Jump to content
Search In
  • More options...
Find results that contain...
Find results in...

Hi all,

 

Apologies in advance for the lengthy post.

 

I'm fairly new to water cooling, so decided to construct my own custom loop just over a month ago for my CPU and GPU. However, according to my digital flow meter, I'm having trouble with a low flow rate (around 50-60 LPH) at 100% pump speed, and believe it's causing some issues with abnormally high CPU and GPU temps.

 

From doing a little digging around online, I should be looking for anything between 0.5-1GPM for maximum efficiency, and I'm a long way away from achieving that at the moment.

 

My i9-9900k CPU has been de-lidded and had liquid metal applied to the die, and a copper ISH installed. It's been overclocked to 5.1Ghz all cores (@1.34v) and my Radeon 6800XT has been overclocked to 2.7Ghz at the moment.

 

The CPU is thermal throttling during OCCT SSE small size and Prime95 non-AVX stress tests with temps of between 90-100 on all cores. Eventually it causes the system to freeze.

On a separate test, the GPU is topping 60 degrees during GPU-only 3D Mark stress tests, and from research, I should be getting anything from 48-55.

 

I've observed the flow myself around the tubing and then dribbles into the reservoir, so I know the flow meter reading is relatively accurate.

 

System Setup - https://ibb.co/QF2dtgh

 

Here are the components used within the loop :

 

Lian Li 011-XL Case

Barrow Distro Block for 011-XL

Barrow DDC Pump (960L/H and 5.2m Head Pressure)

2 x Radiators (1 x EK-XE360 and 1 x EK-PE60)

CPU - EK Supremacy Evo Block

GPU - EK Quantum Velocity Block

 

As you can see from the image above, there are plenty of right-angled fittings (14 in total if you count the connections on the distro-plate) and I've tried to do a few bends myself to cut down on those right-angles, so I understand that there would be some restriction of flow, but would it explain the lowly 50-60LPH I'm currently experiencing in this setup when the pump is rated at 960L/H?

 

For reference, my temps at idle and load are as follows :

 

Ambient - 22 (idle)    25 (load)

Water Temp - 27 (idle)    34 (load)

GPU - 29 (idle)    61 (load)

CPU - 33 (idle)    90-100 (load)

 

The Radiators do get warm to touch on load, and I have set up a fan curve based on the water temps, so I've ruled out any issues with the radiators as they seem to be doing a good job of keeping the water cool based on the above deltas. When running 100% and then 50% pump speed tests, the temperatures increase by around 5 degrees on the CPU.

 

I'm thinking of buying another DDC pump to add in series to the one I currently have and increase the to improve the head pressure, but I'm unsure if the pump I currently have is faulty or not based on the information I've provided.

Any ideas would be much appreciated.

 

Thanks!!

Link to post
Share on other sites

Ddc pumps dosent offer as much head pressure as a D5 pump would, that's why D5 pumps are generally recomended for loops with multiple blocks and radiators. The flow rating and pressure rating is with no restrictions, so the restriction in your radiators, fittings and blocks will lower your flow.

 

Do you have many 90 degree fittings or quick disconnect fittings in your loop? 

Link to post
Share on other sites
1 minute ago, TechJezus said:

Ddc pumps dosent offer as much head pressure as a D5 pump would, that's why D5 pumps are generally recomended for loops with multiple blocks and radiators. The flow rating and pressure rating is with no restrictions, so the restriction in your radiators, fittings and blocks will lower your flow.

 

Do you have many 90 degree fittings or quick disconnect fittings in your loop? 

Including the 90 degree angles at the distro block insert and exit points, I probably have around 14 in total.  No quick disconnect fittings.

 

Aren't DDC pumps usually rated at a higher head pressure with a lower flow rate (<1000LPH) and D5's are the opposite?  The one I have is rated to have 5.2m head pressure with a 960LPH, and some of the 1500 LPH D5's I've been looking at have between 3-4m head pressure.

Link to post
Share on other sites
12 minutes ago, xenosys said:

Including the 90 degree angles at the distro block insert and exit points, I probably have around 14 in total.  No quick disconnect fittings.

 

Aren't DDC pumps usually rated at a higher head pressure with a lower flow rate (<1000LPH) and D5's are the opposite?  The one I have is rated to have 5.2m head pressure with a 960LPH, and some of the 1500 LPH D5's I've been looking at have between 3-4m head pressure.

In general that is correct. But I have seen plenty of ddc pumps not being able to handle two radiators and two blocks, where a single D5 would perform without breaking a sweat. 

My experience is that the gap between a good ddc pump and a bad one is pretty big, where most D5 pumps are almost indentical. 

 

If I were you, I would try to RMA the pump, or at least test it without all the restriction, and see what flow it's producing. Then you would now if it's a flow or pressure problem with the pump. 

Link to post
Share on other sites
21 hours ago, TechJezus said:

In general that is correct. But I have seen plenty of ddc pumps not being able to handle two radiators and two blocks, where a single D5 would perform without breaking a sweat. 

My experience is that the gap between a good ddc pump and a bad one is pretty big, where most D5 pumps are almost indentical. 

 

If I were you, I would try to RMA the pump, or at least test it without all the restriction, and see what flow it's producing. Then you would now if it's a flow or pressure problem with the pump. 

That is surprising. My little SPC pump has no issue with 2 radiators (HWLabs GTS - pretty restrictive in comparison) and 2 blocks. At 100% I get ~170lph / 3/4GPM. 50-60lph is ~25-30% pump speed. The EKWB SPC has a max pressure head of 2.2m and a max flow of 250lph. An EKWB DDC pump has a head pressure of 5.2m and a max flow of 1000lph. I find it very unlikely that a realistic loop is so utterly restrictive that a DDC pump can't push >100lph or 1/2GPM.

 

Since this is one of a few threads regarding low flow rate with Bykski and Barrow components and their built-in sensors with displays, I assume the sensors to be just awfully bad or the software has some issues. Wouldn't be surprised if both Bykski and Barrow basically use/buy the same components and one of their suppliers has just made some mistake.

 

Things to check:

  • Did you use the correct inlet port on the CPU block? (CPU blocks often have jet plates and those are sensitive to flow direction)
  • Do you have a proper mount on the CPU? (all other temps are fine, just the CPU load temps are surprisingly high)
  • Have you tried your CPU before with an air cooler? (this might tell you whether your aftermarket IHS is off spec)

If all that doesn't help: Get a proper flow meter (aquacomputer mps line) and take the readings.

 

P.S.: you can't validate flow visually.

Use the quote function when answering! Mark people directly if you want an answer from them!

Link to post
Share on other sites

Just to throw this into the mix, I am uncertain whether Barrow/Bykski provides genuine DDC/D5 from Xylem/Laing I have a feeling the majority of them are the same form factor but not the real thing.

Link to post
Share on other sites

i have a D5 with 2 rads and 2 blocks and i got plenty of flow. maybe your DDC sucks 😛 

take off the tube that goes from the last area before it goes into the res, turn it on and see what the flow rate really is. i mean gauges and crap are nice and fancy but nothing beats a visual inspection of the flow speed at the slowest point of the loop. Obv have it go into a bucket or a cup or something.

if its not shooting out of there quick as shit then go back one part at a time and do the same and see at that point it speeds up and that will be the thing that's holding up your flow.

Link to post
Share on other sites
On 3/4/2021 at 9:12 AM, TechJezus said:

Ddc pumps dosent offer as much head pressure as a D5 pump would, that's why D5 pumps are generally recomended for loops with multiple blocks and radiators. The flow rating and pressure rating is with no restrictions, so the restriction in your radiators, fittings and blocks will lower your flow.

 

Do you have many 90 degree fittings or quick disconnect fittings in your loop? 

I don't know why you keep saying a D5 pump has more pressure than a DDC.  It does not...DDC has about 50% more head pressure than a D5 does.  The advantages of D5 pumps are higher flowrate at low restriction (which is really irrelevant) and being regarded as more reliable.

 

In any case, a couple blocks and a radiator aren't even close to being "highly restrictive".  I have a D5 pump right now driving 3 CPU blocks, gpu block, radiator, 20 feet of tubing, and a set of quick disconnects.

 

For everyone's education: https://martinsliquidlab.wordpress.com/pump-planning-guide/

 

You'd basically need 15 right angle fittings to equal the restriction of a CPU block...negligible.

 

On 3/4/2021 at 8:50 AM, xenosys said:

The CPU is thermal throttling during OCCT SSE small size and Prime95 non-AVX stress tests with temps of between 90-100 on all cores. Eventually it causes the system to freeze.

If this is instantly happening, you have a bad mount.  Full stop.  Your CPU idling at +6 over ambient is also indicating there's a bad mount, because idle should basically = water temp on a CPU when it's in a lower power state during idle.

Workstation: 10850K @ 5.1Ghz || Gigabyte Z490 Ultra || ASUS TUF 3090 650W shunt, active backplate || G.Skill DDR4-3800 XMP 4x8GB || Corsair AX1500i || 11 gallon whole-house loop.

LANRig/GuestGamingBox: 9900KF @ 4.9Ghz || Gigabyte Z390 Aorus Master || Gigabyte Gaming G1 1080Ti || Corsair SF600 || CPU+GPU watercooled 280 rad push only.

Server Router (Untangle): 8350K @ 4.7Ghz || ASRock Z370 ITX || 2x8GB || PicoPSU 250W, running on AX1200i from Server Storage || 11 gallon whole-house loop.

Server Compute: E5-2696v4 || Asus X99 mATX WS || LSI 9280i + Adaptec + Intel Expander || 5x10TB Seagate Enterprise Raid 6, 4x8TB Seagate Archive Backup, Corsair AX1200i (drives)  || 11 gallon whole-house loop.

Laptop: HP Elitebook 840 G3 (Intel 8350U), Razer Blade Stealth 13" 2017 (Intel 8550U)

Shelved: 9900 nonK || ASRock Z390 Taichi Ultimate || Gigabyte Z270 Gaming 9 || EVGA Titan X (Maxwell) || GTX 780

Link to post
Share on other sites
5 hours ago, AnonymousGuy said:

Your CPU idling at +6 over ambient is also indicating there's a bad mount

While I fully agree with the overall assessment, idling 5-10 over ambient is not uncommon. That solely depends on the CPU. I very much doubt you could keep a Zen2 or Zen3 chip under the mark of 5-10K over ambient on anything else than LN2. And that's just an example of quite a few chips that just run very hot.

Use the quote function when answering! Mark people directly if you want an answer from them!

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×