Water cooling loop. Mcp655 -b Zero / No flow at all.

FaulkofSeagulls

Reputable
Mar 25, 2015
10
0
4,510
Hey first post here and newb with water cooling. My loop is pretty long (ambitious for first try) with two 360 and one 140 rad, 2gpu blocks and CPU block. Everything is hooked up and I tried to run the loop for the first time. The mcp655 I have is hooked up to an external Psu I had laying around. The res is 3/4 full and runs directly to the pump. When I turn it on I hear the pump running (low hum) but the water doesn't move at all. If I open my drain valve water runs through the pump and out (due to gravity) but when the drain is closed water just sits at the bottom of the loop. I have tilted and shaken the case to bleed air, and run several liters of water through the pump but still no flow at all.

Am I missing something? Any suggestions? My initial thought was that the loop was so large th pump can't overcome the pressure of the trapped air in the top of the loop, but I have tried opening plugs and installing pressure valves at different places to no avail.

I knew this would be a challenge but now I am pretty frustrated.
 
Solution
If you have put as much liquid in as possible and still no flow (keep the drain shut), then your pump might not have enough push for the length of tubing. But, I'm using the MCP35X2 to push my fluid through 6 rads and 4 water blocks. You shouldn't have any problem with flow unless you don't have something positioned correctly.

Just going to do a checklist.
Did you wash out the rads before putting them in the loop?
Did you run distilled water through all tubing to clean out particles and make sure there were no blockages?
Have you put the reservoir at a higher position than the pump?
Have you traced the tubing to make sure that you have one loop and are not accidentally stopping the flow by not having it connected properly?
Do you have...

jdcranke07

Honorable
If you have put as much liquid in as possible and still no flow (keep the drain shut), then your pump might not have enough push for the length of tubing. But, I'm using the MCP35X2 to push my fluid through 6 rads and 4 water blocks. You shouldn't have any problem with flow unless you don't have something positioned correctly.

Just going to do a checklist.
Did you wash out the rads before putting them in the loop?
Did you run distilled water through all tubing to clean out particles and make sure there were no blockages?
Have you put the reservoir at a higher position than the pump?
Have you traced the tubing to make sure that you have one loop and are not accidentally stopping the flow by not having it connected properly?
Do you have more than 10ft of tubing and water blocks? If so, you do not have enough power from that pump. Keep in mind that if you have upward flow that counts against the pump as well.
 
Solution

FaulkofSeagulls

Reputable
Mar 25, 2015
10
0
4,510
Thanks for your answer jdcranke.
-yes rads washed out
-clean tubes and blocks
-res is higher than pump
- tubing loop is good. Only thing is that the runs are long with a few long verticle runs. I would say I am at about 10 feet of tubing.
- this pump was recommended as one of the best for higher pressure longer loops. Any other suggestions for pump? Any other troubleshooting before I drop more cash?
 

jdcranke07

Honorable


Yes, if you have a larger inner diameter tubing you will need a much more powerful setup especially with that much tubing. If you need to you can hook another of the same pump up in tandem or you can return the one and get the same one I have. Swiftech MCP35X2. I use 3/8'I.D./5/8"O.D. Primochill tubing. 1/2'I.D. tubing with that much tubing will need two pumps for flow. The max head for your current pump is 10ft. and the MCP35X2 is 27.9ft. Max nominal head @12V is the figure to look for.
 

rubix_1011

Contributing Writer
Moderator
D5 should be fine for this. Radiators are very low flow resistance in almost every case.

I had a single D5 for a CPU block, 2 GPU blocks and 2x360 rads and it ran fine. 10ft tubing also is fine.

Is the pump primed, assuming that your loop is fairly full so the pump cannot run dry?

What GPU blocks do you have and how are they setup...serial/parallel? I've seen where people got the I/O ports swapped so it essentially blocked flow through GPUs...so just wanted to verify.
 

FaulkofSeagulls

Reputable
Mar 25, 2015
10
0
4,510


Ugh that must be it. Poor planning on my part. I dont think when i purchased everything for the loop I envisioned the it ending up over 10ft. I didnt take into consideration the tube diameter as well. My ID is 5/8, and I think i am just over 10 ft overall. I'll check out the mcp 35x2 and any pumps with better pressure. I am assuming there is no single pump solution strong enough.

Thanks for your replies. I appreciate you not calling me an idiot right away....NOW its ok though lol.
 

rubix_1011

Contributing Writer
Moderator
Head doesn't limit your loop length, though; it simply shows you how much power a pump has to push a column of water, vertical without restriction. Think, torque of a car engine.

In a loop, a pump is always pushing AND pulling liquid, much like a toy train on a circle of track that has enough cars to connect engine to caboose. Your issue is that you have air or obstruction in your loop causing the pump to be unable to push past where you see the water. If it were actually an issue with too much tubing or loop length, you'd see the pump get the water further along in the loop than you are seeing. Check for obstructions or kinks first, then see if you can manually prime your loop yourself...like siphoning water or drinking through a very large straw.

If the water isn't making it very far beyond the pump, you have something blocking it.
 

FaulkofSeagulls

Reputable
Mar 25, 2015
10
0
4,510


Hmmm Hey Rubix. Well now i'm a bit more confused taking your reply into consideration.

The GPU waterblocks are 'EK-FC R9-290X (Rev.2.0) http://www.ekwb.com/shop/blocks/vga-blocks/ati-radeon-full-cover-blocks/radeon-rx-200-series/ek-fc-r9-290x-rev-2-0.html. I have them set up in serial. I'll double check the setup again but the water rises past both of those blocks if I fill the res high enough, so water is getting through.
 

FaulkofSeagulls

Reputable
Mar 25, 2015
10
0
4,510


I'll have to go through the loop and piece by piece trouble shoot each leg and block to rule out blockage as you describe. It will be a long painful troubleshooting process (I used rigid tubing so rerouting and isolating the loop will require buying soft tubing to make changes and smaller loops)'

Ill do this before buying another pump.
 

jdcranke07

Honorable


Typically, you are correct that a D5 should be enough. It really depends if there is a blockage which was my first guess and on how he actually has the tubing run through the system. Too many turns or vertical runs would kill the power of the pump if the head is lower. Granted your example is sound also keep in mind that if you try to pull an 53ft 18 wheeler trailer with a GEOmetro engine you will not get farm. This is an extreme example, but it should show that depending on the diamter and the length and the overall resistance that a pump that has only 10ft of head might not be enough. This, again, boils down to if there is a blockage or his overall setup.
 

FaulkofSeagulls

Reputable
Mar 25, 2015
10
0
4,510


My drain valve is off of a T fitting between the pump and my first rad. Yes the res fill port is open when i turn it on (I have tried both). The res only empties if I open the drain port. otherwise the water just sits there in the res (well it also fills the bottom line and runs up to the first rad. After moving the case around a bit and filling the res to the top some of the water travelled up through my first rad and then just through my gpu blocks. It stops there as the water is then level with the res water level. The pump being on doesnt seem to move the water at all (though it is humming nicely)
 

rubix_1011

Contributing Writer
Moderator
Is your reservoir fill port open when you are turning the pump on? If so, this should allow air to purge while the pump displaces air throughout your loop. If your reservoir fillport is open and the same thing results, there has to be blockage somewhere that is preventing air movement.

Is your drain (that you mentioned opening) after your GPU blocks? Where in the loop order is it? It would seem logical that if that is open, the loop flows fine (is this correct?) but closed, it does not? This would point to me that the issue lies after the drain valve in the loop, however that is configured.

Edit: we're posting over the top of each other...just read your last post. Can you post an image of your loop, and list the loop order of your components, including the drain location?
 

FaulkofSeagulls

Reputable
Mar 25, 2015
10
0
4,510


Sorry for the cross posting. I am overseas, and its the middle of the night here so I'll have to take some pics tomorrow (dont want to wake the wife). I will try to upload a quick diagram so you have the basic idea.
]http://[/url]
 

rubix_1011

Contributing Writer
Moderator
OK, you are battling a lot of air. This isn't to say this can't work, because it can, but your loop order is different than most people. Not a problem, really...loop order doesn't really impact cooling. You just want your pump inlet lower than your reservoir.

You are likely going to have to manually prime your loop...think sucking water through a really big straw. Otherwise, you can disconnect your tubing about 1/2 way and prime it, reconnect, and continue onward. Also didn't realize you had a MB block in there, also. To be honest, you also probably have more radiators than you need, but that's really only an issue with purging air, not necessarily flow restriction.

One thing I didn't ask...are you using any quick disconnect fittings? It's been known that using several of these can easily kill a loop by adding excessive restriction. Some are better than others, also.

My loop order would have been:

Res > pump > cpu > MB > 480 > 360 > gpu1 > gpu2 > res. Cutting out the 140mm rad. Otherwise, add the 140mm after gpu2 then to res.

Your loop is a bit more restrictive than I originally though with the inclusion of the MB block, although while it looks nice, it probably isn't benefiting you a great deal unless you are planning on high CPU overclocks.
 

FaulkofSeagulls

Reputable
Mar 25, 2015
10
0
4,510


Ah ok. I thought manually priming just referred to making sure that water was through the pump before it started. I'll try to see if I can manually flood most of the lines as well.

Yes the motherboard part of the loop was def just for looks and overkill. Actually, I realize most of the build is overkill for looks and for overkills sake and the fun of the project (case wasnt meant to house half this stuff so i got to rip up the case and drill holes etc...fun for me).

I have a busy work schedule ahead of me over the next couple days. But this weekend I will try the following:

-re-verify the proper i/o ports are being used at each point
-Flooding all the lines manually (will filling the rads first help too?) and trying to pull air through manually before kicking on the pump.
-Ill remove the 140 from the loop entirely and slowly try to test each segment individually if I can.
-Pull my pump out and test it using a more standard orientation. It has a bitspower pump top installed (again for looks) that allows multiple i/o combos. I should isolate this to make sure the pump top configuration isnt causing issues as well.
-LASTLY if none of this works / fixes the issue, i'll try a dual pump setup (not sure i have any room left in the case though)

Thanks for all the help you guys. I'll try to update this weekend.

Cheers
 

rubix_1011

Contributing Writer
Moderator
Was the pump top shipped installed, or did you install? I know there have been some reports of people having issues with tops and priming, but uncertain if that is one of the culprits.

And BTW, watercooling is typically overkill anyway, so don't let that deter you...it's the reason I do it. It certainly isn't needed but looks great and is fun to learn and install. Plus, you get to answer a lot of 'how did you do that' types of questions from your friends.
 

FaulkofSeagulls

Reputable
Mar 25, 2015
10
0
4,510


Yeah, the pump top came installed. Would pump orientation be a culprit too then? Pump is installed 'sideways'/vertical. the i/o ports on the pump top indicate it can be mounted this way but that may be an issue maybe? Pump oriented like so:

^
|
|
out

I_ in <------
 

FaulkofSeagulls

Reputable
Mar 25, 2015
10
0
4,510


Hey, so I wanted to post an update....good and disaster news!!.

Your suggestions worked. I flooded my tubes and radiators and forced water through the system and finally success!...my pump started pushing the water along. (I may still need a more powerful pump, or two, BUT its working. I'll will have to test flow rate and decide in the future)

THEN disaster struck! I had my whole system set back up and powered up. All was going smoothly UNTILL i freaking knocked over my fill bottle which promply tipped and dumped water directly into my PSU. After a thunderous electrical explosion my psu (AAAAND possibly my whole set up) is dead :-(.

I'll have to order another PSU and go through and test all my components individually to see if anything survived. Needless to say Im pretty freaking upset and angry at myself now.

So anyway....Thanks for your responses. If it werent for me being careless, everything appeared to have been working thanks to your replies.
 

newtowatercooling

Reputable
Nov 12, 2014
105
0
4,710


that sucks thats why some people use a second psu to power the pump while filling . ive spilled stuff inside my pc but since everything was off i was saved . hope its just your psu that dead how much did you still a few cups or liter?