular Forcast product. It gives hour by hour predictions for the next 72 hours. The most useful for this would be Temperature and Sky Cover Percentage. It would be nice for the heat not to turn on in the morning if its going to be a warm day or lots of sun that would warm the house.

An intelegent heating system. I want to analize old data for such things as cost v.s. indoor/outdoor temperature difference to see if there is a sharp rise at a point. I very quickly noticed a sharp rise when the wind blew and took steps to seal up air leaks.

Total out-of-pocket costs so far: <$10.00 (thermistors)

Maybe an application for a scaled down version of dsl since it runs well on old machines, and maybe a chance for me to transition from fee based proprietary and worrying about who is going to steal my code to Open Source.

Comments, Suggestions. I have read the stuff about earning income from Open Source on this site. This could be a way fun long term project.

If there is interest I may post my current code, but it looks horable with a bunch of abandoned code from experiments that didn't work, calibration, and initializing files.

Thanks,
Richard