Holding a good vacuum. Damn it to hell!

Its has been a very hard slog and I have to admit temporary defeat on being able to hold a low enough vacuum to be able to get the water evaporating quick enough to cool the water.

I have been working on putting together a test bed to play with different refrigerants and desiccant material. I have done at least 10 iterations on trying to hold the pressure low enough to get the action happening with water but alas it continues to leak the smallest about but it is sufficient to stop the evapouration process and the temperature of water very quickly starts to rise after the vacuum pump is switched off. The pressure only rises a few tenths of a psi but that is enough to stop the process.

I have been working through each connection point and valve to isolate the leak. A trick I happened upon is to systematically submerge the apparatus one joint or valve at a time. The viscosity of the water is sufficient to plug the hole temporarily and it helps isolate the leaks. I immediately found the largest leak which was a faulty solder joint on one of the corners and it seems that the gas shot off valves are just not up to the job of holding a vacuum. I'm guessing they are engineered for positive pressures to make the seal.

Reading around it seems that the best valves are Globe Valves which also seem to be called shut off valves but instead of the 1/4 turn of the gas shut off valves these taken multiple turns and inside there is an actuator that tightens against a seat to shut it off. It seems that it is much like a normal water tap except that there are o-ring seals on the shaft that turns to close the flow so that it doesn't leak when in the open position. Makes sense in hind sight.

So here is a close up of the appartus with the gas taps.

As you can see lots of teflon tape, buckets of silicone to try to seal it and i even resorted to tacky tape ( the black stuff) for sealing. Eventually I went back and resoldered all the joints, changed the olives in the compression fittings and re-tapped the copper capillary joints i am using to put the quick connects on with and it would hold a vacuum level that was about 0.2 psi higher than when the vacuum pump was running. This 0.2 psi seems crititcal. Unfortunately i can't tell the exact pressure because the vacuum pressure sensor I've got is only rated to 1-30psi.

As an a side here is evidence there is a lot of force exerted on the vacuum vessel. I got down to 1 psi using this fuel can as a vacuum vessel and the bang the fucker imploded in a spectacular way

So I replaced the fuel can with screw top jars, lashing of silicone, tacky tape etc to stop the screw thread from leaking around the jar top. No go, leaked like a bastard. In desperation  (just before giving it away for the day ) I cut a seal out of the anti-fatigue mat i was standing on, screwed one of the plumbing fittings into a flat piece of perspex. The idea being that when the pressure falls the force on the perpex squeezes down on the seal and creates a tighter and tighter seal. Goddamn if it didn't work!!! At last a success! It does solve the issue of the low level leak which i am pretty sure will require the valves to be replaced.

So next step is to replace the valves. i might see if there is any way that I can reduce it down to one.

The plumbing

Finally have some progress to report.

The wine bottle and tacky tape set up was a great way to see the cooling effect in action. I didn't have any check valves or the like in the system so I had to have the pump running all the time. Regardless, it at least demonstrated the cool effect of stimulating evaporation by lower the pressure above the liquid.

Time to move on and get a bit better set up to use as a test bed for playing around with different combinations of working fluid and desiccant material. The simplest seemed to be 1/2" copper tubing and plumbing fittings. I made a few false starts with the valves. At first I tried the type that just just push on over the pipe ( no compression fittings to tighten down). They have an o-ring seal on the inside of each end of the valve and a locking mechanism that prevents the tube from pulling out once its inserted. The problem with these is that they don't prevent the pipe from rotating and so the whole structure had to be secured to a board to stop things moving around.

The second problem I encountered is it seems that for low gas pressures these things leak. With the benefit of hindsight I'm figuring that they rely on the viscosity of water to form the seal. With gas at low pressure the air outside the valve will be about 14.5 psi higher than on the inside and that is enough for the air to get in.

Second iteration was was some gate valves, again a plumbing fitting, that secured to the pipe using compression fittings. These fittings have a nylon washer with top and bottom surfaces that taper towards the edge ( called an 'olive') that compresses as a nut is tightened over them and both squeezes in on the pipe and creates a seal. The seal around the pipe was great but these valves leaked worse than the previous ones. from the design is seems that any grit that collects in them will prevent gate from closing and allow gas to be drawn down from outside through the thread that supports the valve handle.

Finally I went to out local gas appliance repair shop and got a couple of gas shut off valves (yellow handles below) and secured them with compression fittings from the last try and the result was much better. Not perfect. It is still leaking about 0.1 psi each 15 minutes or about 6 mBar each 15 minutes. This would be way too much for a real system. according to one paper I read a change of 20 mbar in the pressure during the absorption phase was enough to almost entirely eliminate the cooling effect. However, I think that the pressure sensor connection and the plug i was using for one of the open ports is most likely the culprit.Anyhow it will do for a test bed as I will just periodically evacuate it.

I added a couple of air compressor quick connects for the vacuum pump and here is the plumbing part of the test bed. The next step will be to make the evaporator and the vessel for holding the desiccant material.

The tap on the vertical shaft is for shutting off the vacuum line and the one on the horizontal pipe is to isolate the evaporator (to go on the left) and the desiccant material (Zeolite, silica gel or calcium chloride will be the different ones I'll try)

The experiment setup

This whole area just smack of DIY and I think that is a great part of the appeal. In addition the science behind is just on the cusp of understandable for me and lends itself well to garage experiments.

The first thing that I wanted to do was just to see one of the key affects in action. Namely, chilling water by simply dropping the pressure above it. This just blows me away. Even more out of step with everyday experience is that if you keep dropping the pressure do extremely low pressure the boiling point of water approaches and then goes below zero. There pressure has to be very low: at 0.3 psi its around 17 degrees, drop that to 0.09 and boiling temperature is .. wait for it zero degrees!!! The boiling point is very sensitive to the pressure level at pressure below 1 psi. At 1 psi the boiling point is around 40 degrees and then at 0.09 its zero degrees.

In my experimental set up in the garage, although the pump I have says that it can go to 5pa which is about 0.0007 psi I have my serious doubts about it being able to achieve and maintain that. Time will tell what is needed to get that low.

As a side note its important to make a distinction here between evaporation and boiling. Evaporation is a phase change from liquid to gas that occurs just at the surface of the water. Boiling is a phase change from liquid to gas that happens in the bulk of the liquid and not just at the surface. Boiling would clearly liberate a far greater number of molecules than evaporation alone and would be very desirable but I have no idea about the practicality of achieving and maintaining that pressure in a set up that bumps around off road in Australia. Again, time will tell.

So here's my garage level understanding of the process of chilling by evacuating the system.

When we measure the temperature of the water we are measuring the average energy of the water molecules that are colliding with the temperature probe. However, the energy of the water molecules throughout the water has a distribution of energies. I'd hazard a guess and say that like so many naturally occurring distributions the distribution of energy is a Gaussian distribution / Bell Curve.

Anyhow, so as you drop the pressure above the surface by removing air the higher energy water molecules have a greater chance of escaping the surface of the water as there is less air to get in the way.  As a result the molecules escape i.e the water evaporates or changes state from liquid to gas.
As a result of the higher energy molecules escaping

The energy required to vapourise water turns out to be 2260 KJ/kg. This is a lot of energy and apparently 5 times the amount of energy needed to heat water to 100 C.

I was very keen to see this process in action so I made a simple set up my garage with a pressure and temperature controller that I made for some work that I do with epoxy. It lets be see the temperatures and pressure real time and log them.

So here is the very simple set up. A vacuum pump (bought on ebay and used for evacuating air conditioning systems to recharge them) and the 'evapourator' is just a wine bottle with a small about of water in the bottom. The black tube and wire you see are the pressure probe and a thermocouple wire for measuring the pressure and the temperature. 

Here is the whole set up. The keyboard is sitting on top of the pressure/temp data logger/controller which is made using a great little kit called a Maximite which uses a PIC 32 microcontrol but has a BASIC interpreter built in so you can program it and read from its 20 I/O pins. Great project in its own right.

So here is the cool part.The screen shot below shots the cooling affect due to the pressure reduction. The horizontal line is ambient temperature of about 22 C and the falling line is the temperature of the water in the evacuated bottle. In this case the time axis shows about 1/2 hour.

The bottle was just standing in the open air so when the temperature his about 10 degrees is stopped getting cooler.At this point it was taking in  heat from the surrondings at the same rate I was taking out via the maintained vacuum. The following evening I but the bottle in an insulated bag and it got down to 5 degrees. 

A couple of days later I filled the bottle up with a lot more water (about 1/3 full) and logged the data. after 83 minutes the temperature fell to around 13 C from its initial 27 degrees. After that its rate of cooling was very very slow.

As can be seen the rate of cooling is much slower as there is a great deal more energy is needed to be removed to drop the temperature. However, the cooling was much slow that I expected and one possibility is that the surface area of the water exposed to the vacuum is too small to allow a sufficient number of molecules to escape. I'll play around with this next time.

Towards a cooler fridge

I love travelling in my Hi-Ace van and for the last few years me and my wife and been travelling north for the winter to sunny, windy Far North Queensland as the Trade Winds kick in and Sydney gets cold and shitty.

The humble inverter was the first big lurch toward freedom. No longer dependent on landing in a caravan park every night we were free to drive and compute and kite where ever the winds of fortunes blew us. However, we are not fully disconnected from the Matrix. We still need to service out ice dependency. 2-3 times a week, 2 x 5kg bags of ice go into the Esky to replace the buckets of cold water relentlessly invading every available orifice of our food stuff. The water, usually wafting of gently rancid milk, claims the labels of every can, bottle or jar not designed total immersion. Damn you to hell esky!

As with all problems to be solved I undertook a short research program and the university of Google and as a foundation member of Over Engineering Anonymous I soon found myself fascinated by this age old art of refrigeration.

RV fridge options do exist but the energy intensive nature of them puts me way off them:

Compressor fridges are a very good option. But they are expensive and energy intensive and I would need to upgrade the 100Ah battery or add solar to the van to make sure that we didn't use up the battery and be unable to use it for running out laptops and phones and printer when working on the road.

3 way fridges I believe also have some merits and are widely used. They do however, need a constant heat source which means putting it outside to use gas and running of our precious limited battery where it is not particularly efficient. Moreover, unlike the compressor fridge it cools relative to ambient temperate so probably still needs to be used in conjunction with some ice.

In the course of finding out about the fridges and finding out how they worked I came across a lot of research on a different type of absorption fridge that introduced me to a lot of very exciting ideas and has the potential of a very DIY fridge or ice maker to be knocked up in the garage.

The absorption process in a 3 way fridge uses either ammonia or a lithium salt dissolved in water. The use of these material makes it tricky or risky for the DIY'er. Ammonia, well it can kill or blind you and lithium bromide is difficult to get hold of as I believe the the presence of lithium means the material is restrict because of its use as a 'psychiatric drug'.

In these fridges the refrigerant is initially dissolved in a liquid that is in a low pressure environment. By heating the liquid the refrigerant is driven out of the solution as a vapour which in the process of changing state and by the powers invested in it by the laws of thermodynamics takes with it heat from its surroundings and the cooling effect on your beer ensues. The vapour then passes through a condenser where it expels much of the heat and then drips back down into the original solution from whence is came and the whole process continues. This is fittingly called a 'vapour adsorption' process.

The total amount of energy it can extract from the system is the amount of energy in addition to the heat source required to vapourise the refrigerant less the amount of heat that can't be expelled through the condenser.

The research that I came across is for an absorption fridge that uses water as the refrigerant and instead of the refrigerant being absorbed back into the solution straight after clearing the condenser it uses a variety of common moisture absorbing materials such as activated carbon, silica gel (like in pharmaceuticals or vitamin table bottles) or Zeolite. The role of these 'desiccant' materials is to take the water vapour (which is the refrigerant) out of the system and prevent the pressure from building up as this would increase the temperature at which evaporation would occur until eventually the evaporation would stop and so would the cooling of the fridge contents.

Unlike the ammonia or salt/water vapour absorption fridges the process is not continuous. The cooling happens as long as there is evapouration of the water happening. The ultimate limit on this is the amount of desiccant material there is because once it is saturated with water it can no longer prevent the pressure rising and the cooling slowing down until it stops. In practical systems this can be 12 hours. This compared to relatively short duration of the boiling /condensing/ absorbing process in the 3 way fridges.

This is both a pro and a con. On the pro's side this means there is 12 hours of cooling occuring with no energy input to the system. Practically this means no burning, no electrically heating, no noise just chillin'. Awesome. The con is that the process then needs to be reversed by heating up the desiccant material to drive the water back into the system. During this time there is no evaporation going on so no further chilling but the cold water or ice created during the cooling cycle acts as a cold 'bank' to keeps things cold while the system is reset.

One of the very exciting things about this is the multitude of 'low grade heat' sources that can be used to reverse or 'regenerate' the system. Any source of heat capable of 150-250 degrees will do. Solar collectors, heat off the car engine, wood gas burner, bio gas off a digester or LPG. Most exciting is a solar collector which some of the research suggests has real promise and especially if you use methanol instead of water at the refrigerant then the temperatures that can be easily reached with a collector are more than sufficient to 'desorb' the methanol out of the desiccant. further, the 12 hours cycle time of these fridges matched perfectly with the day time/ night time cycle. Spooky!

All of this has got me very excited about the prospect of making a DIY solar powered fridge or even more appealing is a solar ice maker to replace our 2x3x 5kg habit and allow us to finally be free of the matrix!!!

As the science is so interesting I am planning on running a few experiments to see all of this in action and come up with some real parameters for a prototype fridge. As luck would have it I over engineered a pressure and temperature controller using the truely outstanding single chip computer/ microcontrol developed by Geoff Graham called the Maximite. The add on box I built has 4 temperature sensors and  vacuum pressure sensor and all this can be display real time and the data logged. This is just what I need to run experiments to fine tune the requirements for a solar ice maker.

Bring it on!!!