The Homemade Heat Pump Manifesto

[AC_Hacker]

Much earlier in the homemade heat pump thread, I mentioned how informative I found the work regarding phase change CPU cooling that appears on the Xtreme Systems site.

Well, today I found a page that is an index of that work.

With regards to the niche that these people are persuing, some of this work is right out there in the cutting edge... not available in any store, not available in any school.

The work that is highlighted on that site, focuses on chilling, and regards the heat that is produced to be of no value.

The work we are doing in building homemade heat pumps is the opposite, but the principles are very much the same.

Much to learn there...

Regards,

-AC_Hacker

[sheepdog]

I searched this thread but didnt find this, sorry if has been previously posted but
danfoss.com/United_Kingdom/BusinessAreas/Refrigeration+and+Air+Conditioning/Product+Selection+Tools+Details/DanCap.htm
is software that will help in your cap tube dimensions.
And coolpack
et.web.mek.dtu.dk/Coolpack/UK/Index.html
will help in the selection of your heatpump components all freeware.
I have been messing around with a previously broken dehumidifier for a little while now, using it as a mini air heatpump from a wind turbine generator and found that by replacing the fan with a pid dc version also with resistors on the condenser line help regulate the superheat.
hope this is of some help
sheepdog
it will not let me post the links but dancap is at danfoss, and search for coolpack

[Fred_Fredowski]

AC,

thanks for the data.

On the Goodman technical manual, look at page 20, entitled:
"Split System Performance Data -- Expanded Heating Data"
It gives the hi and low pressures, BTUs, COP, etc... in the heating mode at various outdoor temperatures -- which would correspond to the ground loop water temperature in your system. You can see that they achieve a low pressure in the 65 psig range when the heat source (air in their case, water loop in yours) is around 50 degrees. This means it doesn't have to lift the pressure up as much and the refrigerant carries more heat (since its denser at a higher temp, and heat capacity goes up w/ density). Together these conditions give a high COP (around 3.8 for the Goodman).

I have no actual experience w/ propane, but my understanding is that the operating pressures are very similar to R22. The major difference is that R290 is much less dense, so you put in alot less refrigerant by weight, but the BTU's per volume are very similar between the two. This is why an R22 compressor will work w/ R290. My understanding of this just comes from reading on the internet -- so its possible I'm missing something.

I think your understanding/procedure for superheat and subcooling sound correct. The one thing I didn't understand from your data is that it seemed like you had two different measurements for T(lo-side rfg. out) and T(hi-side rfg. out) . For example in Test #1, you listed:
T(lo-side rfg. out) = 62.7
T(lo-side rfg. out) = 53.1
I'm not clear on the difference between these and depending on which you use you get a superheat of either 26.5 or 36.1. Both of which are pretty high. Tests #2 and #3 seem to show the low pressure is declining and the superheat is climbing as the system stabilizes.

I believe typical superheat settings are closer to 10 degrees on a high efficiency system. Usually these are controlled by a TXV which regulates superheat to around this value. On a cap tube system, my understanding is that superheat is determined by charge level and how restrictive the cap tube is. I would probably try to adjust charge first and see if you can raise the low side pressure w/o causing the high side to get too high. If that doesn't work then you could try to adjust the cap tube using DANCAP or some other design tool.

Your subcool values look OK to me.

[AC_Hacker]

The previous posts have all been quite helpful, but especially the previous two.

Thanks sheepdog for the programs, they are sure to become very useful tools.

• CoolPack can be found here.
  
• DanCap can be found here.

I'm very interested in your mini air heat pump. Please post photos and further details of what you have done.

And thanks Fred_Fredowski, and also bigsmile for your suggestions and help.

Testing

I have been doing quite a few tests since my last post, trying to determine the optimum charge level. It has been slow going, partly because I'm logging all my data by hand. I'm currently looking to get digital pressure sensors, and also additional temperature sensors, which will make logging data automatic. It is also slow because it seems that it takes a while for a changed charge level to 'settle in'. I'm doing maybe three tests per charge level to make sure things have stabilized. I tell you, that's a lot of manual data logging.

I'm learning a good bit just about the testing procedure itself, and gradually recognizing details that will make for more repeatable tests. I've found that the ambient temperature alone plays a huge role regarding measurements general and performance data in particular. I'm also recognizing test improvements (like covering my test bucket with plastic film to stop eveporation-based heat losses) that give a more correct measure of performance variables, most importantly COP.

I do wish I had built a heat pump unit with an ATV, it would make optimization testing and adjustment much easier. In my parts bin, I have a Danfoss ATV with selectable sleeves (which cost mere than I paid for the ATV) that I will incorporate into the next unit.

Compressor Heating

But one thing I am seeing from the tests so far, which last no more than an hour, is the temperature rise in the compressor. I have noticed that varying the charge level will affect the rate of temperature rise, but in no case did it stabilize. In one test, the compressor's thermal limit switch even activated and shut of the compressor.

When I first got the de-humidifier, from which the compressor came, I ran some tests to see what its performance cycle looked like. Now I regret that I didn't assault the original unit with sensors and logged its operation to within an inch of it's life...

But it had two heat exchangers, and a fan which drew air past both of them. The front exchanger was the evaporator exchanger, and had a sensor attached not too far from where the cap tube sprayed refrigerant into the coils. The area where the sensor was placed, was the first area to frost up.

The second exchanger was behind the first and was the condenser exchanger, it received the air after it had passed through the evaporator.

So when running the de-humidifier this was the repeating cycle:

• compressor-ON
• frost forms on the evaporator coils
• compressor-OFF
• frost melts from evaporator coils and falls into the condensate bucket,
• (begin again)

The cycles were spaced about seven minutes apart.

So, there are several possibilities to consider:

• Perhaps this compressor was not designed to run continuously.
• I still have not found the optimum charge level.
• The heat exchangers I am using have excessive capacity and are forcing the compressor to do more work than it was designed for.
• The cap tube needs to be re-sized to allow a larger volume of refrigerant to flow through he system to cool off the compressor.

I'd be interested in what you think...

Heat Pump Pulse Controller

I'm considering that possibly a 'pulse and glide' system may be the ultimate answer, with the heat pump adjusted to maximum COP, and tested to find the length of the running period that produces the optimum COP running period, and using that as the "pulse". Then additionally, using a very well insulated water storage tank (I've also bought 50 pounds of Calcium Chloride to use for Phase Change Material heat storage experiments) to feed the hydronic floor heating surface. (speaking of which, I have discovered that aluminum flooring is available (Alumafloor), which could be a game changer regarding lightweight hydronic floors)

I have a friend who has designed an Arduino based 'pulse controller' that will function like a very slow PCM (pulse width modulator). In it's current configuration, it will have a programmable cycle length, and a programmable pulse length, as a percentage of cycle length. I have purchased a solid state relay for the power handling. Tonight I'm going to meet up with my friend and take along the heat pump for testing.

My target configuration would be that this unit monitors heat storage tank temperature, indoor and outdoor temperature and adjusts the length of the cycle period as required.

* * * * *

By the way, I'm getting close to building my next unit. I have several compressors to choose from, and I will be ordering the heat exchangers to go with, very soon. I'll post the particulars and I invite your design input.

Best Regards,

-AC_Hacker

[sheepdog]

My compressor also switches off for a period of time after an hour or so even with the humidity stat on max, a timer on the pcb board seems to trigger for the defrost cycle. I will remove it completely on the next prototype.
If your high temp stats tripping (too high superheat) it probably points to either a low load on the evaporator or that your charge level is to low. Better this way though, than too low a superheat and liquid trying to be compressed, try adding a little more charge at a time if your evap load is satisfactory.
Regarding your ssr if you are going to burst fire your compressor be aware that with such a large inductive load, the current demands on each pwm cycle especially if you are going to be zero crossing will be massive and it might not last long.
I will have to find my camera and put up some photos of my unit but it does not look much.
Keep up the good work,a very interesting project.
All the best
sheepdog

[AC_Hacker]

Hey, I have a dandy program for you Sheephead, in case you ever want to get your feet wet and experiment with water loops...

Here's a brazed plate sizing program for refrigeration:

FlatPlateSelect

You have to give them somebody's name & email, etc., but it seems to work pretty damn well.

It's pretty interesting, when you plug in various values and then see how it affects the results, you really begin to understandow this stuff works.

Originally Posted by sheepdog

I will have to find my camera and put up some photos of my unit but it does not look much.

You should know by looking at my heat pump on the first page of this thread, this is not a beauty contest.

Regards,

-AC_Hacker

[AC_Hacker]

I have re-configured my heat pump to work as a Ground Source Shop Warmer. It is now running, warming my uninsulated basement shop.

I borrowed an automobile radiator from Bruce-the-Pirate and hooked up the output water loop (and circulation pump) to it. The setup, crude as it is, is working.

I'm using a box fan to blow air through the radiator.

The ground loop temperature pulled down to about 46 degrees... the radiator-loop temperature stabilized at around 73 degrees and after running it for several hours, the temperature in the unheated basement stabilized at 60 F... not too bad for a shop.

I haven't calculated efficiency for the setup (my estimate is COP =2), but it is clearly better than the 220 V, 2400 watt resistance heater I was previously using.

-AC_Hacker

[AC_Hacker]

I have decided to build another heat pump. The first one I pretty much put together by guess, supported by a lot of reading. This time I'm going to try to actually design it.

So I have been looking over my compressor stable. These are compressors from units I bought at garage sales, friends have given me, or in some cases, just showed up in my front porch. I listed them in a spreadsheet to better see what I have to choose from:

The little compressor in the bottom of the list, the LG QS064CAB is the one that I built the first heat pump from, it was originally in a 25 pint de-humidifier I got at Goodwill for $25 as I recall.

Bearing in mind some advice I got from Fred_Fredowski, I've kept in mind that there would be different loads on the eveporator HX and the condenser HX. The columns for evaorator load is equal to the compressor's rating. The column for condenser load is the compressor's rating plus the heat created by the running of the compressor(electricity turned into heat, and the heat generated by compressing gas).

I'm also looking over the selection of brazed plate heat exchangers on ebay, and I sorted out the exchangers that 'dudadiesel' has, since his offering is pretty varied.

I know that in the final analysis, the sizing of the HX comes down to area. So the selection process will be to determine area, then find the best price for a HX with the required area. I included the "Pressure" ratings, because Propane pressures can reach 300. R-22 can reach 400, R-134a can reach 450 or better, and R-410a even higher.

It is very interesting that there is so much variation in price/square foot. I left some of the exchangers off the list because their prices were substantually higher for no apparent reason.

Also, I got some advice that it might be advantageous to put more than one HX in series with another, to get the required area. I'm really thinking this one over... the price might be lower, but the trouble would certainly be greater.

And as an additional factor, the HX shoud have sweat fittings on one side. My curent HP has a slow leak, and it looks like its coming from a pipe-thread fitting. There may be a way to seal it up but I don't know what that way us yet... teflon tape seems to have its limits.

I have searched far and wide for brazed plate selection software, and the only thing that I have found is the FlatPlateSelect program. If anyone knows of other programs or selection sheets, or other methodologies, I'd sure like to know about it.

I have spent a good bit of time with FlatPlateSelect and it is very interesting what can be learned by changing various values and seeing how those changes affect the outcome.

Factors that I hadn't previously taken into account were the rate of flow of the refrigerant and also the rate of flow of the loop water. Both of these will have an effect on the required size of the heat exchanger.

Here's a selection chart I just came across for Alpha Laval Brazed Plate HX.

• This page has links to detailed product information for the "AC" series
• This page has links to detailed product information for the "CB" series

I have come across the rule of thumb that loop flow rate should be in the neighborhood of 3 gallons per minute per Ton. I noticed that if I adjusted the water flow rate down, since my heat pumps are a fraction of a Ton, the FlatPlateSelect program adjusted the HX area up, sometimes quite a bit.

I was reminded by a fluids engineer that turbulent flow is desireable in heat exchange applications. more on that later.

This is going to take some consideration...

Comments are invited.

Regards,

-AC_Hacker

[BradC]

Originally Posted by AC_Hacker

And as an additional factor, the HX shoud have sweat fittings on one side. My curent HP has a slow leak, and it looks like its coming from a pipe-thread fitting. There may be a way to seal it up but I don't know what that way us yet... teflon tape seems to have its limits.

Have you considered 45% silver rod and just brazing up the fittings? You need some good flux and quite a bit of heat, but it should be doable with Mapp gas. I've certainly brazed copper-brass-steel here by just brazing up the fittings after they are screwed together.

You have to be _really_ careful brazing up to brass though. I melted the first two fittings before I got the temperature right. It became _much_ easier when I used 45% silver rod though.

I know you can braze pretty much anything to anything with the right rod and flux.

The problem you get with interference fittings like you are using (even with teflon tape) is the uneven coefficient of expansion of the metals when subjected to extreme temperatures. You screw them together nice and tight at ambient, then you shove evaporating propane into them and drop them right down below the frost point. The brass shrinks slightly more than the steel and you have a leak.

For leak detection, add plenty of glycerine to the soapy water. It sticks around _much_ longer and you get to pick far slower leaks. There is better stuff available from refrigeration suppliers that appears to defy the laws of physics with how long it leaves a workable film, but glycerine is cheap and you leak sounds like it's big enough to be picked a bit more easily.

I've been watching this thread pretty closely as it's similar to something I've been working on myself. I'm lucky enough to have a huge aquifer about 8M under my place, so I have a couple of separated bores and loads of 20 Degree C water to play with. I'm rejecting heat into it rather than sucking it out however.

[AC_Hacker]

BradC,

Hello there in Oz and welcome BradC, to the conversation.

I really appreciate your idea about brazing. It had already occured to me to do that, but the thought of how much heat it was going to take to braze the threaded pipe connections stopped me in my tracks... I was afraid I might un-braze the HX. Now that I know that it is possible to braze this up correctly, I'll give it a try.

Your GSHP cooler project sounds wonderful.

• What are the daily high temperatures you have there?
• And what is the temperature of the water you are bringing up from your aquifer?
• Are you also using Propane?
• What is the capacity of your compressor?
• What type of metering device are you using?
• What are the details on your heat exchangers?
• How about your water pumps?
• Are you using one well for a source and the other as a sink?
• Are you using this system to cool the air?
• How big is the air HX?
• How many CFM are your fans?

If it is possible for you to do so, I would really like to see some photographs of your project.

If I recall correctly, the very first GSHP was done by a guy in New Jersey who buried copper condenser lines directly in the ground to get heat. But the first really organized development took place in Oklahoma, where it gets very hot in the summer. The Oklahoma work was originally started as a way to cool, and only later was the heating potential of such a system developed. So your work, using the ground for cooling, has a very strong tradition.

And thanks so much for writing, I'm up here in Oregon struggling with the cold, I had completely forgotten that in the other half of the world, people are struggling with the heat. Thanks for the reminder, it makes my plight just a bit more tolerable.

Good Luck to you,

-AC_Hacker