AC Hacks a Heat Pump

[AC_Hacker]

Originally Posted by randen

Is the reversing valve spring loaded or would it have a pilot solenoid on each end.

It has one solenoid on one end, and only two wires, so I'm sure it's spring loaded.

You can see the single coil and the two black wires coming out of the green coil structure.

Originally Posted by randen

What is your thought on the output tank? I find that the floor responds so slowly short cycle won't be your problem and the temp is so low. I would think you could plumb your floor directly from your HX.

Well, my house is about 120 years old, and the floors I'm going to heat are suspended, so I don't think I'll be pouring a slab. It looks like I'll be doing some kind of low mass floor with aluminum spreaders, along the lines of what Vlad on his floor.

So the tank will give me the thermal mass that concrete gives you.

Chop, Chop!

-AC

[Xringer]

Originally Posted by AC_Hacker

Yeah, this sounds great. One problem with that approach that I'm wrestling with is, assuming that the TEST KEY is a toggle (ON & OFF),
then how will the system know what state the heat pump is in? I mean, if the process skips a beat, then the unit will go OFF when heat is required, and go ON when too much heat has been reached.

Best,
-AC

I've been thinking about the test button, and what the control table enables you to do.?.

• Once push : Ignore Communication error.
• Twice push : Comp. ON, / Outdoor Fan ON / Reversing Valve Coil ON(only Heating Model) / Crank Case heater Coil ON(only Heating Model)
• 3 Times push : Comp. ON, Outdoor Fan ON / Reversing Valve Coil OFF(only Heating Model) / Crank Case Heater Coil OFF(only Heating Model)
• 4 Times push : Same
• 5 Times push : Error LED(Red) ON
• 6 Times push : Comp OFF, Outdoor FAN OFF
• 7 Times push : Reset


The diagram shows where you need to attach a logic zero (ground).
Each button press puts a Ground pulse (0 Volts) into R58, that feeds Pin 1 of the micro controller.
I would use an Opto Isolator, to keep outside power supplies, and static voltages off Pin 1.


By experimenting with the button, you should be able to tell, if there is a time-out for a test mode run. (My guess is No Time-out).
Leave it in heating mode for hours.. If it stays, you don't have to re-command it..

~~~
Next item, is there a standard 'command' time-delay? After you push twice (heating mode),
can you just wait 5 seconds and push it twice again? And stay in heating mode??

If there is a few seconds of command time-delay, then control will be easier.

The nice part about using a Micro controller (like the CAI board),
it can generate the required number of pulses, very accurately.
If the programming of the pulse width & spacing is done right,
(to simulate a finger on the button)
there will be no need for feed-back, you can just assume it's in the requested mode.

If it did fail to be in the mode requested, the pulse program needs tweaking.

This is the perfect app for the CAI board! It's temperature inputs, used with
a TTL input from a thermostat would be super easy to implement.

[AC_Hacker]

Originally Posted by BradC

Have you been *really* lucky and got a unit with its factory gas charge still in it?

I went down just now and loosened the hex screw valves, and alas. there was no hissing sound.

On the cheerful side, the valves were closed, so moisture & crud wouldn't be able to get in.

I have no clue as to what the conditions were when it was de-gassed.

Best,

-AC

[ham789]

Interesting project.
I have some suggestions.

Before you go too far, can you test the compressor?
AT least pump it down and leak test the system?
Maybe try to test for plugged capillary?

As for the electronics, I'd
disconnect the juice to the compressor.
replace it with a light bulb.
put an indicator light on the power to the 4-way valve.

Now, you can pulse the test switch to reverse engineer
the operational parameters without having to run the compressor.
If it looks good, parallel the switch with an opto-isolator and
start banging it with the arduino.

For a single pin, the serial or parallel port of a laptop will let you
experiment with the signals. I bet you know someone who has
a pile of old laptops ;-)

[Geo NR Gee]

Ham, thats a good idea with the light bulb. I am going to have to remember that one!

[Xringer]

Yeah, that's a good idea! I should have come up with that one, since I've been a Ham since 1968..

I would also rig up a second light blub with spade lugs for the crankcase heater element.
That will tell you if the ASHP is operating in winter mode..
And avoid overheating the crankcase, in case it's dry..

[AC_Hacker]

Yesterday I attempted a pump down, and never even achieved 1 psi difference from atmospheric pressure.

At first I though that my Little Wonder (wonder how long this little thing will last?) vacuum pump I bought from the local Chinese Tool Store had left the world... but a test of it shows it to be working just as it should be.

So now I need to find the leak... maybe there's a bullet hole in the heat pump after all.

I'm thinking of blowing some compressed air into the system... that would at least give me some hiss noise to help localize the leak... then I could use soap bubbles to locate the exact spot.

This is also bringing into question the oil in the system, as it has been open to air for as few as three and as many as five years.

-AC

[Xringer]

Pressurizing should give you a hiss to find. If you have a 2 foot hose that you can hold up
to one ear, while fishing for the hiss with the other end. Might make it faster.
I hope the hole isn't so big you can't get enough pressure to create some hissing noise..

CO2 might even give you a visual on the leak.. Good luck!

[BradC]

Originally Posted by AC_Hacker

I'm thinking of blowing some compressed air into the system... that would at least give me some hiss noise to help localize the leak... then I could use soap bubbles to locate the exact spot.

Compressed air would do the job. You could always use a little propane and a lighter for a leak detector . If you can't get any Vac at all on it, it must be a pretty big hole.

Originally Posted by AC_Hacker

This is also bringing into question the oil in the system, as it has been open to air for as few as three and as many as five years.

Being an R22 system, it'll likely be Mineral Oil. Just fit a *big* liquid line drier when you eventually fire it up.

[AC_Hacker]

I tried again today to find the leak...

I made an adapter to get from bicycle tube valve fitting to 1/4" flare so I could blow a lot of air into the heat pump, and hooked it up to the heat pump.

I have an air compressor with a pretty large tank, but I wasn't able to find the problem. I put a pressure gauge in line with the air compressor, and it's pressure would rise to about 80 psi, and quickly fall back when I let go of the air lever... but it seemed to hold at 50 psi, at least for a while... so strange.

I should be able to hear all that air leaking out, but so far, I haven't. I guess it's leaking out so fast that it doesn't build up pressure for me to hear any hissing sound.

I tried to find the leak with a smoke stick... but so far, it's evading me.

I tried the vacuum pump again this morning, but it just ran and ran and ran and after maybe twenty minutes, I could see from the manifold gauge that some kind of vacuum was occurring, but all the while, there was a major amount of air issuing out of the vacuum pump... lots of air... like the vacuum pump was just running free.

I also tried a micron gauge/vacuum pump test, which tells me that I need to get a fresh battery, and clean out the thermistor sensor with alcohol, and get fresh new vacuum pump oil and get a solid vacuum reading.

But there is something strange going on with the vacuum/pressure thing.

I wish I could just hook the air compressor up it and throw the whole thing into a tank of water and look for bubbles.

This should be the easy part.

-AC