Got my phase change rig redone, need some help with charge

[Russell_hq]

Ok, here's the deal. I can only measure high side and low side pressure/Saturated temperature and temp 4" from compressor inlet. My temp probe cant measure the compressor outlet cos I think its over 50ºC which is the limit for my probe (-40ºC to 50ºC).

I know that the air dT across the condensor is around 4ºC and room temp is in the region of 18ºC. I know Gary, you will be wanting the whole list but I dont have it right now. So think you can help?

Here is what i have so far:
R290

SST = -45ºC (4"Hg)
SCT = 26ºC (127psi)
Suction Line temp near compressor inlet = -27ºC

CPU temp 1ºC (this is take from socket diode, not the diode in the cpu die)

Processor running conditions = 1.95v @ 2740Mhz equates to:

142Watts and 88% 126Watts from http://www.benchtest.com/calc.html

Think you can help me get things cooler?
one more thing, I have 25mm thick armaflex hose insulation on the suction line, plenty insulation me thinks

[berkut]

Evaporator temps ?

Compressor ?

drop over @ irc.quakenet.org , xtremesystems

[Russell_hq]

I dont use irq sorry, just msn messenger. russell_hq@hotmail.com

Compressor temp is nice and hot but not too hot i cant leave my hand on it for any length of time, compressor is an electrolux ML80FB, R404a low temp 1/4hp. Evap temp I have no idea.

[Russell_hq]

here's some pics;

[berkut]

I love your insulation

i would like to have some evaporator temps :\

[Russell_hq]

Ok, i'll take the probe off the suction line, where abouts on the evap do you want it? Just on the pipe at the evap exit, thats probably the best I can do. The probe head looks like this:

[Russell_hq]

Well all day I have been tweaking vcores, MHz and refrigerant charge to see what I can do. My conclusion is its not cold enough and I want colder. Dont get me wrong, I can get the old barton to 2.8GHz. Prime stable, well maybe stable But I want more speed and I rekon the cooling is holding me back, thus lower temps are called for. Problem is I can't really get refrigerants without spending the £££'s and the best im going to get is R404a or R507. The improvement over propane i'd say is not worth it, maybe 5-10ºC colder. Where I want to go is -80 or colder, and I want to do it with an autocascade

I was going to use propane and CO2, but I just found out the triple point is -56.6ºC, therefore no more liquid at temps lower than that

So now I have to think up what gas to use that I can get my hands on thats not going to cost too much.

[Russell_hq]

Evap temp is around -37ºC

[nas]

under load ?

[Russell_hq]

yeah, full load. Prime 95 running

[MartO]

Russel_Hq how long it`s the captube you are using?

I`ll asemble my system in the next days and I have the same compressor as you, but I`m using R404a.

[Russell_hq]

Im using 0.031" cap tube and 180" of it. Most folk use 10ft of 0.028" (or is that 0.026", I cant remember) for R404a. Just have a look around the forum. I know that Gary is always telling us, but I always forget. I assume that you are going direct die, if its a water chiller then you will probably need less. Hope that helps.

[Gary Lloyd]

It all looks like it is running as well as it can at the moment. In the list of temps you report, include the liquid line temp a few inches from the condenser outlet.

The difference between SST and CPU temp (1C minus -45C = 46C TD) tells us that the heat transfer isn't very good. It seems like I have seen 30-35C TD on these systems. This could be the contact, or it could be the evaporator design.

[Russell_hq]

I dont trust the temperature reported from the sensor on the board. Previously I used the same evap in the same system but with 150" of 0.031" ID cap tube, and I tested it on a board that could read the temperature of the actual cpu die (as they have a diode built into them) and the temps that it reported were much lower, 15-25ºC lower if I remember right. The motherboard I have at the moment uses a thermistor that measures the temperature on the back of the cpu packaging so its not that good an indication. i am confident that I have good contact with the cpu and that my evap design is sufficient.

[Gary Lloyd]

Your cap tube size sounds about right. A cap tube is stable between 5-16 feet, so if you wanted to make it more restrictive, you would need to go to .028 cap tube. 15 feet (180") of .031 is equivalent to about 9 feet of .028 cap tube. With more restrictive cap tube, the evaporator temp drops, but the capacity also drops. The ideal is to have just enough capacity to handle the heat load. None of this is chizeled in granite. The jury is still out on ideal cap tube sizes for these systems.

[]JR[]

Couples on the way russ, first class only had 50p on me nevermind...

]JR[

[Russell_hq]

First class is perfect. Cheers mate.

Gary, if you get a bigger compressor then you could use a longer cap tube and get lower temps, as the compressor could handle the capacity at the lower evaporating temps. Does that sound right?

[Gary Lloyd]

Essentially with a bigger compressor, the high side pressure remains the same, but the low side pressure is less, which increases the refrigerant flow a little. The two most important factors are the heat load and the refrigerant. With the larger compressor, we might be able to go slightly more restrictive on the cap tube, but not much. Compressor size is not a major factor.

Assuming we are going for lowest temp, as opposed to fastest pulldown, cap tube size is about refrigerant flow, not compressor capacity. Conventional cap tube charts are geared to compressor size because they are going for fastest pulldown.

[Russell_hq]

OK, got my Fluke and thermocouples (cheers ]JR[) today. I put one thermocouple underneath the evap, right beside the core so thats the best indication of core temp I think I will get. Right now im running stock speeds, 68Watts.

MBM core temp: -7ºC
Fluke Temp: -36.9ºC
SST: -51ºC (10"Hg R290)

Thats runnung idle.
Running Prime95 for 15 miutes I get the follow:

MBM core temp: 1ºC
Fluke Temp: -30.1ºC
SST: -51ºC (10"Hg R290) not much change dial isnt accurate enough.

I'm going to bump up the speed and v's and see what happens

[Russell_hq]

I forgot about the pics

Evap temp with computer off nice and cold (I know, I know, no load means nothing but its nice to see how cold it can go. -50ºC )


Pressures (running pressure not much different from these)

[Bra!nFreeZe^]

Damn thats nice! I would love to have the tools and such to build such a system!

[Russell_hq]

Ok, im stuck now. My mobo wont let me select multi's higher than 12.5. It used to do it, and I've not really changed anthing apart from putting the thermocouple under the evap. Its an Abit NF7-S v2.0 and it should be able to use all the multipliers, I was using 15 and 16 the other day no problem. Anyone got any ideas? Im using an unlocked Barton 2500+ and I just reflashed the BIOS. Everytime I go from 12.5 to 13 the mobo wont POST, even with low bus speed. Can anyone help?

[Devilman]

Hmm, if I remember correctly, there were some NF2 boards that wouldnt go above 12.5 unless you modded the chip, at which point you could use above 12.5 but never go under.

In theory any NF2 board should be able to access the full range of multi's on a barton anyway, its 99% always a bios issue.

Try "downflashing" to your previous bios version and see if you can access 12.5+, Although Im guessing that will fix it right up.

They probly just broke something in the new bios

[Russell_hq]

The multipliers were working fine, I was using all the way up to 16. Since then I mounted a thermocouple to the processor and I havent been able to use the multipliers above 12.5

The first thing I did to try and fix it was to reflash with the same BIOS, but that didn't help. I then flashed with a previous BIOS version and that didnt help either.

I've seen a site that shows how to unlock a Barton, but my cpu has never been locked so i dont know whether to try this method.

http://www.athlon-unlock.co.uk/5202/5283.html

[Russell_hq]

Gary, if you could, what are the delta T's that you are looking for when you ask as to report all the temperatures. I've been reading the forums and come up with this so far for what to look for and why when tuning.

SCT - Air In Temp: 12-13K ( So we can get subcooled liquid out the condensor)
SCT - Condensor Out: 8K (our subcooling value)
Suction Line near compressor - SST: 8-11K (Superheat, enough so that all droplets evaporate and nice cold gas flows into compressor )

Is this correct? Maybe I have missed something. What I am trying to get across is what we are looking for and why when we are tuning these systems. i think this information would be very useful, especially in a sticky