I wonder how oil will flow in that huge accumulator.
Printable View
I wonder how oil will flow in that huge accumulator.
in such small systems like we do its really not a problemQuote:
Originally Posted by Unknown_road
the velocity in the accumulator will be far to low to push the oil up.
blaster: because the system is so small with such little refrigerant flow it is a problem when you use an accumulator like that.
ok let him build it and we talk laterQuote:
Originally Posted by Unknown_road
Other little idea's coming into play here, the accumulators just bent down at the moment and will be coming up to a horizontal plane before its finished. As for .036 I will be testing that vs 6 feet of .031. If .036 12ft is truly 6ft of .031, it should work better to have the longer subcooled length.
I think its time to be a bit more innovative, too many are stuck in the old same patterns.
And: Velocity on its own doesn't push.Quote:
the velocity in the accumulator will be far to low to push the oil up.
:confused:Quote:
Originally Posted by Blaster
just saying, take this into considiration, if it will go wrong it might be to late for the compressor and it will take quit some time before it will happen.
@nol: you know what I mean
Personally I already was going to have the accumulator mostly horizontal, a slight tip but less than 20 degrees.
the accumulator of a rotary has a very small hole in the bottom were the oil can go in, I recommend making something like that. Even when you have the accumulator horizontal there is lots of room for an oil pool
Yes this is true, as for when I said experimental, I'm considering trying to burrow a capillary in to the bottom and run it to the top.
There is nothing wrong with trying experimental stuff, otherwise we'd get stuck and never end up progressing. Looks like good stuff and if all else fails we will simply prove that this concept won't work for us.
It should though, and I DO know those using .036 for higher loads with small units.
6-7’ of 0.031” / 0.8mm capillary line is about what ya need for the feed on the first stage of a cascade. Good luck though and look forward to your results on holding 300w. Even if it holds 200-230w thats a nice effort. Noob can we have some pics and info on this load tester you will be using. What sort of cold plate do you use between the cpu head and the resistor load?
WHen charging these small systems keep in mind to get -60 with R507/404 we need to operate in the 3 to 5 inch vaccum region this is very low flow unless you have very large displacments, oil return is at issue as well as compressor cooling, stick with hi effeciency compressors at about 1.3 to1.5 W/W Rotaries are fine for bencing runs but 24/7 use you need cooler compressors, for instance an 18,000 btu rotary is running at 2 inch vaccum rotaries are cooled by the refrigerant as is most compressors so motor cooling is not very good due to teh flow rates, this in turn heats up the small amount of refrigerant entering the compressor to 95 to 110C causing the oil to start to break down.. the extra heat also has to be removed from teh refrierant discharge gas causing you system head to rise making the system runhotter and higher suction pressures... so not always is larger better. There is a trade off unless you can get higher effeciency compressors.
@Kayl,
I do not use a resistor load, they are not as accurate as a load cell burrowed with thermal TIM into a copper block!
For a resistor load, Id say take off 10-20watts of heat, and thats a bit closer to what your tuning for. Its also easier to insulate the load tester on a 2" x 2" x 3/4" copper block.
Maybe he means like 1.3-1.5 to 1 as a ratio of heat moved to power consumed? Hard to tell what your saying Chilly.
And thats why I'm wondering what Reg is saying :P
Incase you guys dont have this, here's a nice chart to save!
http://img249.imageshack.us/img249/5...ersions1dg.jpg
Guys my last phase change unit was the original Vapochill, so I guess I need some help with phase cooling quad cores.
I am currently running a Q6700 on a P5B Dlx and a X3220 on a Commando, both at 3.8 GHz with 1.6 Vcore. For cooling the Q6700 has a Swiftech kit with their Apogee GT block, while the X3220 uses a custom kit, also with the Apogee GT block. Both run at 100% load on all cores 24/7/365 with DC apps.
What I plan to do is bench with them at much higher speeds and Vcore.
The question is, what would you recommend for phase change setup?
DDTUNG:cool:
Well DDTung -
It's going to have to be a custom cooler because no commercial coolers on the market are doing well with the quad cores. They just don't handle the load.
That's pretty much the long and short of it Ron.
Something I've been meaning to post for a while now, in relation to the 'we need a better block' statements.
I've used just about everyone's blocks now, from Kayl's to Chilly's to Joe Cool's, and while I haven't used everyone's I'm getting pretty close :rolleyes:
Heat exchange is what makes tuning for higher loads easier, especially where smaller systems are concerned, we already know that. Plate heat exchangers have proven that quite handily on cascades and water chillers already.
Enter Ssilencer/@itor's evap.
http://www.picturehosting.com/images.../photo0157.jpg
Yes I know, bad pic :p:
Once you get past the 'bling factor' of the clear enclosure and LED's (and it is quite pretty :D) you get to see the 'guts'.
I didn't realise but the evap is a 'press fit' and I shouldn't have been able to get it apart easily. It wasn't 'easy' but I like to check things to be sure about them :D
Thin plates, lots of them, thin base. Very tight refrigerant return channels. Sandblasted internals. Very very effective heat exchange.
I had the pleasure of making a V2000 system for Messiah Kahn using this evap, and tuning it for higher loads even in a small system only using a long suction return line was very easy.
The amount of floodback even at full charge/no load was next to nothing. That's with the unit tuned for about 230w. Why? Heat exchange.
The head makes so much more efficient use of the refrigerant passing through it than any other head I've used, that I need less actual refrigerant in the system to make it hold load.
When you see a clear evap 'test bed' and how the refrigerant is working inside it really opens your eyes to refrigerant behavior in a system. I think it was Tyrou or one of his French compatriots that gave me that insight at one point.
It let's you realise that there's more than enough refrigerant passing through an evap to cool 5 cpu's, if we could only make use of it more efficiently.
The @itor evap makes far better use of it from my testing, and for higher loads, ie anything over 200w, I firmly believe it's the best one available at the moment. I'll be watching for others as I go, but it's behavior and the systems response in having it confirms it.
I was amazed when I got to see inside it, by how small, how thin etc and I thought it shouldn't work by everything I've come to know and use in phase change.
I was proven wrong, and was very pleasantly surprised by how effective it was. I'm quite sure that tuning for even higher loads would be possible and quite easy, and I'm looking forward to having another look at it's performance as I have another @itor evap here to test when I can get the time to work with it.
Cheers guys
Gray
Gray, I agree fully.
I see with @itor evaps on my last builds that what you are saying is completely true. I did not think about it this way but it makes sense now after your explanation.
I found on both units that static pressure was much lower than other systems I have built. This obviously means that the systems both have less refrigerant inside them than other typical builds.
I also found that when holding load (230w) pressures was lower than normal which in turn produce better temps. Normally when running this kind of load you have to charge much heavier to make up for evap inefficiency. This is a snowball effect leading to high discharge pressure, high compressor temp and bad temp at lower load. You often have to charge so heavy that lower load temps completely fall of the wagon.
The blue unit ("Monster" in my thread) holds 230w load to -35c while maintaining a idle temp of -64c and 150w temp of -40c.
5c delta from 150w to 230w
This makes you think..
Im not going to post the length of cappilary used in Monster because it will stirr up flaming and it changed it 10 times to get it rigth but I can tell you that its longer than with other systems, not an inch or 4 but ALOT more.
This say something, why does a unit with ALOT longer cappilary tubing hold 230w to a better temp than anything else I have built or most seen around here?
The cappilary used is against everything we have become used to over here at XS and completely opposite of the direction that the cappilary discussion in this thread has been heading too.
This made me realise that what I thought I knew about cappilary tubing lengths up to this build was completely way off, problems lie elsewhere, mostly evap design, once you have to go too short in cappilary you have to realise that the problem lies in the design of the evaporator since you need more refrigerant to remove the heat than needed.
We have been using ALOT shorter cappilary than is needed, even 3m or more should be more than enough to hold loads of 200w+ (If evap heat transfer is good)
Makes sense? :D
I mentioned better evap designs earlier (cant remember if it was this thread), but I think I'm going to want to test an @itor.
I certainly didn't want to say that it 'can't' be done, that is, tuning a system to 250w of load with another evap.
It's just that with the @itor evap design, it's very easy, efficient, effective.
If you can't get the efficiency to happen at the evap, then adding enough heat exchange on the suction line to allow for a heavier charge is the only safe and effective way to do so.
Johann, it's nice to have your experience to show this as well. Pretty amazing though wasn't it? To be able to charge it for such high load and not be fighting the pressures and condensing temps that go with heavy charging. It was great!
I'm also looking forward to seeing Chilly's new evap. He's mentioned some of the revision changes and I'm hoping to see this kind of result on his ones too. Evaps are getting to a point where they're quite good, and the technology behind their manufacture is maturing as well. Hopefully this will trickle down to us poor folk that have to rely on other people making evaps :D
Cheers
Gray
PICSPICSPICS!!!!!!! :DQuote:
Originally Posted by Gray Mole
I think what he's referring to is the 'project aurora' from some guys over at cooling-masters. (French site).Quote:
Originally Posted by s7e9h3n
I've seen the video before of a clear evap and it shows how the refrigerant behaves once inside :).
I had a quick look but couldn't find the video :(.