My journey from liquid to vapour

Beccara

n00b
Joined
Jun 1, 2004
Messages
41
Well, i posted this during the down time so its all dondensed up

For some time I’ve wanted to get into phase change cooling and I’ve started to take the plunge and I’m getting my tools together, this thread is a journal of sorts that I plan to update with my learning curve and progress in building, I will also use it to ask questions etc etc
If I list any prices they will be in $USD

The first thing I needed being a total newbie was the basics. I went to refrigeration basics and started with the free samples section given I have a very limited background in physics I skipped the intro and went straight to http://www.refrigerationbasics.com/1024x768/rb2.htm

This proved to be most helpful in the basics, which is all I really needed. At this point in time its not worth being over confused with cascades or HX Exchangers, Start small, Start basic.

I decided my first project would be cheap in components so I wouldn’t waste money should I break it. A common place for people to start if with a old fridge, so I got one. It was working and the freezer section would frost in 20min so I just cut suction and discharge with a pair of garden scissors, Its all I had at the time and did the job well.

NOTE: Discharging to the atmosphere is illegal is a lot of countries and some gases may be toxic, I had access to a respirator at the time and cut the pipes outside on a windy day.

The next step was to pull the compressor out as well as the wiring that goes with it, to this point I still don’t know how to power the thing yet as I’m still learning about compressor start-up and running coils + caps

To recap,
At this point I have limited knowledge of refrigeration
No tools
1 used unknown specs compressor with electrical bits
Some Copper piping

At this point it was clear before I moved along I needed tools so I started to save to buy them. Condensers, evaporators etc weren’t much use right now so I put the hunt for parts on the back burner, yet as it turns out I ended up with my condenser being given to me from a old AC unit long before I got any tools

I needed to make a list of the tools I would need, after looking on forums and thinking about it this is the list I came up with

Hot Gear
Firebricks
Torch of some sort
Brazing rods
Flux
Mild acid
Dry nitrogen
Welding Clothes and goggles

Pipe Tools
Flaring kit
Swaging tool
Pipe benders
Pipe Cutters

Other tools
Vacuum pump
Charging gauges
Vacuum gauges

From that list I set out to price up the tools and also had to decide on the type of torch I would use

Hot Gear
Firebricks $5 each
Welding/Cutting/Brazing Kit $150
Oxygen/Acetylene Tanks + Gas $80
Pro Silver 15 Brazing Rods $1/each
No Flux Needed
Dilute H2So4 Acid $19.95
Dry Nitrogen Tank + Gas + Regulator $80
Shade 5 goggles + 2 sets of lenses $19.99
Welding gloves $10.99
Welding apron $40

Pipe Tools
Flaring kit $50
Swaging tool $10
Pipe Springs for 1/4” to 5/8” $40
Pipe cutters (normal + mini) $60

Other tools
2hp Air Compressor with threaded intake to use as vacuum $100
R410a Manifold with 60” 2000psia Hoses $130
Vacuum Gauge $40

All up that’s $855 WOW!

Now you make ask why I don’t just spend say $200 and get the parts together and ship them off to of an AirCon place to get assembled. That’s a good point and I thought about it long and hard but in the end I decided it would be best to learn this all myself and I would be able to make a few units and sell them cheap to the people around my region (Wellington, New Zealand) to recover the costs. Thankfully I don’t have a wife or house to pay off yet and it will only take a month to get the cash together.

Right now as I save up I’m reading up as much as I can and welding/braze/cut pip as much as I can in order to get my skills and knowledge up to the level required.

At this very point in time I’m working with a program called CoolPack (http://www.et.dtu.dk/Coolpack/UK/) and the wide range of P/T charts I have converted from PDF to XLS and HTML formats to compose my system design and look into fail-safe

As luck would have it during a brain picking chat with a HVAC engineer on the phone I was offered and have picked up 3 bottles of supposed R406a, R407c and R410a which when I get my gauges I will control the temperature to 15°c and measure the pressure to confirm correct gas and purity. These bottles which are 50-70% full would have cost over $900 to buy full here and I’m lucky to have scored them

That’s all I can think of for now, no doubt more questions will arise and I will get more gear together



Ok, time for another update.

I went to BOC gases today and had a meeting with their account manager and walked through what i would need blah blah blah and after 2 hours of really good talking and business advice i walked out with a tank each of oxygen,acetylene and dry nitrogen along with some welding filer and welding goggles.
Next stop with the local hardware shop to find a air compressor, The air compressors were plentiful but air compressors with intake ports weren't, i had to drive by about 4 shops to find one but i did, 2 hp with a 25L tank.
After all of this my wallet had gone from 5y.o kid at ironclads to a catwalk model, I only had enough cash left to buy the welding kit from a local auction site. I got all the gear home and tested it out, sure enough every worked except for a acetylene regulator, which on its output valve had developed a leak. I haven't gotten around to fixing it yet, I'm sure all it needs is some thread tape

I've been given a total of 4 tanks of refrigerant by now and i came to the task of choosing which one to use. I had r406a, r407c and r410a. The 406a was ¾ full, the r407c was ¼ full + a new full tank and the r410a was barely used.
From what i have picked up you can tell a lot of basic info from a gases P/T chart and what i work with right now is the pressure at -40
r406a is 16” of vacuum
r407c is 3PSIG
r410a is 11.1PSIG

now in order to have a context for those numbers lets throw in something you might be familiar with

r134a is 22.2” of vacuum
r404a is 5.4PSIG

Now after looking at what i had i should just throw the r406a away which i may do but its not costing me anything to keep so this leaves me to choose between 407 and 410, to do this i look first at what i know and what pressures i feel safeish working with so i look up the scale at temps of 55f (44c) and see that

r407c 108.2 PSIG
r410a 157.4 PSIG

after much thinking i have decided to use r407c for a number of reasons, the first being the lower pressures involved and among other reasons i have a lot of it and would prefer to keep the good stuff until such a time i can put it to go use.

A rather productive day even if i do say so myself, payday is coming up in 2 weeks and that will most likely go towards a compressor + some tools, which i am only left with

Hot Gear
Firebricks $5 each
Pro Silver 15 Brazing Rods $1/each
No Flux Needed
Dilute H2So4 Acid $19.95
Welding gloves $10.99
Welding apron $40

Pipe Tools
Flaring kit $50
Swaging tool $10
Pipe Springs for 1/4” to 5/8” $40
Pipe cutters (normal + mini) $60

Other tools
R410a Manifold with 60” 2000psia Hoses $130
Vacuum Gauge $40


The most important thing i need to get this payday is the r410a manifold then all of the pipe tools, which i should be able to afford with this payday's cash. Right now I'm going to get some sleep and start planning some basic layout and case design

Good Night All



ok well looks my big ass P/T chart is in the wrong temp range :p
Teach me for doing this while i'm tired
www.wtfhax.com/ptchart.html

Based on Cool Pack

Temp is C and pressure is PSIA
@ -40c

134a = 7
404a = 19
406a = Vacuum
407c = 12
410c = 26

@+40c
134a = 147
404a = 263
406a = 126
407c = 219
410c = 347

So as you can still still the same result i'm going with the 407c
 
dude i am selling my vapo xe for 550 shipped if you are interested.
 
If you have a lisense to buy R22 its the best price to performance ratio refrigerant...
 
Why go through all this? Just tear up an old air conditioner/fridge! You'll get yourself some nice tubing with copper fins soldered to it. You also get a compressor. Building your own phase change refrigirator is very dangerous and is bound to fail. You don't even know the specifications for your pump! What if it doesn't generate enough nominal pressure, or has a low feed rate? Or what if it's too high? Your loop will burst. And you're spending twice the amount of money it costs to buy a top of the line phase changer already made for you, while the results of what you will end up with are highly questionable!

Sigh. And people say I have no common sense... :eek:
 
Its dangerous but not "very" dangerous unless your a retard with R290. Its not a bad idea either and a direct die system can get much better temps than a window AC chiller.

But yea this guy has ALOT more to learn.
 
Anyone can buy a phase change unit and unpack it...

This guy is learning something new by building it. Props to you, and good luck.
 
M4d-K10wN said:
5 bucks says it won't work.
Quickest $5 i'll make

Why do you think i'm posting here and not ripping a fridge apart? I'm sitting down learning the physics of it, Reading the p/t charts for r407c and talking nightly with people who have around 130years HVAC xp?

Loops busrt some time, mainly during pressure testing of joints - so they burst, all it is a pop in a shed - YOU DONT pressure test with the item near you!

I guess i really posting this to show the level of skills and reading required to pull it off, you cant just go rip a firdge to bits and expect it to work, but you can rip a fridge apart and get a feel for whats invovled.

I know all the specs of my compressor i have braught, 1/4hp, 7cc displacement and with r404a gas it can handle 139w's at -35c, Single Phase 240v that takes POE and MO oils

I'm not building this thing with duct tape, why do you think i have got all the tools before i even touched the compressor?

Before i even trust myself i will have my design and specs checked over by pros, while i'm designing it i will take the advice of the pros, if i'm in doubt i wont do it and i'll ask for help

Either your trolling or you assume everyone is dumb enough to duct tape a PC system together :mad:
 
z3r0- said:
Theres no such thing as overkill.
Yeah there is. Some CPUs perform worse at sub-zero than they do at ambient temps. Some better, though. For example an A64 3200+ reaches it's peak at 2.6GHz at 30C, but won't go past 2.5GHz at -45C.
 
Anyone else sense the irony in the "Phase Change is overkill" statment on the [H]Forum? Please go away, this thread is about teching and learning, not trolling
 
What you may find interesting if you don't already know about his work is the creations of Steve Foster. He was a sort of pioneer to this end. Sadly he seems to have dropped of the face of the earth in the past few years. I managed to restore some of his old website content and have re-posted them on my domain. He did some pretty impressive phase change vapor cycle cooling work back in the day on P-II systems that may inspire you...]

check it out here:
http://www.unofficial3d.com/atx/stevefoster.html

BTW, if anyone know where I can find Steve Foster or knows what happened to him please let me know!
 
Some of those system, wowo, if ever there was art born from computers that would be it!
 
Who's chip? Contact may have been bad. And since the thread is on A64 DEATHS it might have been dying from VDIMM volts.
I'm still waiting for a
"Yeah there is. Some CPUs perform worse at sub-zero than they do at ambient temps. Some better, though. For example an A64 3200+ reaches it's peak at 2.6GHz at 30C, but won't go past 2.5GHz at -45C."
With perfect evap contact and a good chip thats not possibly dying from voltage. Even so thats ONE chip. Show me dozens that do better on air or water than phasechange.
 
I've never had a chip that didn't like phase change cooling the best (with respect to overclocking). The prommy or vapo usually got me a couple hundred mhz more at practically same vcore.
 
Yeah, its only 1 chip!

isnt it a law of physics that the cooler a semi-conductor is the less resistance it has??
 
Lots of people on [H] think phase-change is too extreme or a waste. I hope that attitude will change for the better. In the past I had to go to other forums for phase change info. With this forum maybe the [H] can be a one stop shop for me!
One of these days I'll have to post pics of my pos getto homemade phase changer. Works good, but it's ugly. :D
 
Hey Beccara, you keep going dude, your bound to learn a LOT of things with a project like this. I give you props for taking it on.
Sure you could buy a system for less , but hey ANYONE can buy a system.

Goodluck, work safe and keep us posted.
;)
 
No; those were the actual temps; so -45 is still a lot lowe than +30... A lot of them on that forum are using phase changers (such as prommy mach-II..). Didn't help the overclock any, and didn't stop the cpu from dying either :eek: lol.
 
Jonsey said:
Lots of people on [H] think phase-change is too extreme or a waste. I hope that attitude will change for the better. In the past I had to go to other forums for phase change info. With this forum maybe the [H] can be a one stop shop for me!
One of these days I'll have to post pics of my pos getto homemade phase changer. Works good, but it's ugly. :D
There's no fucking way i'm paying $900 for a phase changer, no matter how good it is.
 
M4d-K10wN, you are taking one forum thread out of context. They are not sure what is killing the chips, but no one is saying it's the temps.

In fact, phase changing is a very cost effective way to get speed. For example, my 3.0 C would only do 3.5 on air - nothing special. Under a phase change system I made for under 300 bucks, it will do 4 GHz. I think that's pretty good bang for the buck. People spend more money on fancy motherboards and very fast RAM but think that phase change is too expensive.

As for reliablity, I've been running this system every night for six months. No problems yet.
 
M4d-K10wN said:
There's no fucking way i'm paying $900 for a phase changer, no matter how good it is.

If you had the tools one could be made for around $300-$350 USD
 
Jonsey said:
M4d-K10wN, you are taking one forum thread out of context. They are not sure what is killing the chips, but no one is saying it's the temps.

In fact, phase changing is a very cost effective way to get speed. For example, my 3.0 C would only do 3.5 on air - nothing special. Under a phase change system I made for under 300 bucks, it will do 4 GHz. I think that's pretty good bang for the buck. People spend more money on fancy motherboards and very fast RAM but think that phase change is too expensive.

As for reliablity, I've been running this system every night for six months. No problems yet.
No, that's not what I meant. For most of them, phase change didn't quite stop the cpu from dying; just somewhat delayed it. For the other guy, he got a better overclock with temps above zero.
 
M4d-K10wN if you wish to discuss the virtues and/or evils of phase-change cooling in respect to overclocking, you're more than welcome to do it in another thread, or through PMs.

Please leave this thread to its original purpose, namely the (to some) fascinating description of a single person's attempt to design and build a phase-change cooling system from scratch.

Thank you.


Beccara some might call you crazy, but isn't insanity one possible cause of progress? ;)

Good luck with your project!
 
Electical resistance in FETs tends to increase slightly the lower you go, until you reach around -178'C, at which point it lifts it's skirts and starts superconducting exponentially the lower you go. The guys at xtremesystems.org are shooting for -180'C...

The point is that sometimes semiconductors will stop working if they're too cold...but then they simply arent cold enough! :)
 
M4d-K10wN said:
Yeah there is. Some CPUs perform worse at sub-zero than they do at ambient temps. Some better, though. For example an A64 3200+ reaches it's peak at 2.6GHz at 30C, but won't go past 2.5GHz at -45C.

Please post a reputable link which proves this beyond a reasonable doubt. Your own experience doesn't count. Please feel free to explain *why* this phenomenon occurs (if it does) - If you can't, which I'll bet money is the case, then please leave the thread. You've already shit on it enough.
 
I think at this point he is simply joking around trying to get flamed like a smore! :D

I think the phase change is the next step in the evolution of cooling from water chilling. I am currently working on my first home made chiller after having several aquarium chillers. My next chiller is going to be made using a 5,000 btu AC system. I found several available for cheap on E-bay using the local search function. (that way I can pick the unit up and avoid shipping) Doing this allows me to get me hands on a AC unit for like 40 bucks. Then I drop the coil in some water, build a case and blam, ice cold and oh so frosty. So far with aquarium chillers I have only gotten to about 10 C. This is due to the unit being only 1500 btu's. Perhaps after this I will begin looking into phase change. Check out the forum at PRO Cooling. They have all sorts of good information and much less hate when it comes to this sort of topic. :p I even found a link to a site selling pre-built direct die blocks for your CPU. Of course they were pretty expensive.
 
PhyberOptik said:
Electical resistance in FETs tends to increase slightly the lower you go, until you reach around -178'C, at which point it lifts it's skirts and starts superconducting exponentially the lower you go. The guys at xtremesystems.org are shooting for -180'C...

The point is that sometimes semiconductors will stop working if they're too cold...but then they simply arent cold enough! :)

Sorry, no. Silicon freezes out around -150C. The closer to 0K, the fewer electrons in the Conduction Band. The Valence Band at -180C, which is 93K, is nearly full. Electrons are stuck in the Valence Band because they do not have enough energy to jump the gap to the Conduction Band. Si at low temperatures is an INSULATOR. (No formula for this one, it's way the hell too complicated for a forum post) Your comment about the "resistance" is only correct below freeze-out. Resistivity = 1/(q[µnN+µpP]). {µp and µn are mobility, and N and P are carrier concentration} Mobility increases the lower the temperature, but the carrier concentrations don't drop until freeze-out so the resistance DROPS until you hit freeze-out.

If you need a good reference, I suggest Pierret, Semiconductor Device Fundamentals, I believe is the title. Streetman's book Solid State Electronic Devices is also decent.
 
M4d-K10wN said:
If you turn your cpu superconductive, it will never work.
Agreed. The whole point of a semiconductor is that it isn't superconductive.

Cooling has limits, where chip architecture does not.
 
Back
Top