sacha35 came over today to start our joint efforts to improve performance
of one of these units using water cooled condensing
http://www.benchtec.co.uk/1/watercoo...h/DSC00865.jpg
http://www.benchtec.co.uk/1/watercoo...h/DSC00867.jpg
Printable View
sacha35 came over today to start our joint efforts to improve performance
of one of these units using water cooled condensing
http://www.benchtec.co.uk/1/watercoo...h/DSC00865.jpg
http://www.benchtec.co.uk/1/watercoo...h/DSC00867.jpg
the mod done we just had time to vac and gas the unit today
with 402a
tomorrow we will start to tune the unit under load on a 6800
parts:
5 foot of 1/2 with 6 foot of 1/4 coiled HX
high pressure read point
1m suction return
0.031 cap
Little devil chilly type evap
danfoss high pressure cut off KP 5 (not fitted yet) used to shut unit down if water cooling pump fails
R402a
mounting brackets still to be made to support the HX
and insulate the pipe work
http://www.benchtec.co.uk/1/watercoo...h/DSC00882.jpg
http://www.benchtec.co.uk/1/watercoo...h/DSC00884.jpg
water cooling being used in these tests is the same as i used in cascade 5 and 4
water barrel and pond pump
this would not be the final solution for the water cooling but will work well for the these
experiments
more on that here
http://www.xtremesystems.org/forums/...d.php?t=152749
http://www.benchtec.co.uk/1/cascade5...g/DSC00781.jpg
http://www.benchtec.co.uk/1/cascade5...g/DSC00783.jpg
Quote:
Originally Posted by SoddemFX
my R23 bottle has a screw on top that was just right for the job:D
Really nice job ;) What lenght of captube did you use on that unit? And mybe you will get even better results with condensor + watercooling ;)
Regards
i have 2.8m for nowQuote:
Originally Posted by LittleDevil
i will see how it works under load then maybe longer maybe shorter:D
you would not believe how quiet this unit is without the fans
No need to use fan on the condensor, only try to put it before ''watercooling'' and it will work like pasive cooler ;)Quote:
you would not believe how quiet this unit is without the fans
Regards
A D-Tek DB-1 would be great for this situation as they're tiny! Plus cheap
Quote:
Originally Posted by Freddie123
i would need to take advice from guys like you on the water side as i have very little knowlage of whats available
but i have seen some rads in the water cooling forum that look like they could do very well
Nice one baz you got it posted up, i must admit baz done most of the work and i just observed, but i was truly amassed at how well this works, i cant say for loaded pressures but the unloaded pressure is far better than my twin condenser run unit using the same compressor.
I was totally blown away by the amount or should i say lake of heat generated by the compressor and feed line, this seams to have a big impact on not just cooling the compressed refrigerant, but also the cooling tracks back to the compressor keeping this cool as well.
As Freddie123 said the D-Tek DB-1 would be a good pump to use together with a Black Ice GT Stealth 360 and 3 Yate Loons running at 7v, it would be very quite unit.
will look forward to seeing loaded results mate and what a pleasure it was working with you today.
sacha
Looks like a lot of fun guys :up: Not much of a Mach anymore. :D
hehe maybe when it gets its clothes back again it will look like a mach:DQuote:
Originally Posted by runmc
Very interesting. Cant wait for the results
Sounds fun :) Time to use geothermal source maybe shortly? :O
Wow Baz, you never stop trying to push the limits do you.
I was going to PM you latter asking what if you mixed WC with a DSHC.
I guess a Thermochill PA120.1 be idea as there low CFM and high heat disapation, the D-tek pump is great, cheap, small and Tom at Chilled PC has them instock.
Infact with the PA120.1 you might even be able to run passive.
I got an old BIX out of an Astek kit here if you wanted to bottor it to test with, G1/4' threads for what ever fittings you wanted/needed.
I'm really really intrested in this
Or the Thermochill PA160 :)
Or just a PA.360 on the side of the case :D
PA160 on the base with barb going vericly up, where condser would go, undet the coil now that be a good location as your fan could even blow over the coil too so water + airQuote:
Originally Posted by Nosfer@tu
idea?
make a chiller, one to the condensor(watercooled/chilled) and other to the rest of the setup. Would be expensive, but could get near cascade temps(not quite) and also be able to chill other stuff like gfx and nb...
And for rad, best bang for buck by far is the swiftech mcr-320...
WOW this is looking great.
Thermochill Rads are the best Rads. Can use lower CFM fans "Quieter" so if possible fit a PA160 Rad at rear and have exhaust fan pulling through the rad. This saves adding a extra fan to the rad so just use the exhaust fan from the case to do the job "if can be fitted there that is".
So Baz is the idea to have this connect to the Loop of a existing water setup Or have a reservoir also so this is a loop of its own ?
Again great work, no hanging around with you 2,. Only see that unit yesterday morning and now its striped and modded :clap:
4Qman
PA120.1 is betterer as the PA160 needs the shorud to fit the 120mm fan!
Thermochill PA120.1, 1/2" T-Piece , Danger Den Fillport 1/2" barb , Masterkleer 7/16" ID (5/8" OD) PVC Tubing, D-Tek dB-1 Compact 12v Pump
Use the DD fill port and T-line as a res, simple, cheap, small and very very effective. The fill port can be fitted in the top of the case even and drop down to the base.
PA120.1 could go at the front or the rear. I'd of said front so the NL11F has some active cooling going over it.
http://www.chilledpc.co.uk/tom/toms/...%20008.jpg
Only decent photo i can find of a T-line setup for you guys.
Hay guys what about just having the rad/pump/res just in your PC case and then just linking the water cooled condenser up with some fast fix fixings that could be put in the back of the mach unit, do you think this would be a good idea ?
Your correct i remember now :p: . I also agree on the T-Line as this will be cheaper and easier to fill.Quote:
Originally Posted by Pete
So it will be a Loop in the Phase unit itself :yepp:
Will be watching this with excitement :shocked:
Mate the only concern would be the heat dump from the condensor :confused: Id imagine :shrug: it will be enough to heat the rest of the loop.Quote:
Originally Posted by sacha35
Great till you have someone who watercools their GPU and NB. Would get a bit cramped.Quote:
Originally Posted by sacha35
I think one idea would be to attach a slimline rad to the bottom and raise the Mach up. A 3 fan rad could be used passively with this setup.
Ahy it make sense too. It's what i'm thinking of! I'd use a 240v pump, Eheim 1048 maybe. Only reason is i have 2 spear here.Quote:
Originally Posted by 4Qman
Thats exactly what i was thinking. I just got a PA120.3 and my SS case is 390x210x230 so trying to work out where i can mount it. Top is a no go as PC case will be sat on top so base makes sense.Quote:
Originally Posted by Freddie123
Has anyone ever made a cascade with a chiller on the first stage? I'd imagine it would work wonders.
Can't really beat the direct heat exchange I think in cascade. Why would we want to put 2 heat exchange I wonder. Water cooled condenser in first stage might not be bad idea but since condensers for cascade tend to big, not sure if there will be any gain...
Quote:
Originally Posted by jinu117
I can say for sure that there are gains to be had, take a look at Baz's work log:http://www.xtremesystems.org/forums/...d.php?t=152749
That's looking well smart guys. :up:
I love your experiments bazx.
Uhmmm that's not water chilling... -_-; I just looked at that build and it is PRETTY :)Quote:
Originally Posted by sacha35
I have a question, havent read the whole thread so it might have been answered but it doesnt seem like it has. This doesn't seem very efficient in terms of noise reduction. All you are doing is transferring the heat from the condenser to a radiator. You still have all the heat needing to be removed and therefore still require the same fan setup on the radiator as you would on the condenser. Now with such a massive rain barrel as a res of course you wouldnt need any fan as it would take days for that thing to heat up and become inefficient, which probably wouldnt even happen as it would get passivly cooled outdoors anyway. So practially speaking, this is no better then an air cooled condenser correct? practically speaking that is, as in no giant res.
Hi sneil you have missed what i am try to achieve hereQuote:
Originally Posted by Sneil
This experiment is not about "noise reduction" its about increasing the performance of the mach unit
The noise from this unit does not matter to me just the load temp, pressures
I am only after performance
These are the goals of this mod :
Lower pressures
Much better condensing
Consistent pressures (not affected too much by increase in ambient (Hot Day)or load)
Better refrigerant
Improved load temp
sound output from the units water cooling will be for others to work out:)
well so far things seem to be working ok
first test set
cap tube 0.031 @ 2.80m
6800@1.7v
temps in window @ 4500mhz -51/52
high pressure rock solid @ 13bar
http://www.benchtec.co.uk/1/watercoo...d/DSC00893.jpg
larger image here
http://www.benchtec.co.uk/1/watercoo...h/DSC00905.JPG
http://www.benchtec.co.uk/1/watercoo...d/DSC00905.jpg
http://www.benchtec.co.uk/1/watercoo...d/DSC00903.jpg
larger image here
http://www.benchtec.co.uk/1/watercoo.../mach/mach.JPG
http://www.benchtec.co.uk/1/watercoo...bench/mach.gif
1m @ 4600mhz
http://www.benchtec.co.uk/1/watercoo...d/bench/1m.gif
cpuz @ 4700mhz
http://www.benchtec.co.uk/1/watercoo...ench/cpuz2.gif
some small movies of the tests with the 32m
http://www.benchtec.co.uk/1/watercoo...h/load/movies/
all images taken
http://www.benchtec.co.uk/1/watercooledmach/
i will have a Q6600 for testing soon
but for now want to try and improve the temps some before i move on to the quad
the 2.8m of 0.031 was just a jump off point for the tunning but seems to worked out well
i will get some K probs on the hx tomorrow so we can look at the temps there
as well
the high side pressure did not move up or down thro out the tests it just sat there at 13 bar
the 32m loaded temps stayed at -46 with the probe T1 taped to the suction return about 6mm from the evap
the other reading is from a prob taped to the out side of the evap T2
WoW baz it has worked out better than i thought.
well done again mate, i think as you said you could get some better temps out of this unit with larger loads but altering the cap tube.
This is just great though mate this unit is so quite it is quiter than the best water setup all you can here is the slight hum of the little NL11, quite and very powerfull looks like the S/S has made a come back for quad core benching.
Sacha
Nice results Baz, would quite like to see how it does with standard watercooling though, not a barrel :D If I had a halfway decent CPU/Mobo i'd offer to test it with my WC setup.
P.S. What're the dimensions of a Mach case?
Quote:
Originally Posted by Freddie123
hehe yes the barrel needs to go
once i have done the phase side of this unit i will be returning the mach
to sacha35
and here will be taking over the water cooling side of this experiment
and finishing
From the looks of it you could attach a rad flush to the top with 2 120's on the inside pulling air through it.
Keep us updated once you get it sacha
Will do mate the idea will be to have some fast fix water conections on the back of the unit so that this can be hocked upto any water setup, so you can just bolt on the phase unit as normal and hock it upto the water setup you have.Quote:
Originally Posted by Freddie123
So look at it this way you all ready have a watercolling setup for your CPU but want more, so simply disconect the CPU waterblock and hock upto the back of the phase unit, obviously the better rad you have the better load temps you will get and the quiter it will be, so if you have say a Black Ice GT Stealth 360 and 3 Yate Loons running at 7v this unit would perform very well and be very quite running, which would be nice for most.
Anyway baz will still have a few bits to do but when it comes back i will let you know how i get on.
Look, i will do you a favor. WHy don't you send that to me, ive got a q6600, mcr-320, laing d5, custom res, and a evga 680i almost ready for phase. If you send that to me for free, i guess i could try it out for yuou.... jkjk but i owuld like to see some before/after or comparison in temps, in addition to the temps of that water.
Yes dream on, i will be doing a complet test with diferant water setups and rads, on a Q6600, so when the units finnished i will post what i find.Quote:
Originally Posted by teyber
But i dont think we will be sending it to anyone for free :rofl:
Bazx, do you mybe have load tester?
regards
hi ldQuote:
Originally Posted by LittleDevil
i only ever test load on a cpu
for me this is all that matters
as this is where it will be used:)
Yes i understand that, but you can make one (cartrige heater) 0 to 250W and you can tune your SS or cascade really easl when you know how muh W is OC'ed X6800 or QX.
It is not that big stuff;)
http://ld-phasechange.com/images/load_tester/1.jpg
Regards
jeeze how much$$ on copper is that ld?
if i were you and i made lots of units then i would say
a load tester would be useful to get all units the same
but this will be the only mach i will mod
and as my interests are overclocking
i like to tune on the cpu and see the results of my work as i am working
i enjoy it this way
When your done with the unit baz i will be putting it on my 0 - 500 watt load tester and will post up some results, but as you said you cant beat a cpu for real load mesuerments.
Very interesting!
I would like to know how the HX is coping as well :)Quote:
Originally Posted by bazx
just pm him.Quote:
Originally Posted by teyber
Quote:
Originally Posted by Jack
i will fix um on today
5° idle to load, very impressive bazx :)
Very good performance. I would be interested to see how it would perform using a 120.3 Rad and Pump with some decent fans on it.
There would be a large amount of heat to shift from the tripple rad. Sacha will you be testing with your water setup when you get the unit home?,.
Now thats awesome, I want such a mod once I get the funds ;)
I got some WC kit you guys can borrow if needs to test it all out
PA120.1
PA120.3
BIX 120.1
Eheim 1048 x 2
Laing DDC2
Hydlor pump (asket kit)
Yate Loon med speed
I had the idea of using 2 PA120.1 in a std case. Rads fit on the base of the case okay, this is taking not case is raised off the floor enough to allow fresh airflow in. Anyway radon the base, 1 at the far back and 1 at the far front. Front one can have the coil over the rad so it will have the water flowing through it and also be acitvly cooled via air off the fan on the rad!
Thats basicly a 120.2 setup, fans can run happerly on 7v and 5v and be totaly silent. Get the pump on some aint vibe mounts and your sorted even more! All you'd have is the sound of the compresser.
I think the WC should be in the same box, having it mixing with the PC and outside is going to be too messy.
How ever here is an idea. Gray built me a V2k unit that had a SLHX which i was going to use on my GFX. CPU was -35 at the his testing load and water was -0.8c i think also under load. Now how about it...you make a SLHX add it to the SS so it's all running off the same system and loop and you basicly then have a water chiller and use that to cool the coil down even more. Might be able to get some crazy blends going on then. Then all you'd have noise wise would be the compresser and pump.
I got photos of it on my rig so i'll grab them soon as painkillers kick in!
So manny ideas that could be done here gents!
My thoughts were to build a res around the existing HX (with current fan removed) and then have that piped through a XSPC 3x120mm rad - the resoning being that you might get great heat dissipation by doing something like that.
But as I am not as skilled/brave/knowledgeable as these guy in such things, I just sit in awe at the work that has been done so far - the results gained prove the worth - nice job.
I will be trying out a single 120 rad first to see how well it performs, then i will use a twin and then triple rad and compare the results of these tests and post, the idea is that if we can have something that can just be hocked up to a water loop and removing a CPU water block then there would be know expense to the consumer as he would already have a water setup, but he would then have a water setup with the added benefit of a silent S/S unit providing sub cooling temperatures.Quote:
Originally Posted by 4Qman
Anyway will keep you guys posted, but baz has some more work to do with the unit first so it will be nice to see what he comes up with.
would you see any increase in performance using a plate HX instead of the coil? i have been waiting for someone to try using water for aaages! the next step would surely be to use passive liquid to air heat exchanger on the condenser side for a lovely quiet setup!
Ya but for the rest of us on load testers bazx, we have no idea the wattage that cpu is putting out, or dissipating into your evap, so your temps look great no doubt, but we can't really compare.
Primary advantage of heat exhchanging that way would be the fact water can transmit a lot more heat given surface area vs refrigerant. (and we only have so much surface area inside the condenser tubing)
Worth the added cost? Not sure. However, it does give some option on how to cool water part obviously (like what was done here, maybe even geothermal for cool 15c - or warm 15c, or typical water cooling route).
Quote:
Originally Posted by n00b 0f l337
adam you no my feelings on dummy loads by now :)
as a bencher i am only intrested in cpu overclock
dummy heat loads mean nothing to me
i only want to know how fast i can make the cpu work
Out of interest can you tell me the heat load that would be needed to be applied to this unit
To simulate
6800 over clocked to 4500mhz @ 1,7v Also at 4600mhz and then at 4700mhz
Running a bench mark prog
and at what point the unit would fail to hold load
I'm not sure which is why I wish you had a load tester :P
jk
Dun worry about it maybe someone will chime in.
225W at a guess Baz
Thats round 150-160 watts on my load tester. What you mean point fail to hold? If it fails its no good :shrug:Quote:
Originally Posted by bazx
Ontopic: More than 4-5 degrees subcooling will not give you better temps. :)
Edit: if you mean 1,7 vcore on bios, it will be more like 130-140 watts due to droop
BAZX - Can I substitute that water thingee with this?
http://www.thermaltakeusa.com/produc...y+kandalf1.gif
That would work in place of his waterbarrel. Difference being his semi passive cooling vs. your active cooling
Well I don't know... How many Watts can that Thermaltake unit handle? It doesn't look that powerfull.
Just took a look at that TT radiator, it's positively huge and I imagine it would work. I'm really interested in seeing exactly how many 120mm x 120mm rads it takes to keep the temps under control.
nice work there bazx, some cold numbers there with the water cooling.
not so good 32m spi run though?
hehe i was waiting for that:DQuote:
Originally Posted by kayl
this is just a load test not benching:p:
Quote:
Originally Posted by X_Man
looks good
This could be good in a Lian-Li V2k, PA120.2 or 120.3 mounted the roof!
Ssilencer managed to put a condensor as well as a radiator in a V2k case, so you should be able to put a decently sized radiator in the bottom of that case as well.Quote:
Originally Posted by Pete
Try 2m50 captube, you'll get better temp once loaded (according to pressure drop it involve with that compressor)
few more tests last night before i shorten the cap tube
32m @ 4600mhz with 1.7v on the same chip 6800
Time unit was running 1 .5 hours
Water barrel temp before 16.1c (taken @ the top of the barrel)
Water barrel at the end 19.5c (taken @ the top of the barrel)
Barrel size 120l
Ambient in shed 15c
Ambient outside shed 12c
the tests were run around 8pm in the evening so the ambient temps had dropped quicker the the water temps
as you can see the unit is running under load in a vacuum
this however is where the best loaded temps were to be had -47
when i increased the charge so the unit was running @ 0 vacuum the temp went up by 4c so -43 under the same load
the high pressure stayed @ 13 bar
http://www.benchtec.co.uk/1/watercoo...2/DSC00919.jpg
loaded hx temps 32m@4600mhz/1.7v
HX temps peaked at T1 36c measured at the start of the hx and T2 21c as the water left the hx
at the start of the run as image
http://www.benchtec.co.uk/1/watercoo...2/DSC00932.jpg
The water going into the hx is hotter than the water coming out of it? After it condensed the vapour refrigerant? That doesnt make sense. Unless I am missing something.
Quote:
Originally Posted by TopherTony
the probes are fixed to the copper pipe carrying the refrigerant
the temp on the 1/4 copper pipe with the refrigerant in entering the HX is 36c
the temp of the 1/4 copper pipe with the refrigerant in leaving the hx is 21c
i am not measuring the water temp here
Great work mate, -47 @ 4.6gig with 1.7v going though that chip, wish i new just how many watts that converts to, do any of you guys have any idea of how many watts this chip is putting out at this speed and load?
Again good job Baz.:clap:
Quote:
Originally Posted by SoddemFX
Thanks for the info tom, but thats running a 32m would that not be using both cores, or would it just be the one?
15c reduction! Pretty dam impressive. Did you have a before and after with the original conderser in there Baz?Quote:
Originally Posted by bazx
So when you going to try a Plate HX?
Why should a Plate HX improve performance?
Do you think the refrigerant will be much colder than 21c with a plate hx? The condensor is good enough! What temp. has your water? (around 18°C i think so he is very close to the max.)
I'm thinking when Baz takes it off his water butt and goes to a WC setup like most of us will have about the place and there being more surface area there and to also experiment all in the name of fun and finding out
like moc has said pete the coil hx has work ok so far
the water cooling experiment has work in that it reduced the high side
pressures in the unit at this load by quit a bit
which means to me that i can put in a colder refrigerant now
i think this is where the real gains will come from
What ya thinking gas wise now then? I love it how you never stop pushing the extreams and limts
Quote:
Originally Posted by Pete
hehe that all i enjoy
i would not be any good at 24/7 units
Never say that dude. I think if we locked you and the Mole in a room and gave you all the kit you wanted a very very scary unit could be made!Quote:
Originally Posted by bazx
I just got hold of a cheap Vapo LS and tempted to mod it and WC like your self and see what it does
Quote:
Originally Posted by Pete
its worth remembering that this unit is an experiment so ya might want to wait till we have all the data in to see if its worth the effort:)
Tom your correct dude. SuperPI is Single Thread only.
I only bought it for the case to be frank with ya. It was uber cheap, something to toy with and then bin maybeQuote:
Originally Posted by bazx
unit is ready for sacha35 to work some magic with the water cooling
http://www.benchtec.co.uk/1/watercoo...b/DSC00968.jpg
http://www.benchtec.co.uk/1/watercoo...b/DSC00972.jpg
http://www.benchtec.co.uk/1/watercoo...b/DSC00973.jpg
http://www.benchtec.co.uk/1/watercoo...b/DSC00974.jpg
Nice work my friend :up:
Regards
Very clean, I like it. I hope it turns out to be efficient.
Nice Bazx but erm...has sacha35 gotta fit a suction line etc etc?
Quote:
Originally Posted by Pete
hehe yes
i use that one i had on there for testing stuff
sacha35 will put on his own maybe evan the mach evap he will also be trying it with some 410a
plus the rad cooling
so still much to come from this unit
I cant help but notice that you are using 90 degree elbows on the water inlets to the dsh/hx. Anybody who has seriously water cooled knows that this is a :shakes: in a loop. It takes away from the efficiency of the loop and thus some performance. I believe that you should switch up the vapor and water inlet locations so the water can flow straight though.
you clearly know a great deal about this sort of cooling thanksQuote:
Originally Posted by TopherTony