HELP: Black Magic S**t Going On, Flow Rate Slowly Drops After Boot [Even with 2 pumps]

[Hexenhammer]

I have Black magic sh**t going on, because there is no logical explanation for this.

I have both Digital Flow meter with screen and Transparent square Water tank so i can confirm the flow meter results Visually. [and its not PWM, its disabled and hardware monitor in bios shows constant pump speed]

 

1st Loop Try: 18W LAING DDC Pump inside Bykski Reservoir combo > GPU Block  > GTX360 Rad > Heatkiller IV CPU Block > GTS480 Rad > Back to res

I had water filter + sliding valve that has small holes around the tube + 5 quick disconnects 0.6L/m

2nd Loop Try: Removed water filter and valve, NO CHANGE

3rd Loop Try: Added Barrow D5 pump Full speed 4800RPM after the LIANG DDC 4300RPM the flow starts strong and then drops back to 0.6 L/m

Even with just the DCC by itself enabled, sometimes it would start at 2L/m and the flow is so fast the water looks like boiling, but then it slowly drops.

 

I left to do car test for 1.5 hour, let the system bleed air with 2 pumps on FULL SPEED, came home to see 0.4L/m

 

4th Loop Try: I get psychotic and remove all QD

No change, flow starts strong, sometimes as fast as 4.0L/m and then slowly drops down [again before anyone says the flow meter is broken, i confirm everything visually, when its strong it looks like boiling water and sounds like waterfall, when it goes down, it looks like dripping]

 

MY GTS480 is installed with ports down, I assume maybe air pocket moves and block, i flip the HUGE Phantecks 719 case on its top, small change.

 

5th Loop Try: I decide to  remove the GTS480 from the loop and see if it helps.

 

Now I have DDC 18W full speed/Res combo > D5 Full speed > Heatkiller IV > GPU block > GTX360 > back to res

 

Same thing, flow starts strong and then drops down, I would pull my hairs out if i had any

 

Right now when i type this post, I booted the PC and the flow already dropped to 1.7L/ [102L per hour]and stays there, but its not enough for 2 pumps and thats with LAING 18W DDC running full speed and Barrow D5 running full speed, only GTX360 rad, CPU block and GPU block installed.

 

Guys im loosing my mind here, what can it be?

 

P.S.I plan to disassemble and re-do anyway, need to remove the wifi card from below the shroud, the damn thing has 128 byte payload and it affects all PCIe devices connected to the same switch

But im afraid that when i re-assmeble ill return to same situation, I also plan to install the GTS480 with ports up, the GTX360 is at the top of the case so i cant change its position anyway.

[Hexenhammer]

edit

[vf1000ride]

In between the different iterations have you taken the blocks apart or inspected the cpu/gpu blocks to make sure the fins aren't partially clogged?

 

I had seen this same basic thing in my system earlier in the week.  My situation is I rebuilt mine and ended up with a small amount of plastic shavings in the system after making new hard lines.  The plastic passed through the larger fins in my GPU block but got hung up in the CPU where I couldn't see them.  When the fluid was fresh and had a fair amount of fine bubbles in the system the water seemed to be moving normal speed but as the air settled out it looked like my flow was slowing down.  I ended up having to open up and clean my cpu block to get the gunk out of it.  Been back to full speed since.

[Hexenhammer]

7 minutes ago, vf1000ride said:

In between the different iterations have you taken the blocks apart or inspected the cpu/gpu blocks to make sure the fins aren't partially clogged?

 

I had seen this same basic thing in my system earlier in the week.  My situation is I rebuilt mine and ended up with a small amount of plastic shavings in the system after making new hard lines.  The plastic passed through the larger fins in my GPU block but got hung up in the CPU where I couldn't see them.  When the fluid was fresh and had a fair amount of fine bubbles in the system the water seemed to be moving normal speed but as the air settled out it looked like my flow was slowing down.  I ended up having to open up and clean my cpu block to get the gunk out of it.  Been back to full speed since.

Thank you, What you say makes lots of sense and i really hope it was that easy.

But:

1) Im using Soft Tubing

2) All New never used parts

3) Both rads were pre-clean [outside the system] with tap hot water flow and then filled with Rad clean kit

4) GPU block is transparent and vertical so it leaves just the sealed CPU block

5) Before filing the loop i run 24 hours of that pre-fill clean mix, forgot how its called, small tiny bottle you mix with distilled water and run for 24 hours and back then i had water filter installed on CPU block input and it came out clean

 

Can it be paper towel dust? The only thing that looks similar to what you said, i had to disassemble the GPU block [it leaked from pressure] and cleaned with paper towels, so dust from them?

but since i need to remove the wifi card [which hidden under the rear ports shroud] ill be dissembling everything so ill have a chance to open up the CPU block and just in case ill clean it with tooth brush.

 

[Hexenhammer]

3 hours ago, vf1000ride said:

In between the different iterations have you taken the blocks apart or inspected the cpu/gpu blocks to make sure the fins aren't partially clogged?

 

I had seen this same basic thing in my system earlier in the week.  My situation is I rebuilt mine and ended up with a small amount of plastic shavings in the system after making new hard lines.  The plastic passed through the larger fins in my GPU block but got hung up in the CPU where I couldn't see them.  When the fluid was fresh and had a fair amount of fine bubbles in the system the water seemed to be moving normal speed but as the air settled out it looked like my flow was slowing down.  I ended up having to open up and clean my cpu block to get the gunk out of it.  Been back to full speed since.

Hi,

So i disassembled it and the Heatkiller block has such tiny microscopic channels that they gunked up with what looks like dust particles from paper towels or from inside the black rubber tubes that I use.

What do you think? Should I install water filter on the CPU block Input, i know ill lose some flow because of it but at least the CPU wont gunk-up and maybe its a better alternative to not doing anything.

Or maybe I should buy a cheaper CPU block with bigger water cutouts, whatever they called?? Maybe Byksky or Corsair?

[vf1000ride]

I don't use a filter.  Should be ok now that everything was captured and cleaned you should be good to go.  Anything small enough to get through the fins before should continue to be small enough and with the number of times it's been apart and back together hopefully all of the large stuff had already made it to the CPU block.

[Mick Naughty]

So you used ek matte tube then didn’t you? 

[Hexenhammer]

11 hours ago, Mick Naughty said:

So you used ek matte tube then didn’t you? 

I think i did,  I got a length of 16/10 EPDM tube from Watercool in Germany and then i got some more, also from their brand just from PPCS in USA and even thou both are EDPM and same brand, the one from Germany was super soft and then one from USA was stiffer, i liked it better its more rubbery.

 

Was I supposed to wash it before?

[MaratM]

I’ve run into the same problem a few years ago. I’ve ended to use an automotive fuel filter (it cost me a couple dollars) in order to get all the sh!t out of the loop. I’ve removed all waterblocks from the loop and connected with use some electrical tape and run the loop with it for a few hours with a spare psu (it is possible to use the main one if you can jump start it). That was enough to get it clean.

the best way was to use some soft tube, but I did not have any

 

 

[bowrilla]

Inlets and outlets are correct? No kinks? What kind of flow sensor are you using? How are the pumps controlled? What is the reading of the pump speeds when the flow (allegedly) drops?

[AnonymousGuy]

2 hours ago, Hexenhammer said:

I think i did,  I got a length of 16/10 EPDM tube from Watercool in Germany and then i got some more, also from their brand just from PPCS in USA and even thou both are EDPM and same brand, the one from Germany was super soft and then one from USA was stiffer, i liked it better its more rubbery.

 

Was I supposed to wash it before?

I had "problems" with that tubing (PPCS watercool brand EPDM) in my guest gaming box.  I don't think it did any clogging but it definitely fogged the loop with a brownish color.  I'm guessing it was more of an oil than a solid based on what I was seeing (maybe to release it from whatever mold they use?).  I didn't leave it in long enough to see what would happen I just swapped it back for Mayhem Ultra Clear which was also easier/softer to work with.

[Mick Naughty]

2 hours ago, Hexenhammer said:

I think i did,  I got a length of 16/10 EPDM tube from Watercool in Germany and then i got some more, also from their brand just from PPCS in USA and even thou both are EDPM and same brand, the one from Germany was super soft and then one from USA was stiffer, i liked it better its more rubbery.

 

Was I supposed to wash it before?

Cant say for sure as I dont know that hose. If you have any left, you can check. Idk if cleaning mine would help enough. I actually run something like a pipe cleaner through mine and get all the residual debris out. Clogged my cpu and gpu blocks in multiple rigs. Kinda concerned it did the same to some rads or they just perform badly. 

[Hexenhammer]

6 hours ago, bowrilla said:

Inlets and outlets are correct? No kinks? What kind of flow sensor are you using? How are the pumps controlled? What is the reading of the pump speeds when the flow (allegedly) drops?

 

Yep all Inlets and outlets are fine.

Flow sensor is my eye looking a the flow power in the transparent Bykski CP-DDC-X-TK220-V2 reservoir/pump combo [I swapped the Chinese pump with 18W LAING]  and their latest version of flow meter: Bykski B-TFC-CS-X Digital Flow and Temp Meter, its small OLED that shows 3 parameters: Temps, liters per minute and how fast it revs

Its a new build so i always just boot to bios to test the system, my DDC has 3 pin connection to 12v PUMP connector on the x570 MSI Creation board and by default it has control signal disabled, always run 4300RPM, the Barrow D5 [Chinese pump but its more of a helper to my DDC so its ok to have one] is Molex powered with PWM signal to FAN connector and i disabled the PWM control in the bios, so it always runs about 4800RPM

When the flow drops the pump speeds stay steady unchanged so its not the pump, i found dirt stuck in the CPU waterblock fins, it has micro fins that i cant even put a razor blade between them, so it collects macro particles, looks like it got dust from EDPM tubes.

I also took out all the water and run twice trough "filter", i took a clean wet kitchen fabric towel and put over a jar and run the water trough it.

First time the towel had black dots like head of the needle size and some sparkling shiny dots, maybe from the CPU/GPU block coating

 

Im yet to assemble everything back.

 

What do you think?

 

 

[Hexenhammer]

5 hours ago, Mick Naughty said:

Cant say for sure as I dont know that hose. If you have any left, you can check. Idk if cleaning mine would help enough. I actually run something like a pipe cleaner through mine and get all the residual debris out. Clogged my cpu and gpu blocks in multiple rigs. Kinda concerned it did the same to some rads or they just perform badly. 

I have water filter ill be installing it on the CPU block input this time, if nothing is clogged with 2 pumps i wont feel it

[Hexenhammer]

6 hours ago, MaratM said:

I’ve run into the same problem a few years ago. I’ve ended to use an automotive fuel filter (it cost me a couple dollars) in order to get all the sh!t out of the loop. I’ve removed all waterblocks from the loop and connected with use some electrical tape and run the loop with it for a few hours with a spare psu (it is possible to use the main one if you can jump start it). That was enough to get it clean.

the best way was to use some soft tube, but I did not have any

 

 

 

I filtered the water trough wet clean single use fabric towel, put it over a jar and run trough it, it had tons of micro debri and silvery sparkling dots from the GPU/CPU block, the thing they cover the copper with.

 

The big radiator GTX360 i connected directly to hot tap water and run trough it, same thing i did with GPU

CPU block i dissembled and scarbed

 

Now i have GTS480 and the reservoir to clean

 

I hope after cleaning it up and using same parts it wont clog the system, but i have water cooling filter and ill install it on CPU in since it has the smallest fins

 

[bowrilla]

10 hours ago, Hexenhammer said:

Flow sensor is my eye looking a the flow power in the transparent Bykski CP-DDC-X-TK220-V2 reservoir/pump combo [I swapped the Chinese pump with 18W LAING]  and their latest version of flow meter: Bykski B-TFC-CS-X Digital Flow and Temp Meter, its small OLED that shows 3 parameters: Temps, liters per minute and how fast it revs

Probably the same sensor internals  with the same software/firmware. Could very well be a software bug creating false readings.

 

10 hours ago, Hexenhammer said:

Its a new build so i always just boot to bios to test the system, my DDC has 3 pin connection to 12v PUMP connector on the x570 MSI Creation board and by default it has control signal disabled, always run 4300RPM, the Barrow D5 [Chinese pump but its more of a helper to my DDC so its ok to have one] is Molex powered with PWM signal to FAN connector and i disabled the PWM control in the bios, so it always runs about 4800RPM

Monitor the pump speeds with HWInfo64. It uses the onboard sensors to read TACH signals. If the pumps stay at their speed it's almost impossible for the flow to just drop that far down. If it was debris/particles clogging up the loop, it would not start with a regular flow rate and then break down.

 

10 hours ago, Hexenhammer said:

When the flow drops the pump speeds stay steady unchanged so its not the pump, i found dirt stuck in the CPU waterblock fins, it has micro fins that i cant even put a razor blade between them, so it collects macro particles, looks like it got dust from EDPM tubes.

A soft toothbrush and soap is all you need to clean the microfins. I can't imagine there to be so much dust that it totally clogs everything up. Seriously, I've ran opaque fluids without a change for over a year once and the clogging was not severe. I mean, you can try ripping everything apart, flushing every piece of tubing with water and then distilled water (or just distilled water) and rip apart the blocks and the distroplate. If it still happens it's definitely not clogging.

 

10 hours ago, Hexenhammer said:

I also took out all the water and run twice trough "filter", i took a clean wet kitchen fabric towel and put over a jar and run the water trough it.

First time the towel had black dots like head of the needle size and some sparkling shiny dots, maybe from the CPU/GPU block coating

Probably from the radiators. Did you prep them before buiding your loop? A few tiny particles cannot make the loop come to a virtual standstill after a few minutes. If that was the problem, your loop would be clogged permanently.

 

 

The only reliable and accurate sensors I know of are the ones from aquacomputer. The rest is more flash toy stuff (sorry to be blunt here). At best they are reasonably accurate at a specific flow rate and sometimes come up with a huge error well in the double digit percent range. If the software/firmware has a bug, the reading might not represent the actual flow rate. Could also be a shot BIOS or broken mainboard giving unstable voltage. Or it could be a software interfering with fan/pump control.

 

Do you have any thermal issues? With flow coming to almost a standstill there should be a noticeable heat gradient in your loop. With 2 temp probes you could measure this. Radiator inlet and outlet can have potentially a ΔT of ~10K with a very low flow rate while it will probably be more around 0.5-2K (ish) with 0.5-1GPM of flow. Coolant temps before the radiator(s) will also rise noticeably. On average the loop temps will not be very different, but the gradient is a lot bigger.

 

There's no way for the flow rate to suddently totally collaps and the pumps running at the same speed they did before unless something else changes - as long as the pumps aren't sucking in any air (which would be audible and the revs would most likely go up due to less resistance) and nothing kinks under heat (unlikely since the kink would stay).

 

Potential points of issue:

1. F-ed up sensor firmware
2. F-ed up BIOS/UEFI that messes up the revs of fans and pumps
3. Damaged mainboard 
4. Damaged PSU (the 12V rail might have an issue)
5. Some fan control software in Windows messing up the settings

You will need to make sure the pump stays at the right revs. Again, use HWInfo64 to make sure they stay there.

Next, monitor the temps, ideally adding 2 more sensors (ideally in front of and after the first radiator).

[Hexenhammer]

6 hours ago, bowrilla said:

Probably the same sensor internals  with the same software/firmware. Could very well be a software bug creating false readings.

 

Monitor the pump speeds with HWInfo64. It uses the onboard sensors to read TACH signals. If the pumps stay at their speed it's almost impossible for the flow to just drop that far down. If it was debris/particles clogging up the loop, it would not start with a regular flow rate and then break down.

 

A soft toothbrush and soap is all you need to clean the microfins. I can't imagine there to be so much dust that it totally clogs everything up. Seriously, I've ran opaque fluids without a change for over a year once and the clogging was not severe. I mean, you can try ripping everything apart, flushing every piece of tubing with water and then distilled water (or just distilled water) and rip apart the blocks and the distroplate. If it still happens it's definitely not clogging.

 

Probably from the radiators. Did you prep them before buiding your loop? A few tiny particles cannot make the loop come to a virtual standstill after a few minutes. If that was the problem, your loop would be clogged permanently.

 

 

The only reliable and accurate sensors I know of are the ones from aquacomputer. The rest is more flash toy stuff (sorry to be blunt here). At best they are reasonably accurate at a specific flow rate and sometimes come up with a huge error well in the double digit percent range. If the software/firmware has a bug, the reading might not represent the actual flow rate. Could also be a shot BIOS or broken mainboard giving unstable voltage. Or it could be a software interfering with fan/pump control.

 

Do you have any thermal issues? With flow coming to almost a standstill there should be a noticeable heat gradient in your loop. With 2 temp probes you could measure this. Radiator inlet and outlet can have potentially a ΔT of ~10K with a very low flow rate while it will probably be more around 0.5-2K (ish) with 0.5-1GPM of flow. Coolant temps before the radiator(s) will also rise noticeably. On average the loop temps will not be very different, but the gradient is a lot bigger.

 

There's no way for the flow rate to suddently totally collaps and the pumps running at the same speed they did before unless something else changes - as long as the pumps aren't sucking in any air (which would be audible and the revs would most likely go up due to less resistance) and nothing kinks under heat (unlikely since the kink would stay).

 

Potential points of issue:

1. F-ed up sensor firmware
2. F-ed up BIOS/UEFI that messes up the revs of fans and pumps
3. Damaged mainboard 
4. Damaged PSU (the 12V rail might have an issue)
5. Some fan control software in Windows messing up the settings

You will need to make sure the pump stays at the right revs. Again, use HWInfo64 to make sure they stay there.

Next, monitor the temps, ideally adding 2 more sensors (ideally in front of and after the first radiator).

 

Its new build all new parts, i dont even use windows and boot into BIOS to use build in hardware monitor that controls pumps, fans and shows temps.

They Pumps run at constant speed [well you know how it is, its not locked but goes 50rpm up down].

PSU is HX1200i

 

The flow never came to a stand still, once after settling down and all air came out it was 0.4L/m but usually its 1.1 L/m for both pumps but starts much faster as far as 4L/m

 

I took everything to clean so ill see if it stops, ill use water filter on CPU block inlet to be 100% sure its not some debry this time

 

BTW, can you take a look at this post, what do you think?

 

 

 

[bowrilla]

2 minutes ago, Hexenhammer said:

 

Its new build all new parts, i dont even use windows and boot into BIOS to use build in hardware monitor that controls pumps, fans and shows temps.

They Pumps run at constant speed [well you know how it is, its not locked but goes 50rpm up down].

PSU is HX1200i

 

The flow never came to a stand still, once after settling down and all air came out it was 0.4L/m but usually its 1.1 L/m for both pumps but starts much faster as far as 4L/m

 

I took everything to clean so ill see if it stops, ill use water filter on CPU block inlet to be 100% sure its not some debry this time

 

BTW, can you take a look at this post, what do you think?

 

 

 

You've got your answer already in that thread.

 

I still think you need to verify those readings from the Bykski sensors. 

[Hexenhammer]

2 minutes ago, bowrilla said:

You've got your answer already in that thread.

 

I still think you need to verify those readings from the Bykski sensors. 

So you also think its OK to use it that way and it wont choke the pump? I mean the flow rate will be pushing on the same flow that pump uses for intake.

 

About the bykski sensors, i confirmed everything visually, you see my reservoir its transparent, so when flow is fast you see it hitting the opposite side of the pump and jumping back to left side and you can hear it.

When it goes down, it doesn't hit the opposite site of the res, and the slowest the flow the farthest its from the right-side wall, at slowest point it flows over the left reservoir wall just below the inlet

[bowrilla]

Just now, Hexenhammer said:

So you also think its OK to use it that way and it wont choke the pump? I mean the flow rate will be pushing on the same flow that pump uses for intake.

 

About the bykski sensors, i confirmed everything visually, you see my reservoir its transparent, so when flow is fast you see it hitting the opposite side of the pump and jumping back to left side and you can hear it.

When it goes down, it doesn't hit the opposite site of the res, and the slowest the flow the farthest its from the right-side wall, at slowest point it flows over the left reservoir wall just below the inlet

Could be either trapped air being pushed out or pump speed dropping - or the pump sucking in air. At a given rpm and constant restriction a pump will always have the same flow rate as long as it is not sucking in air or the fluid changes its properties dramatically. One of those factors needs to change.

 

You need to check those aspects. I've listed several points of failure. Now it's on you to find the culprit.

 

And yes, it's fine.

[Hexenhammer]

2 minutes ago, bowrilla said:

Could be either trapped air being pushed out or pump speed dropping - or the pump sucking in air. At a given rpm and constant restriction a pump will always have the same flow rate as long as it is not sucking in air or the fluid changes its properties dramatically. One of those factors needs to change.

 

You need to check those aspects. I've listed several points of failure. Now it's on you to find the culprit.

 

And yes, it's fine.

Can you give example how pump can suck air but no leaks, can this be? I also had similar idea but then everything is clean, i colored my liquid red just for that reason to easily see leaks

[bowrilla]

1 minute ago, Hexenhammer said:

Can you give example how pump can suck air but no leaks, can this be? I also had similar idea but then everything is clean, i colored my liquid red just for that reason to easily see leaks

Too much turbulences and bubbles.

[Hexenhammer]

3 hours ago, bowrilla said:

Too much turbulences and bubbles.

Hey, now that everything disassembled and cleaned, should I install the GTS480 ports up? It will be uncomfortable and hard to connect the tubes thou

Ports down is the easiest way and connection from radiator-out top res-in is close to each other.

[For Science!]

20 minutes ago, Hexenhammer said:

Hey, now that everything disassembled and cleaned, should I install the GTS480 ports up? It will be uncomfortable and hard to connect the tubes thou

Ports down is the easiest way and connection from radiator-out top res-in is close to each other.

Ports up = easier to bleed, harder to drain

ports down = harder to bleed, easier to drain

 

pick your poison.

 

with respect to the thread, you really need to start posting photos and videos of the issue, there is a lot left to the imagination at this point.

[Hexenhammer]

35 minutes ago, For Science! said:

Ports up = easier to bleed, harder to drain

ports down = harder to bleed, easier to drain

 

pick your poison.

 

with respect to the thread, you really need to start posting photos and videos of the issue, there is a lot left to the imagination at this point.

When ill start to assemble everything back ill post pics for troubleshooting