July 25th 2014 - The WAN Show Document

[LAwLz]

What the hell was @Barnacules on about with the coding? You don't have to use pointers or other low level memory management with C. You can keep all your variables in the stack if you want.
Here is a "hello world" program in C:
 
#include <stdio.h>

int main(void){
    printf("Hello world");
return 0;
}
 
You don't even need the return function, so excluding all the necessary stuff (int main and include libraries) the entire program is 1 line of code...
I get that he was talking in hyperbole but if he continues like that he will scare people away from good languages (like C) and instead go for much worse performing and more locked down (in terms of compatibility) languages (like C#).
The last thing we need is more terribly coded, low performance programs in the world. C# is almost as bad as Java with the additional drawback of barely being supported outside of Windows.

[Internet Swag]

HOLY CRAP DID YOU GUYS SEE HOW IRRITATED LINUS GOT WITH BARNACULES WHEN HE KEPT INTERRUPTING HIM?

 

Guess Linus' has a fiery temper for a Canadian. 

[lament]

Hello guys,

I wait for workstations cooling project, and I think somebody (probably from LMG) should make separate topic for that project.

On WAN show casters told their idea, which is interesting, but I see some flaws.

 

First thing is idea of using separate pumps for each PC. I think it may cause some trouble. Flow of entire system will depend on amount of PCs being turned on, another problem may be pumps not being powered on, while having fluid flowing. It will act as turbine / energy generator so it will slow current and it may damage electronics in a PC.

Next thing is idea of using multiple pump type; industrial grade pumps before radiators, and MCP-655 in each PC. Using multiple pump types will cause inefficiency due to mixed flow rates and maximum head (how high pump can push water). Other thing is that water blocks on al CPUs and GPUs shoudn't cause more resistance than all radiators, so you don't need so many pumps. Also you need to decide where you will keep pumps. Inside office you will get extra noise, outside office pumps may face humidity, which may damage them.

All pumps, fans and controllers need to run on independent PSU, so it can work even if most PCs aren't working. PSU must be very high quality. Some industrial-grade pumps run on 24V DC to decrease current, so it will be harder to use ATX standard PSU

 

How I see this project: System would have 2 pumps, first before radiator, and second before PCs. Temp shouldn't be a problem. A delta temp water-air should be no higher than 20 deg Celsius. so even if outside temp would be 40 deg, the fluid would be 60, which is at 60 deg max temp for most pumps. Example pump I've seen in some overkill builds: Iwaki RD-30. Reservoir, which should contain at least 5l of coolant (which of course should be de-ionized water). would be before first pump. There would be two tube types; thick, which would go in this order  "thin tubes(PCs)>reservoir>pump>thin tubes(radiators)>pump"  and thin tubes which will be spitted from thick ones into radiators or PCs, and go back into thick after going though rads/PCs.

 

To make my idea easier to understand, I have made a scheme:

Black Dots are Quick disconnect, I didn't put those on radiators, because they aren't necessary. Also there aren't any of these on main stream (thick tubes), because they will slow the current.

 

PS. Sorry for language errors, I'm not English.

[onionring]

isnip-

your solution has some problems.

1. condensation:

the water in the loop cannot go below ambient temperature inside the building than means, 25C (77F) (there s a safety margin here) so you need to control the flow thru the radiators independently from the flow to the pcs.

2. winter happens, which means if you use water, it will freeze.(the comps aren t always on)

so you can either

use car grade antifreeze (prestone) which isn t great for computer parts,

use a water/water heat exchanger to separate pc water from the car antifreeze.

use two sets of radiators one outside for summer and one for winter and antifreeze.

now for the individual pumps in the computers, you could easily bypass them when plugged to the central cooling by putting the quick connects before and after the pump

(quick connect out) (pump inlet-outlet) (check valve) (quick connect in)

or since they want to put the computers as stand alone capable, you could put the cpu/gpu block on quick connects inside

with the individual rad-pump-res on quick connects and a power hub, both of which will only be power when the compiter is out of the big loop.

and the big loop just plugging to the gpu-cpu blocks.

now to control the flow rate, you need a pressure regulator to control the pressure at the pc inlet manifold connection which will allow different numbers of pc running without a relatively constant flow

[lament]

...

Okay, I didn't thought about freezing (32 deg today, horrible) I think idea of 2 loops seems the best one, another with two radiator setups seems too complicated and easy to screw it up, actually 2 loop reminds how thermal electricity plant works, just at much smaller scale. I still think doing pump for each PC is bad idea, and having just pump (and possibly radiator) on quick connect is good replacement. Using pressure regulators on PCs is not bad idea, but is another piece that can fail, also main loop will work with too much power, and in not powered PCs flow will still power pumps. Just don't put loop going back to reservoir. And I don't think external loop pump can work with so low temperature.

 

Also, why is it so important to keep water above room temperature, computer elements like CPU or GPU should work in temp -10 deg or higher?

[onionring]

Okay, I didn't thought about freezing (32 deg today, horrible) I think idea of 2 loops seems the best one, another with two radiator setups seems too complicated and easy to screw it up, actually 2 loop reminds how thermal electricity plant works, just at much smaller scale. I still think doing pump for each PC is bad idea, and having just pump (and possibly radiator) on quick connect is good replacement. Using pressure regulators on PCs is not bad idea, but is another piece that can fail, also main loop will work with too much power, and in not powered PCs flow will still power pumps. Just don't put loop going back to reservoir. And I don't think external loop pump can work with so low temperature.

Also, why is it so important to keep water above room temperature, computer elements like CPU or GPU should work in temp -10 deg or higher?

if the temperature of the water goes below ambient, the humidity in the air will condense on the pipes (like the windows of your car)

but in this case the pipes and blocks will drip on your mobo/psu/cpu/gpu.

what i proposed for individiual pc is to put the individual pump and reservoir as a unit that will only be put in when the comp is not on the main loop.

so either

quick connect 1 - cpu - gpu - quick connect 2 - quick connect 2 - indivial res - individual pump - individual rad - quick connect 1

or

main loop in - quick connect 1 - cpu - gpu - quick connect 2 - quick connect 2 - main loop out

so even if the pc is off liquid will flow thru but that won t have any negative what so ever...

pressure regulators are quite reliable from my experience and there would only be 1 of them (since all comps loop are similar, the flow in each would be too) the regulator would only enable you to keep relatively constant flow when comps would be removed from the loop. (it would route excess water back to the res)

putting two sets of rads is quite easy btw, it s only 2 valves (or even 1) but you could put more if you want to vent it for winter...

[lament]

...

Yes, I understand it, I'm just saying there are two way of doing it, both with pros, and cons. The idea of having two sets of rads is bad due to cost of getting another set of Noctua fans and radiators, or changing its position at least twice a year.

One on the left is what you think is better, I think the second seems to be cheaper and more efficient.

Thanks for explaining temperature thing.

[onionring]

Yes, I understand it, I'm just saying there are two way of doing it, both with pros, and cons. The idea of having two sets of rads is bad due to cost of getting another set of Noctua fans and radiators, or changing its position at least twice a year.

One on the left is what you think is better, I think the second seems to be cheaper and more efficient.

Thanks for explaining temperature thing.

nop that s not what i proposed, see my drawing, there s 2 pumps to better control the temperature of the fluid and the individual pumps would only be plugged in the conps when they re out of the main loop.

[lament]

nop that s not what i proposed, see my drawing, there s 2 pumps to better control the temperature of the fluid and the individual pumps would only be plugged in the conps when they re out of the main loop.

Sorry, I misunderstood you, but I'm talking about system with heat exchanger, and i skipped external loop on those drawings, due to my laziness. And I agree, that second pump should be controled to keep temp right. The only thing I disagree with you is putting loop going back from pump/heat exchanger (whichever is last before loop splitting to each PC) to reservoir. 

[George Hart]

On the freezing topic, if you left the main pump on all of the time, the flow would prevent it water from freezing (unless it got really cold), and this would also prevent algae growth. Just an idea.

[Syntaxvgm]

On the freezing topic, if you left the main pump on all of the time, the flow would prevent it water from freezing (unless it got really cold), and this would also prevent algae growth. Just an idea.

I live in Indiana and it got to like -20 last year. I can't image what it'll be there. 

[George Hart]

I live in Indiana and it got to like -20 last year. I can't image what it'll be there. 

 Coldest temp last year in Vancouver was -10 Celsius, if the liquid went back through the house it would warm up the coolant, as long as the inside temp wasn't below freezing. Just an idea, might not work.

[onionring]

Coldest temp last year in Vancouver was -10 Celsius, if the liquid went back through the house it would warm up the coolant, as long as the inside temp wasn't below freezing. Just an idea, might not work.

nop that wouldn t work, if the water temps get colder than ambient temp inside, condensation will happen, and water on cpu/gpu blocks isn t somethkng you want...

and the surface are inside is way smaller than outside so there s no way it ll heat uo the fluid.

[dreamss]

wow linus.... at least do it geo cooling style ( i be glad to help with this project so it done properly cause it sounds like you guys got high and were like "man we should do this..." but dint really put tought into it)

im guessing you people are the kind newbs that use compression fittings, transparent tubing, colored coolant..  sound like a very dumb implementation.. not to mention outdoors will clog the rad SO quickly

 

what a waste of hardware  gezzuz h chist

 

or get a 4u and rack and have the pcs outside the room like someone mentioned, for example my buddies rig http://www.xtremesystems.org/forums/showthread.php?201013-Norprene-and-4U-lovin

 

i need to install my phasechange and show u how its done... worst geeks ever

 

you wont get condensation unless ur temps go into the negative, and even then u can easly protect a rig from it

 

 

Copper and Aluminum 

 

love people who dont know what happens when u use an aluminum rad with copper waterblocks (take a basic chemistry class plz) god i hate those aio corsair loops people get... underperforming as heck and use alu rads

[sickgr33k]

hey guys new to the forum here but just posting this as i thinks its the most relevant place to post it...

 

In regards to the whole room watercooling loop, i think if i recall correctly Linus said that in one of the wan show episodes that they struggled to significantly raise the temperature of the water in the loop and that the radiator might not even need fans in the winter...

 

Anyway my question is, how do you know the copper pipes you have running along the walls wont in themselves "radiate" heat into the room? and negate the desired effect of the loop, so instead of the loop transferring heat to outside the house, the loop is also heating the room up due to the amount of copper pipe that runs along the walls.

 

I have a feeling you should have just used PVC pipes instead as they have far greater insulation properties than copper, and would hence allow the most heat transfer to occur at the radiator... If you are finding the copper pipes are radiating a fair bit of heat, you may have to insulate the "hot" pipes with foam. Anyways hope it all goes well, i am looking forward to seeing an in depth video of the watercooling loop in operation

[sickgr33k]

Once again I'm relying on the community to save me from watching my own show... If you're watching the show and wouldn't mind filling in the timestamps in the below Google Doc... You're AMAZING!

 

News Items

 

 

 

hey guys new to the forum here but just posting this as i thinks its the most relevant place to post it...

 

In regards to the whole room watercooling loop, i think if i recall correctly Linus said that in one of the wan show episodes that they struggled to significantly raise the temperature of the water in the loop and that the radiator might not even need fans in the winter...

 

Anyway my question is, how do you know the copper pipes you have running along the walls wont in themselves "radiate" heat into the room? and negate the desired effect of the loop, so instead of the loop transferring heat to outside the house, the loop is also heating the room up due to the amount of copper pipe that runs along the walls.

 

I have a feeling you should have just used PVC pipes instead as they have far greater insulation properties than copper, and would hence allow the most heat transfer to occur at the radiator... If you are finding the copper pipes are radiating a fair bit of heat, you may have to insulate the "hot" pipes with foam. Anyways hope it all goes well, i am looking forward to seeing an in depth video of the watercooling loop in operation