D5 Pump Not Strong Enough?

[turok_t]

Hi everyone, I just got a Singularity Computers Protium D5 reservoir and an alphacool VP655 D5 PWM pump. I?m testing them out of my system to make sure that the pump works fine before installing it in my build. Water travels from the radiator up to the inlet of the pump/reservoir and then it leaves the pump into the green tube (testing only) going back into the radiator. My question is this, there is a huge air bubble in the green tube close to the outlet of the pump and it looks like the pump isn?t able to pump the water strong enough causing the massive air bubble. It looks like water is just trickling down the green tube. Is this normal? I made a video and posted it in the link below.

At first, I thought this might be caused by the 90 degree inlet bend restricting the water flow into the pump. But then again, the pump is gravity fed by the reservoir so there is plenty of water feeding the pump and this shouldn?t be a feeding/intake problem.

I tried using a different D5 pump (Laing D5T Vario) and I get the same result. I also tried a different PSU (EVGA 850W) and get the same result.

So is that massive air bubble normal when water is leaving the pump? Or do I have to wait for it to naturally bleed itself out of the loop? I?m worried that my pump isn?t strong enough, but then again, I know D5 pumps can push through a lot. Any comments/suggestions would be helpful, thanks.

 

[pendragon1]

you sure you got the ports right? looks like it. res is opened to let the air out?

 

[turok_t]

Yeah, res is opened at the top. Both of my pumps are a bit noisy (sounds like crackling) as well, can it be due to excessive air bubbles? If so, how can i get rid of air bubbles except for tilting, rotating and doing the radiator dance?

 

[pendragon1]

Yeah, res is opened at the top. Both of my pumps are a bit noisy (sounds like crackling) as well, can it be due to excessive air bubbles? If so, how can i get rid of air bubbles except for tilting, rotating and doing the radiator dance?

you didnt over tighten and jam up the impeller(?)? did it work at all? how are you powering it, 12v molex? am i seeing air bubbles coming out of the inlet tube?

 

[turok_t]

you didnt over tighten and jam up the impeller(?)? did it work at all? how are you powering it, 12v molex? am i seeing air bubbles coming out of the inlet tube?

Nope that at all, the impeller can spin freely on both my D5 pumps. This is the first time im using my pump which is being powered by a 12V molex.

Now, my radiator is higher than my reservoir, so im guessing as i fill in my reservoir, the radiator will be filled up to the same level as well. However, once i fill my reservoir to the max, there is still a gap at the top of my radiator. Will this top section be filled with water too as the pump pushes the water in? or will the water level of the radiator always match the reservoir? The reason why im asking is because i want the radiator to be completely full even though my reservoir is lower.

 

[pendragon1]

Nope that at all, the impeller can spin freely on both my D5 pumps. This is the first time im using my pump which is being powered by a 12V molex.

Now, my radiator is higher than my reservoir, so im guessing as i fill in my reservoir, the radiator will be filled up to the same level as well. However, once i fill my reservoir to the max, there is still a gap at the top of my radiator. Will this top section be filled with water too as the pump pushes the water in? or will the water level of the radiator always match the reservoir? The reason why im asking is because i want the radiator to be completely full even though my reservoir is lower.

yeah you rad should be full. try topping it(res) up more and the lean/tilt it around. it looks like its kinda flowing but it looks like air is getting suck in on the right, or it that the impeller spinning?

 

[turok_t]

yeah you rad should be full. try topping it(res) up more and the lean/tilt it around. it looks like its kinda flowing but it looks like air is getting suck in on the right, or it that the impeller spinning?

I think its probably both- air bubbles and turbulence from the water. I wish I can fill it up from the top of the radiator itself so that I know it is completely full!! Actually, now that I think about it, my D5 pump is probably running at full speed (connected to molex only) which may be producing a lot of the air bubbles/turbulences. Is there anyway I can change the speed of the pump without plugging in my PWM into the fan header of my mobo? I can't turn on my computer right now and I also don't have an external PWM fan controller.. Anyway I can control the speed of the pump outside of my system?

 

[pendragon1]

I think its probably both- air bubbles and turbulence from the water. I wish I can fill it up from the top of the radiator itself so that I know it is completely full!! Actually, now that I think about it, my D5 pump is probably running at full speed (connected to molex only) which may be producing a lot of the air bubbles/turbulences. Is there anyway I can change the speed of the pump without plugging in my PWM into the fan header of my mobo? I can't turn on my computer right now and I also don't have an external PWM fan controller.. Anyway I can control the speed of the pump outside of my system?

when you started filling it, it should have pushed all the way through the rad. my setup is a small 50mm res with a ddc knock off and it pushed through my gpu's reused aio block, a 120, my cpu block, and a 240 no prob. might need to just drain it and start over. only put about 200ml into the res and see if it can push through the rad.
you can drop it to 7v by moving the ground pin to the outside

 

[turok_t]

when you started filling it, it should have pushed all the way through the rad. my setup is a small 50mm res with a ddc knock off and it pushed through my gpu's reused aio block, a 120, my cpu block, and a 240 no prob. might need to just drain it and start over. only put about 200ml into the res and see if it can push through the rad.
you can drop it to 7v by moving the ground pin to the outside

View attachment 366218

Thanks for the diagram Pendragon1 ! The 2 wires from the molex on my pump are RED and BLACK. Is the RED one the same as the YELLOW wire shown on the right half of your diagram?? So basically, all I need to do is swap out the wires within the molex to decrease the voltage? It's not going to short out or fry my pump right ?

 

[pendragon1]

no it wouldnt cause a problem BUT i just wanted to double check and apparently DDCs need min 9v. so no way to do that without a rheostat. so id drain it, make sure everything is tight and start again with less water. push it through, add more and repeat.

 

[turok_t]

no it wouldnt cause a problem BUT i just wanted to double check and apparently DDCs need min 9v. so no way to do that without a rheostat. so id drain it, make sure everything is tight and start again with less water. push it through, add more and repeat.

Are you saying add a bit of water into the res, push it through until its finished, add more water into the res, push it through again and repeat? Or is it better to fill the res to the max and let the pump circulate all the fluid at once and only add more if the res depletes?

Here is something that might be helpful. I have a 200mm res. Initially, I filled it up to 150mm and then i turned on the pump and the water level drops to 70mm and over time it drops to around 20mm and stabilizes here when some of the air bubbles bleed out. This was for a total of maybe 1.5 hours? Im not sure if the radiator is completely filled?

 

[pendragon1]

Are you saying add a bit of water into the res, push it through until its finished, add more water into the res, push it through again and repeat? Or is it better to fill the res to the max and let the pump circulate all the fluid at once and only add more if the res depletes?

Here is something that might be helpful. I have a 200mm res. Initially, I filled it up to 150mm and then i turned on the pump and the water level drops to 70mm and over time it drops to around 20mm and stabilizes here when some of the air bubbles bleed out. This was for a total of maybe 1.5 hours? Im not sure if the radiator is completely filled?

yeah but not so low as to run dry. its supposed to work either way.
im guessing the rad will hold no more that 200ml, if its 25mm, its a surprisingly small amount for the size of the things. maybe it is full and you just need to tilt it around to finish bleeding it. its hard to tell with that setup if its flowing. are you planning on adding a flow indicator of some sorts?

 

[turok_t]

yeah but not so low as to run dry. its supposed to work either way.
im guessing the rad will hold no more that 200ml, if its 25mm, its a surprisingly small amount for the size of the things. maybe it is full and you just need to tilt it around to finish bleeding it. its hard to tell with that setup if its flowing. are you planning on adding a flow indicator of some sorts?

It will but the air will still get trapped in there. The tilting and shaking will free it up to move as you do it though. It just takes time to get it all out. It took me most of a day of on and off tilting, powering on and off and an overnight run to get all the air out when I built my current loop. At least yours is out if the case. I had two radiators the res and blocks in a huge Enthoo Primo to muscle around.

I was debating whether to fill the res bit by bit or push it through all at once. I thought if i did it all at once, the water would be more continuous within the loop, and hence, less gaps/air bubbles. Im using the black ice nenemis 360GTS, the website doesnt show the actual capacity/volume of the radiator, it just says

• 398mm x 133mm x 29.6mm (L x W x H)

Actually, here is a picture of my completed build. I took out one of my radiator and pump/reservoir to test it out since there was a leak and that's how i was able to make the video in my original post. Anyhow, you see how i have a horizontal radiator at the top? im worried that water can't fill that radiator completely since it is at the highest point.

 

[pendragon1]

I was debating whether to fill the res bit by bit or push it through all at once. I thought if i did it all at once, the water would be more continuous within the loop, and hence, less gaps/air bubbles. Actually, here is a picture of my completed build. I took out one of my radiator and pump/reservoir to test it out since there was a leak and that's how i was able to make the video in my original post. Anyhow, you see how i have a horizontal radiator at the top? im worried that water can't fill that radiator completely since it is at the highest point.

View attachment 366225

it should still push through but will need bleeding. yes in that setup, id mostly fill the res, push it through and then top it up as it gets low.

 

[turok_t]

So my pump should be able to push through that setup completely (ie. fill up both radiators completely and keep the coolant flowing) right? Im worried that the pump wont be strong enough

 

[pendragon1]

So my pump should be able to push through that setup completely (ie. fill up both radiators completely and keep the coolant flowing) right? Im worried that the pump wont be strong enough

yeah it should. ive seen plenty of dual 360s with a single DDC and res or distro plate.

 

[turok_t]

Alright, when i guess i will try it out tomorrow and report back to you. But thanks so much for all your help and suggestions, I appreciate it greatly !

 

[pendragon1]

no prob!

 

[Nobu]

The way radiators are built varies, but generally the internal structure looks like this:

_____
| | |
| | |
| | |
\o|o/

Water goes in one side, travels up several tubes to the top, crosses over to the right side in an open chamber, then back down another set of tubes on the other side before exiting the radiator.

Problem with this orientation is (unless the radiator is full) the water will tend to trickle/cascade on the second half as air tries to climb back up into the cavity at the top. Of course you have the same problem in the opposite orientation, just with one side. Best possible orientation would be sideways with the inlet on the bottom and outlet on top–you'd still get some air trapped, but most of it would naturally pass through.

But since flipping and rolling the rad isn't really an option here, the next best thing would be to open the top, fill it to the brim, close it up and agitate gently (or start and stop the pump in 3-6 second intervals) to try and move whatever air is still trapped in the radiator to the res. Do this a few times, then run it for 30-60 seconds (or until you hear air get into the pump). Pop the top again and if it's low, put a bit more (enough to cover the tubes at least). The rad should be good at this point–any remaining air should make it's way to the res eventually, and there should be enough water that it won't sound like a waterfall.

 

[turok_t]

Alright, I think I fixed it after some testing today.

Here is a video of the pump spinning, notice how its spinning quite fast and generating a lot of air bubbles? The air bubbles are then fed back out from the pump into the water loop. Also, notice how loud the pump is? With all the pumps I used in the past, it never sounded this loud.



After connecting it to a separate computer, this is how my pump sound like now. It’s completely inaudible (the background noise you hear is from the computer at the back). There are minimal air bubbles in the reservoir and inlet and outlet tubes from the pump. This is how quiet I am used to with my previous pumps. When I went into the BIOS, my pump was spinning at 1800RPM (Normal Speed). If I increased the pump speed to 4787RPM (Full Speed), it sounds exactly like the loudness in the first video which sounds like a turbine. I also tried used a different inlet (top of my multi-port cap) on my reservoir and the water seems to be flowing well and not trickling or splashing back in.



So all in all, it was indeed the pump speed/RPM. After modulating its speed, there are less air bubbles, less turbulence and the pump is much quiet now. I still did do a lot of tilting and rotating to bleed out the air bubbles as much as possible, but this looks and sounds much better than before. I’m more confident with my pump now and will reinstall it back in my system.

 

[pendragon1]

Alright, I think I fixed it after some testing today.

Here is a video of the pump spinning, notice how its spinning quite fast and generating a lot of air bubbles? The air bubbles are then fed back out from the pump into the water loop. Also, notice how loud the pump is? With all the pumps I used in the past, it never sounded this loud.



After connecting it to a separate computer, this is how my pump sound like now. It’s completely inaudible (the background noise you hear is from the computer at the back). There are minimal air bubbles in the reservoir and inlet and outlet tubes from the pump. This is how quiet I am used to with my previous pumps. When I went into the BIOS, my pump was spinning at 1800RPM (Normal Speed). If I increased the pump speed to 4787RPM (Full Speed), I can replicate the loudness in the first video which sounds like a turbine. I also tried used a different inlet (top of my multi-port cap) on my reservoir and the water seems to be flowing well and not trickling or splashing back in.



So all in all, it was indeed the pump speed/RPM. After modulating its speed, there are less air bubbles, less turbulence and the pump is much quiet now. I still did do a lot of tilting and rotating to bleed out the air bubbles as much as possible, but this looks and sounds much better than before. I’m more confident with my pump now and will reinstall it back in my system.

nice. those vids are private though... fixed. that looks much better.

 

[turok_t]

nice. those vids are private though...

Hey sorry about that, i just fixed the links. You should be able to view them now.

Pendragon1, i also filled up the res more so its 80% full. I tilted the radiator so that the inlet/outlet ports are elevated and i see the water line on my inlet/outlet tubes so i know the radiator is indeed full of water. Once I revert it back to the normal position, water starts to fill where the waterline was. So im pretty sure the radiator is full now.

 

[pendragon1]

Hey sorry about that, i just fixed the links. You should be able to view them now.

Pendragon1, i also filled up the res more so its 80% full. I tilted the radiator so that the inlet/outlet ports are elevated and i see the water line on my inlet/outlet tubes so i know the radiator is indeed full of water. Once I revert it back to the normal position, water starts to fill where the waterline was. So im pretty sure the radiator is full now.

saw that, good stuff. youll still need to run it for a while to get all the air out but your res will cover that no prob. mine burped after a couple weeks running and now i need to top up my res by about a 10mm.

 

[turok_t]

saw that, good stuff. youll still need to run it for a while to get all the air out but your res will cover that no prob. mine burped after a couple weeks running and now i need to top up my res by about a 10mm.

Wow 10mm, thats a lot! I am definitely going to tilt and rotate my case as much as possible when everything is installed.

You know the vertical rad in my picture? Is it bad that the inlet/outlet are located at the bottom vs the top? I have heard mix opinions about vertically mounting a radiator and whether the ports should be on the top or bottom.

 

[Nobu]

Wow 10mm, thats a lot! I am definitely going to tilt and rotate my case as much as possible when everything is installed.

You know the vertical rad in my picture? Is it bad that the inlet/outlet are located at the bottom vs the top? I have heard mix opinions about vertically mounting a radiator and whether the ports should be on the top or bottom.

It doesn't really matter–both have the potential to trap air at the top. Might have to add water occasionally to keep it from waterfalling.

 

[turok_t]

It doesn't really matter–both have the potential to trap air at the top. Might have to add water occasionally to keep it from waterfalling.

Alright, that's good to know. I don't want to redesign my hard line tubes again if i reverted the inlet/outlet valves.

 

[pendragon1]

Wow 10mm, thats a lot! I am definitely going to tilt and rotate my case as much as possible when everything is installed.

You know the vertical rad in my picture? Is it bad that the inlet/outlet are located at the bottom vs the top? I have heard mix opinions about vertically mounting a radiator and whether the ports should be on the top or bottom.

oh i did the tilt game, thought i got it all, didnt move for 2 weeks, then one day it dropped. no leaks so it much have been a burp.
no, doesnt matter. my front rad is tubes down.

 

[turok_t]

oh i did the tilt game, thought i got it all, didnt move for 2 weeks, then one day it dropped. no leaks so it much have been a burp.
no, doesnt matter. my front rad is tubes down.

Would the water evaporate over time?

 

[pendragon1]

Would the water evaporate over time?

it can sometimes but not that quick and my loop is sealed up tight.

 

[turok_t]

Hi everyone, here is an update on my build. Pump and reservoir has been installed back into my rig and I’m just leak testing everything. So far, there are no leaks, however, I’m wondering if my pump flow is being restricted. As you can see from the picture below, I have a sharp 90 degree angle as the pump outlet, and sharp 90 degree angle as the pump inlet. I’ve seen other builds with these angled fittings, so I thought it would be okay. Would the 90 degrees outlet restrict the water coming out of the pump? The reason why im asking is because I have a flow meter right after the water leaves the pump and it is showing 10 revs/second and 0.4L/min which is really low, especially when the pump is on full speed at 4700 RPM. My flow meter works perfectly as I tested it out before on my other rigs. On my other rigs, when my D5 Vario was on setting 3-4, it would show 40 rev/sec and 1.8L/min. But now on my D5 PWM pump it is showing much less flow at max speed. I also included a video of this at the bottom too, along with my entire water loop. I have a CPU and GPU waterblock and two 360 rad.





Here is the video:

 

[hititnquitit]

Does your mb have a pump specific header rather than a pwm only header? If so give that a try. It sounds like your pump isnt getting enough power. Even with it connected only to the psu it may go into a lower power mode.
How does the pump sound?

 

[turok_t]

Does your mb have a pump specific header rather than a pwm only header? If so give that a try. It sounds like your pump isnt getting enough power. Even with it connected only to the psu it may go into a lower power mode.
How does the pump sound?

Yeah my mobo has CPU_OPT which is meant to measure pump speed. On full speed, my pump sounds very loud and generates a lot of bubbles. It sounds like its in overdrive mode. Adjusting the speed to full speed in my BIOS sounds exactly like if i just use the molex connector. So i know it is indeed full speed.

 

[Nobu]

It looks like a lot of air is coming out of the pump, although there doesn't appear to be much going in. I'd hazard a guess that it's got air trapped inside the impeller.

Is the top of the reservoir open? Opening it may allow water to flow more freely into the res, and hopefully allow the trapped air out. Other than that, try turning the pump on and off a few times for short (2-4s) intervals to try to release the air.

 

[Tsumi]

Yes, 90 degree elbows will restrict your flow. How much? I think some very old testing said it took 8 90 degree elbows to match the restriction of one CPU block, so I would not worry about a few 90s at all.

I would not say that using the bottom inlets for your main flow path is a very good idea. The reason is that based on the design, the inlet goes directly to the pump inlet, which makes it very difficult for any trapped air to escape. It would be a lot of work to rework your design to feed from the top though.

I can't find any articles on it, but this is why the XSPC D5 tank reservoir went through 3 revisions (I have a revision 2). The first revision had an inlet tube that extended most of the way through the reservoir to direct flow to the pump inlet, presumably to reduce pumping losses and improve reservoir performance. What happened is that it was extremely difficult for any air to get out. The second revision then curved the tube up slightly, directing the jet of water to hit off-center so air could actually escape the system. It looks like the third revision shortened that inlet tubing, but hard to tell from the pictures.

Also, for future reference, the maximum head pressure of the D5 pump is 3.9 meters. That means with a tube pointed straight up, it can push water up to a maximum height of 3.9 meters. It's got plenty of head pressure to push water through 99.9% of watercooling systems.

 

[cpufrost]

It's plenty strong for typical builds. If you fall below the 1gpm baseline, add another. My rather complex SR2 build of yesteryear needed more than two so I got an Iwaki RD30 and all was well.

 

[hititnquitit]

Yes, 90 degree elbows will restrict your flow. How much? I think some very old testing said it took 8 90 degree elbows to match the restriction of one CPU block, so I would not worry about a few 90s at all.

I would not say that using the bottom inlets for your main flow path is a very good idea. The reason is that based on the design, the inlet goes directly to the pump inlet, which makes it very difficult for any trapped air to escape. It would be a lot of work to rework your design to feed from the top though.

I can't find any articles on it, but this is why the XSPC D5 tank reservoir went through 3 revisions (I have a revision 2). The first revision had an inlet tube that extended most of the way through the reservoir to direct flow to the pump inlet, presumably to reduce pumping losses and improve reservoir performance. What happened is that it was extremely difficult for any air to get out. The second revision then curved the tube up slightly, directing the jet of water to hit off-center so air could actually escape the system. It looks like the third revision shortened that inlet tubing, but hard to tell from the pictures.

Also, for future reference, the maximum head pressure of the D5 pump is 3.9 meters. That means with a tube pointed straight up, it can push water up to a maximum height of 3.9 meters. It's got plenty of head pressure to push water through 99.9% of watercooling systems.

Good eye Tsumi, thats what the problem is. Because of how close the inlet and outlet are to each other your pump is recycling the air bubbles in your loop causing your pump to cavitate (killing flow). With ek reservoirs they have a small plastic insert that goes in between the two to give the air a chance to disperse. Or with alot of reservoirs like Tsumi is describing they employ a tube to give the air a chance to go up the res before it gets pulled back into the pump. Some even have sponge like mediums to catch the air or both.
Does your res have any of these that just isnt installed by chance turok_t? I cant tell what brand your res is, bp?

 

[cpufrost]

Another issue with D5 pumps is their absolute intolerance to run dry or partially dry for any period of time. Doing so will irreversibly damage the bearing! 90s should be avoided close to the suction side always.

 

[turok_t]

Thanks for all your comments guys! anyways, I fixed the issue. The flow meter is much more accurate now and increasing the pump RPM doesn't create bubbles from the impeller. All the air bubbles are gone and the pump is inaudible even at high RPMs.

 

[pendragon1]

I fixed the issue.

how?

 

[turok_t]

The flow meter was not reading properly because the bubbles were obstructing its measure. As i continued to allow the system to bleed out the air bubbles with a lot of tilting, rocking and changing the pump RPM, the flow meter showed gradual increase in flow rate. Initially, the pump would be generating a lot of air bubbles (hence, cracking noise) even at 60% speed. But now, even at 75%, it seems to be quiet. I keep my pump at 60% speed as i dont see any difference in temps as i increase the pump RPM. Same thing with my 6 fans on my radiator. However, if my pump stops for a few seconds, temps will begin to rise quickly.

My old EK res had an anti-cyclone in the reservoir to give the bubbles more time to escape before being sucked into the pump again. I couldn't find any insert, anti-cyclone device that fits my singularity pump top, but it seems to be fine now. Before i can see a lot of air bubbles where the impeller is, but over time, they vanished.