D5 Pump Not Strong Enough?

str · 16 Jul 2021 at 06:30

Forgot to say it looks amazing.

Also, the loop might look like it's full but evidently the radiators still have a lot of air in them. The loud noise is a symptom of air still in the loop.

After the loop is filled you'll be able to reduce the pump speed to make it virtually silent while still having a good flow rate.

Also to help visualise why it isn't working properly yet. Think of a partially filled loop as pump outlet -> [rads full of air] <- pump inlet and when the loop is full it's now pump outlet -> [rads full of coolant] -> pump inlet. The pump needs our help to replace the air in the rads with coolant and usually all we have to do is carefully tilt the case back or forward while the pump is running.

dave1983 · 16 Jul 2021 at 15:13

Hi I had similar issues with my water-cooling loop, the flow was really low even at 100% speed. I suspected a faulty pump so i cleaned out the propeller on the dc5 and installed it back in, but still restricted flow.

Tried different coolants and was really baffled as to why there was no almost no flow in my system.

So i carried out a deep clean of the loop and the suspect was the jet plate in the CPU block, the amount of gunk that was causing back pressure and low flow rate.

The jet plate is really easy to get to, however it does mean emptying the loop. If however the jet plate is clean i would suspect that there is a lot of air in one of the rads, the only way to really solve it is a deep clean, i normally give the rads a good shake to get rid of anything that may block flow.

Another suggestion is to change the the pump speed to 5 to get rid of all the air then decrease it to 3-4

flea.rider · 16 Jul 2021 at 15:56

now in both my res i have a pipe that go's about half way up this is where the water comes into the res ..hence keeping bubbles away from my outlet way easier to fill it ..

MikeTimbers · 16 Jul 2021 at 16:33

Lots of gentle rotation of the case in all directions while keeping an eye on the reservoir level. One of the problems with this sort of build is that the sheer number of bends to keep lines "straight" (I count >15 90 degree bends) is that even bleeding the system becomes a PITA.

My tips for bleeding. Pump on 100%. Only pump connected to any power. Short bursts of the power whenever the res is full while rotating case. As soon as air appears in the res, power off and refill res. The other thing - almost the most important thing - is the loop really can't be closed during this process. That sounds like a contradiction to the earlier points but a closed loop seems to fill more slowly.

Last tip: you have a fill port on the top of the reservoir? Fit a barb/fitting, some soft tubing to that. Suck on the tubing.

im4gine3 · 16 Jul 2021 at 19:58

There was a thread similar to this one a while back where the guy had identical issue to yourself and that his flow meter was trotting along snail pace.. TLDR, bleed the system and check each block internals for gunk. That is usually the main culprit. Once cleaned refill the loop with the pump at 100%, lots of tilting later and you should be good to go. Having multiple 90 degree connected shouldn't be an issue for a D5... one of my rigs has about the same number if not more than yours and is operating fine

x-X-MIKEEE-X-x · 17 Jul 2021 at 09:03

After having DDC pump in my system for five years i decided to switch to a new res pump combo.

Now i have a Aqua Computer ULTITUBE D5 200 PRO pump res combo and I'm just not impressed with the performance I'm getting temp wise CPU GPU so far.

The D5 is nice and quiet at full speed but i could run the DDC just as quiet and get better performance with the same loop configuration.

Is it too early to judge as its only been running for about 19 hours at full speed? Or could it be there is still trapped air somewhere in radiator?

MikeTimbers · 17 Jul 2021 at 12:16

highly unlikely to see any temperature difference because of a switch from a DDC to D5 or vice versa,

x-X-MIKEEE-X-x · 17 Jul 2021 at 14:48

MikeTimbers said:
highly unlikely to see any temperature difference because of a switch from a DDC to D5 or vice versa,

Higher pressures in water blocks generally cool better (turn the pump speed to lowest and see temps climb a bit). Water flow is definitely down.

il have to just give it time.

LuckyBenski · 17 Jul 2021 at 20:15

D5 pumps dump their heat into the loop but this is only in the order of ~10W so I'd be very surprised if it made a noticeable difference to temperatures.

turok_t · 17 Jul 2021 at 20:45

Anyhow, I fixed the issue. The flow meter was reading inaccurate readings because the bubbles were obstructing its measure and blocking the turning wheel inside. 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. I did NOT want to put the pump on 100% (yes, i know it helps with the flow rate to fill the circuit and to push the bubbles out) because the impeller was creating a lot of air bubbles and reintroducing them back into the loop. It was generating more bubbles then bleeding them out at high RPM. That is why i would leave the pump on normal/silent with rocking and tilting, and occasionally, I would put the pump on short bursts of full speed to bleed out the bubbles. Before, the pump was making lots of cracking noise (from air bubbles) at speeds above 60%. But now when all the bubbles have bled, i can increase the pump speed to 75% without any bubbles or crackling noises. I leave the pump at 60% now since i dont see a big difference (if any) in temperature drops. Even with my radiator fans (6 corsair fans on 2 black ice BTS 360 rads), increasing the RPM also doesnt help with temps. So far my 5900x is idling at around 32C to 34C and when it loads, 60C with cinebench, and max 70C when playing games (ie. Battlefield V)

Edit: All my parts were brand new. I cleaned and flushed both rads before installing them and my cpu jet plate/fins were all clean as well. It took approximately 12 hours for all the bubbles to bleed out. I have 4 x 90 degree bends in my loop. The other bends are more curved and gradual from the tubing.

MikeTimbers · 17 Jul 2021 at 20:52

x-X-MIKEEE-X-x said:
Higher pressures in water blocks generally cool better (turn the pump speed to lowest and see temps climb a bit). Water flow is definitely down.

il have to just give it time.

Never said pressures don't make a difference. I said you wouldn't see much temperature difference between a DDC and a D5. Their performance is too close.

x-X-MIKEEE-X-x · 17 Jul 2021 at 22:48

MikeTimbers said:
Never said pressures don't make a difference. I said you wouldn't see much temperature difference between a DDC and a D5. Their performance is too close.

100% vs 100% power the DDC is better in my system by around three or so degrees at same coolant temperature using aida64 FPU and GPU stress.

Before with the DDC at full speed i got no more than 80 Celsius Tdie 3600x.
D5 I'm getting around 84-85 Tdie. (same coolant temp)
Aida64 FPU and GPU simultaneous.

It could possibly be my CPU block is just too restrictive and not optimised for ryzen 3000 series (corsair xc7).