If it's any consolation, I do have the Bitspower tops mounted "upside down to accomplish this" ... at least as far as the logo is concerned.Quote:
Originally Posted by SNiiPE_DoGG
Printable View
If it's any consolation, I do have the Bitspower tops mounted "upside down to accomplish this" ... at least as far as the logo is concerned.Quote:
Originally Posted by SNiiPE_DoGG
yes see I presumed you had drawn the diagrams incorrectly in the picture earlier on but that is not the case. My apologies ;)
still I say just use two pumps in series for the GPUs and you will have more than enough redundancy as long as you monitor them frequently enough.
If the rig is planned to run 24/7, then you will get no temperature gain whatsoever from adding more water to the loop. The water will (eventually) reach equilibrium just the same as if you had a T-line and no res. If you only run the machine for a few hours out of the day, then yeah, you might see gain from a huge res. But if you have the multiple loops running into and out of the same massive res, it no longer becomes a dual-loop system, just a single loop with a whole bunch of pumps.Quote:
Originally Posted by xaxis
why not run a third pump?? well, u can give it a go, im just not entirely sure u would get that much of a benefit from putting a third in there!
if u have a spare u could go for PUMP>RAD>RAD>RAD>RES>PUMP>PUMP
so that you get the pressure up going into the radiator!
I like that idea... I will shortly have 5 D5s... two as hot spares as I detailed a bit in previous posts (for "unnecessary" redundancy and experimentation's sake)Quote:
Originally Posted by exemay
Hmmm, yes it will reach an equilibrium but with more volume for heat to diffuse thus surface area to escape I question whether that equilibrium would be less so to a non-negligible extent.Quote:
Originally Posted by YMAA
The redundancy is just for fun... doing something I've never done, etc. Plus I've already completed 90% of the controller schematic... just looking for some electrically pneumatic ball valves I can mod into the loop as failure point diversions. :)Quote:
Originally Posted by SNiiPE_DoGG
A closer look at the bitspower top/bottom/front.
http://www.xaxis.net/downloads/image...ation/0020.jpg
http://www.xaxis.net/downloads/image...ation/0021.jpg
I'm not sure what you mean by "volume for heat to diffuse thus surface area to escape." Do you mean heat dissipation through the res? The amount of heat dissipated from a closed acrylic res is insignificant, even a metal one would be the same for that matter. If the reservoir is exposed to the air and large enough it might make a difference, but then you'd have the same contamination problems bong users have.Quote:
Originally Posted by xaxis
Or do you mean heat dissipation from the radiator surface area? If that's the case, then it doesn't matter how large your res is because the air -> water delta will be so low due to the radiators. The fact of the matter is, running the machine for long periods of time will cause the water to reach equilibrium. How long it takes depends on the res, how high or low the temperature depends on the radiators.
Either way I don't recommend planning a gigantic res for the intent of cooling better. T-lines may actually be a better way to go considering the amount of stuff you're trying to fit in the case(s).
i heard you do, because the air inside your res is compressible.Quote:
Originally Posted by exemay
Unless someone want to show me physics on how it doesnt.. :P
U know me, im a total physics person, and physics will shut me up.
I've been saying this since Martin did his test with the Coke bottle. ;)Quote:
Originally Posted by NaeKuh
http://www.xtremesystems.org/forums/...14&postcount=6
From
http://www.xtremesystems.org/forums/...d.php?t=152428
ok....Quote:
Originally Posted by NaeKuh
physics says that im right and your wrong ....happy?:yepp: :rofl::rofl::rofl:
not without a general proof using constants solved out mathmatically. :rofl:Quote:
Originally Posted by exemay
Solution: FURTHER segregate the inlet and outlet to upwards of 75% the height of the res "coke bottle".from the top derlin, insert two 1/4" thick nickel plated copper coils into the res, with stems breaching the derline top yet 100% air tight. attached to the coil apparatus affix a heatsink/fan. Hypothesis: Thermal diffusion will be provided via the coil assembly and the increased segregation in the reservoir will decrease flow from rushing in and out while allowing for near complete air bleeding for the EDIT: air will have time to "settle" and escape. The system flow will stabilize. Additions of sealant tapes and extra measures to ensure air tightness will be a must as well.Quote:
Originally Posted by Waterlogged
EDIT: the heatsink/fan would of course be brazed and affixed outside of the res... basically an external heatsink with "tentacles" coiled into the res.
Sounds fun. Thoughts?
lol if u really want i can investigate it for ya!!Quote:
Originally Posted by NaeKuh
but this is what i understand it to be:
common sense would say that the pressure would be effected overall, because its like pumping into a radiator, you get a pressure drop such as when you pump into a radiator, its a volume of water behing held in a the small confides of the tubing going to an area of much larger volume, becasue the loops are selead PHYSICS tells us that it requires more effort for fill up an area of larger volume under pressure!......but i cant see that it really applies to this, from testing we can see that having a setup such as pump>rad>pump will keep the pressure constant because the pump after the radiator picks up the pressure lost!
going on that, id say that you wouldnt really loose much pressure overall having the 3 pumps/res in this setup such as has been posted on this thread becasue from what i can tell, the final pump will pick up the pressure which has been lost form the first and second pumps, pumping into the second and third res.
From what i can tell, you wont loose any pressure because he hasnt got the pumps in series....he hasnt built up any extra pressure or flow therefore from what i can understand, this setup would have the same flow having 1 pump, or 3!
so basically, the only way to set this loop up effectively is to have 3 pumps in series......OR have the pump dotted about the loop to pick up pressure lost through blocks or radiators OR have a monster sized res and have 1 pump after the res!
exemax u forgot tho, in a fully bled system there is no air, and water isnt compressible.
That is why loop order has no impact because flow going from A to B is uniform...
however if you have a res b4 the pump, the pump will exert pressure on the res, compress the air, and eat your flow and pressure.
Martin tested this and showed it.
As long as you can get 100% of the air out of the res, it "should" act like an over sized piece of tubing. It should also have absolute minimum effect on pressure.
in that case im confuzzledQuote:
Originally Posted by NaeKuh
but the way he is proposing to set up the pumps and res's he wont get any benifit in flow...would he??
...just have the pumps pull water from the res like a normal person...
Just a small fraction but that would be due to the pumps achieving near full rated pressure. A serious gain in flow would be from joining them in parallel into 1 bigger tube (3 x 3/8" I.D. into 1 x 5/8" I.D.).Quote:
Originally Posted by exemay
Look what just dropped into my lap.
http://www.xaxis.net/downloads/image...ation/0022.jpg
http://www.xaxis.net/downloads/image...ation/0023.jpg
Lots of modding fun to be sure. How, what, where, why? Certainly not combined with any of the other loops... But I have big plans for this little unit, plans I may need help with when the time comes, but plans none the less.
SS meets TEC meets LC w/ a little LN2 sprinkled ontop.... Or something like that... :D
Boreas.... :\
Ima keep quiet about that as that is the one of the things i hate the most in TEC's. The other is the freezone.
Xaxis you mix metals, and ima shoot you... then resurrect you only to shoot you again... because your butchering a queen rig. :\
Quote myself: I just wanted to point out that from the start I was fully aware pressure increase would be negligible if any. I was orienting the setup the original way I proposed because of my interest in building a redundant loop with logic controlled electrically pneumatic valves...Quote:
Originally Posted by xaxis
(yes, I agree, see above)Quote:
Originally Posted by exemay
normal person??? Have you lost My mind? :wierd: never! :)Quote:
Originally Posted by YMAA
(Also, agreed).Quote:
Originally Posted by Waterlogged
:) I wouldn't speak such words to early on... this is just a toy and though my suspicions of your telepathy are growing I still strongly doubt you know what I have in mind. ;)Quote:
Originally Posted by NaeKuh
Out of all this insanity one thing I can promise you, is: I WILL NOT MIX METALS, PROMISE!
It uses alu in the wetted parts.
That means your almost screwed in mixing metals unless you use it for ram blocks.
also xaxis...
http://img199.imageshack.us/img199/3849/mistakeh.jpg
found a ton of errors in your original plan.
I thought the same thing - turns out the BP tops work differently than a normal d5 topQuote:
Originally Posted by NaeKuh
Quote:
Originally Posted by NaeKuh
Are you talking about aluminum in the TEC? Read below, I said it won't be part of any of the loops.
As for the D5s, your probably right, but I didn't explain it well but way way down I said I have them turned sideways for viewing purposes... The pumps when mounted on the UN brackets as I have planned I believe will be purposed as they should be... so yeah, this exact loop order etc is not at all what will be seen.. I was planning revisions based on all of yours and everyones helpful suggestions/criticisms below... The only thing I won't do is leave a pump out or leave a rad out or leave a res out, what have you... They will all remain, but how and in what order remains to become a constant...
Again, again.. TEC is just a toy... all shall be revealed.
Adding the Boreas to your loop is absolutely pointless and would probably result in a temp INCREASE thanks to the amount of rad space you'll be using. Not to mention it's loud, produces a lot of heat, and IIRC uses 1/4" barbs.
Honestly, for a tribute to XS this build sure is ignoring a lot of XS' principles, with the exception of a low price/performance ratio.