Please explain "Phase Change"

  • Thread starter Deleted member 134350
  • Start date
D

Deleted member 134350

Guest
What refrigerant is used in most phase change (evap/condense) systems? Are they actually running an evaporator on the block or are they using it as a chiller to circulate chilled water? And if they are doing an evaporator on the block is it a flooded evaporator or a metered flow evaporator?
 
lol this guy can't catch a break.. even when he moves it to the right forum... lol

check out xtreme systems dude...

mal
 
Quick run down for you.
Phase change falls into 3 group really for CPU cooling at the moment
1) Waterchiller - cools water using refrigeration loop and cold water cools computer components.
2) Direct Die - could be single stage, cascade, autocascade. Evaporator (cold element) in this case contacts what you are cooling directly (cpu most of time). Much colder temperature can be expected than waterchiller.
3) Dry Ice/LN - basically you have container attaching to cpu like direct die but pour these into the container and natural evaporation will cool the container and cpu. Can be cheaper and colder than #2 solution (well within a reason) for quick benchings. Not rally suitable for long term usage nor cheap when you keep buying more containers of dry ice or liquid nitrogen.
basically you will get cpu temperature well below 0c most of time doing these cooling solution which helps with top end overclocking. pelter do fall a little short and more of in between watercooling and waterchiller though. And I have to say, it isn't cost efficient at all.
 
Thanks for the reply, I am a heating and air conditioning tech and wanted to verify that they are talking about an evaporative block. I also wanted some specifics on the way the refrigerant is metered into the block and what metering device is used, also where are they sensing the suction line temperature to determine the amount of refrigerant to feed to the evaporative block and what refrigerants are used? I have worked on cascade systems where the air temperature in the box reaches -100d F but have never applied it to computers.

What you call "Phase Change" we call refrigeration. Latent heat of vaporization and latent heat on condensation are the vehicles of heat transfer. ;)

I found the information that I was looking for. Most just use a fixed lenght cap tube which is fine for constant load situations. I believe that I would like to try it with R408A and see what is possible. The real low temp cascade refrigerants are very expensive. Now to start thinking about a good block design. :D
 
SiGfever said:
Thanks for the reply, I am a heating and air conditioning tech and wanted to verify that they are talking about an evaporative block. I also wanted some specifics on the way the refrigerant is metered into the block and what metering device is used, also where are they sensing the suction line temperature to determine the amount of refrigerant to feed to the evaporative block and what refrigerants are used? I have worked on cascade systems where the air temperature in the box reaches -100d F but have never applied it to computers.

What you call "Phase Change" we call refrigeration. Latent heat of vaporization and latent heat on condensation are the vehicles of heat transfer. ;)

I found the information that I was looking for. Most just use a fixed lenght cap tube which is fine for constant load situations. I believe that I would like to try it with R408A and see what is possible. The real low temp cascade refrigerants are very expensive. Now to start thinking about a good block design. :D
Click to expand...

Since you are not noob per say on this :p
What you are looking into is cooling 150-170w heatload. (disregard those cpu heatload calculation people mock up... i've tested and can verify :p)
R408a sounds like decent choice (I usually use R404a, R507, or R402a) considering PT chart.
What I find with R404a (well closest to R408a out of all gaese i use :p) is that cap tubing around 9ft 2 inches works great for 150w heatload (can handle 180w too but best temp around 150w). Look for compressor from 1/5 to 1/3 hp. Evap need to be designed in mind with surface area, mass, turbulance if you can put some in (no evap design currently has good turbulance model thanks to complexity of machining). There are commercial evap available as well.
Yes cascade gas can be expensive (still cringing at r23, r116, r508 i have here...)
A cheating way could be using ethylene which is dirt cheap compared to most refrigerant (well short of r-123 i guess)... only catch is, it is flammable and explosive under pressure. Even lower temperature can be expected than r508 properly tuned (need bit of r22 or something to carry oil though).
Feel free to pm me with more specific questions (doubt you will have much). Oh and you usually will run this things under mild vacuum vs what is used on field thanks to smaller heatload we deal with... we take advantage of it as better temperature. As long as you keep superheat under control, condition it is used makes it possible for our uses on this at least.

PS) 0.031 cap tubing bleh. I knew I was missing something.
 
Man, I was gonna check out this thread to see if I could throw in any basic info...but it looks like you're already miles ahead of me jinu...I'm actually gonna give this a read when I had time...cus I do want to get very "into" phase change myself.
 
I myself wanted to build a system. I reviewed the concepts and such for a while but never really got into where I would get parts and such. Is R134 good refrigerant to use? And is it stronger than R404A?
 
Jinu117,

Thanks for the great reply. Some people just seem to not want to share the info? :(

There is so much to adapting my trade to computer cooling meaning the evap block and the possible ways to keep the refrigerant changing state as long as possible to be able to extract the most heat. As you know, it don't do crap being pumped around the system. If you can not keep "Latent Heat of Vaporization" going as long as possible you will not get the maximum benefit.

Have you tried using a small accumulator to be able to handle lower superheat settings?
 
You guys both seem very "on the money" and I'm glad this thread came around. I dont know if I'll ever be to the point of being able to build and diagnose phase change systems...but I sure as hell want to be knowledable enough to use them properly and identify common terminology.

Thx jinu again for all the good information. I think I'm gonna go check out some of the FAQ and instructional threads over at XS again to give my urge for more ever growing.
 
In our case, what you don't really want is to keep the latent heat of vaporization as long as possible :)
You want it to happen in very rapid way. More violent the better.
You see, evap in our case has limtied total size it can be due to mounting and insulating consideration. So we have to work with small evap that is transmitting quite decent amount of heat for contact surface area. What becomes really important now is... consistent good flow rate at high speed so heat is "sucked" in essence, without boiled refrigerant blocking new fresh liquid to come in. Another thing is avoiding oil trapping. What researchers have found is... jet or impringement on these refrigerant has huge impact on increasing usable surface area for this to happen rapidly. These will help reducing overflooding of evap while increasing the rate of evaporation meaning more capacity or colder temperature.
However, I still to this date haven't figured out jet type evap design that can be machined without extra ordinary cost... even some minor turbulance model would be nice... but not easy to machine by cnc again. Got any idea?
 
jinu117 said:
In our case, what you don't really want is to keep the latent heat of vaporization as long as possible :)
You want it to happen in very rapid way. More violent the better.
You see, evap in our case has limtied total size it can be due to mounting and insulating consideration. So we have to work with small evap that is transmitting quite decent amount of heat for contact surface area. What becomes really important now is... consistent good flow rate at high speed so heat is "sucked" in essence, without boiled refrigerant blocking new fresh liquid to come in. Another thing is avoiding oil trapping. What researchers have found is... jet or impringement on these refrigerant has huge impact on increasing usable surface area for this to happen rapidly. These will help reducing overflooding of evap while increasing the rate of evaporation meaning more capacity or colder temperature.
However, I still to this date haven't figured out jet type evap design that can be machined without extra ordinary cost... even some minor turbulance model would be nice... but not easy to machine by cnc again. Got any idea?
Click to expand...

Wow...I'm just going to assume that you're not talking to me on this one...because while I did understand every word of what you just said...that's WAY over my head. :p
 
jinu117 said:
In our case, what you don't really want is to keep the latent heat of vaporization as long as possible :)
You want it to happen in very rapid way. More violent the better.
You see, evap in our case has limtied total size it can be due to mounting and insulating consideration. So we have to work with small evap that is transmitting quite decent amount of heat for contact surface area. What becomes really important now is... consistent good flow rate at high speed so heat is "sucked" in essence, without boiled refrigerant blocking new fresh liquid to come in. Another thing is avoiding oil trapping. What researchers have found is... jet or impringement on these refrigerant has huge impact on increasing usable surface area for this to happen rapidly. These will help reducing overflooding of evap while increasing the rate of evaporation meaning more capacity or colder temperature.
However, I still to this date haven't figured out jet type evap design that can be machined without extra ordinary cost... even some minor turbulance model would be nice... but not easy to machine by cnc again. Got any idea?
Click to expand...

Im fairly new to phase and ive only seen the basics, and a few cascades, but, maybe you could use simple angles in the design to create turbulence? Just a suggestion, trying to help :p
 
Emission said:
Im fairly new to phase and ive only seen the basics, and a few cascades, but, maybe you could use simple angles in the design to create turbulence? Just a suggestion, trying to help :p
Click to expand...

You don't want turbulance outside of evap (well bit maybe on suction but our flexible suction line tend to take care of that)...
The evap is where turbulance is needed to aid in removal of heat as well as preventing pooling of refrigerant. Not an easy thing to do.
 
jinu117 said:
You don't want turbulance outside of evap (well bit maybe on suction but our flexible suction line tend to take care of that)...
The evap is where turbulance is needed to aid in removal of heat as well as preventing pooling of refrigerant. Not an easy thing to do.
Click to expand...


Why not? I'm unfamiliar with the internals of typical evaporators for CPU cooling, do you have any pics of what you're referring to?
 
jinu117 said:
You don't want turbulance outside of evap (well bit maybe on suction but our flexible suction line tend to take care of that)...
The evap is where turbulance is needed to aid in removal of heat as well as preventing pooling of refrigerant. Not an easy thing to do.
Click to expand...

I meant using angles within the evap.
 
Would be nice... but how would you machine it is ultimate question. I guess one can get file out of hand and start biting on evap, drill bit... even router.
Bottom line always comes up though... is it feasable to do it in reasonable price? Probably not. By doing it in hand on most current evap design I would think 3-4 hours of work if you are lucky. This might work well with maybe few units but once we start talking about more... well... anything manual that takes that long to do do get old quick :p
 
Cant you just machine the inside of the block and seal it up? Dosn't sound hard.
 
Emission said:
Then there shouldnt be any problems with making some angles :D
Click to expand...

I'm honestly not familiar at all with the machining process myself or the dynamics of how things work (jinu might better field this) so I'm not sure if it'd help as you think it will, or if it'd be easily done. Sorry bro, but this is one of my future projects, learned more about the workings of phase change.
 
Its pretty simple using 90-degree turns, looping in a spiral out from the spray point.
 
You must log in or register to reply here.
Back
Top