Wow, thanks for the flood of input everyone! Didn't expect so much response to that post. Jeff5may, it sounds like my dehu defrost is very similar to what most central heat pumps run, with the possible exception of the timer to kick it out of defrost if the thermistor isn't satisfied. Still going to have to run some further testing to determine if that is indeed the case. I got cold and went inside for about 10 minutes and there is a slight chance the compressor cycled on and back off in that time with the way I had it set up, and I just missed it.
Little history, last year I had this thing running and heating my hottub succesffully, but frosting was a continual problem. Here's the thread that details my journey:
http://ecorenovator.org/forum/geothe...ater-hp-5.html
This year instead of just having the condenser dipped in the tub, which was inefficient, ugly, and somewhat dangerous, I've built a tube-in-tube type heat exchanger around the current condensing coil. I finished stripping the fins off, cut the metal brackets off(very painstaking process not to nick a tube) and over winter break will encase them in PVC pipes. I know it seems like a lot of work to avoid simply opening up the system, but for not i REALLY DON'T want to open it up. I don't have the equipment or know-how to take that step and am not ready to make the investment at this point, and a tube in tube should be pretty efficient as-is. There's over 30 ft of tubing from that coil.
You all will be happy to know I finally invested in a Kill-A-Watt as suggested in my other thread, so I can get accurate power draw readings.
Last year in below freezing weather, when I'd get frost, I'd simply turn off the unit totally for a few minutes and it would melt the frost, but usually not completely. I attribute it to heat in the refrigerant migrating to the evap. To control this I tried a simple Klixon thermostat from another air conditioner, with the sensing bulb on the suction line of the compressor. Problem was that spot would warm up quickly enough after the compressor shut off that it never defrosted much. My hope is the heat from the compressor will be enough to bring my dehu board out of defrost, and its timer system will keep the compressor off long enough to allow the coil to melt and dry. If not, Fischertechnik microcontroller here I come.
Below freezing temps will probably still be a problem with melting the frost and last night I was thinking of manual ways to defrost it. I had the idea of soaking the coil with hottub water, but I don't think it would dry, leaving me with a block of ice when the comp restarted. Resistance would be easy, but with my small evap, I may be defrosting often enough that it would seriously impact my total COP numbers. Another option I got from research is a second coil, or some tubing in front of the evap I could pump hot water from the hottub through with the fan running. It would warm the air and melt the coil quickly and aid in drying it. I already own the small mag-drive fountain pump I would use, and a coil, from the same dehu the board I'm using came from. I think I just found my solution. The relay setup for this shouldn't be too complicated either. Some dehu's accomplish this by simply reversing the fan direction so the air passes over the condensing coil then evap, with the compressor still running. The heat is enough to bring the evap above freezing. I would turn off my compressor to speed the process up.
I wish they made window heat pumps for the US market. I've seen them in many Youtube videos in other countries(like Australia) where they have 220v mains, but never 110v versions. That would make my job ALOT simpler right now. A hot gas bypass or reversing valve system would be a god-send, and at some point I will do that if I ever open up this system. I'll also double(at least) the size of the evap to increase efficiency, and build a proper ref-water hx for the condenser side. I'll also probably add a ref-liquid(glycol) HX on the evap side to take advantage of the solar water panels I have, but that's another story.
Only 26 days until winter break and I can get back to work on this project!
Adam