Seconds from watercooling disaster:

[Lemlurker]

2 days from disaster: my new christmas present of a 3TB hard drive was installed. first off both old 1tb and new 3 in together to copy over then swap back in the temporarily removed 2tb for the one and im at my running capacity. doesnt put on back pannel incase i need access to drives again.

30 minuets from disaster: i was discussing with my boyfriend about what we could do so that we could play more together and decided that maybe we could move his pc into my house on my long ass desk so he could game whilst im on my pc whist over here.

20 minuets from disaster: having collected his pc the space clearing begins. 3d printers moved, under desk cleared, cables managed and tidied. i realise that since my pc is now backing up to open space i should reinstall my back panel i had removed a few days. things are a little stiff around the drive bay cables but it goes on okay.

5 minuets from disaster: i turn on my pc as im setting up the boyfriends to let it boot. seems fine first boot but it keeps turning off after about 5 minuets, each time booting slower and slower, at one point it even tried to disk check. first sign of issues.

30 seconds from disaster: i keep trying to boot the pc, it gets to windows but feels dogshit slow, i wonder if i had disturbed my bootdrive in the reinstall of the backpannel and it was booting off a forgotten windows install from the 2tb drive, it still keeps shiutting down before i can check the c drive. i try different wall sockets incase the cable management has caused an iffy connection

5 seconds from disaster: still booting i catch motion from the corner of my eye... OH SHIT LEAK, slow dripping leak from the inlet port of my EK velocity CPU block., i scramble for power as the life blood of my pc is dripping on my gpu backplate and dribbling around the pcie slot and down the motherboard, not too much fluid but a decent little puddle. i get the pc on its back and begine trying to think how i can drain the loop with a dodgy fitting and trying to figure out wtf caused this loop and why the pc was booting so slow

investigations begin

5 minuets after disaster: i get the side panel off and grab a cloth on hand to mop the fluid off the motherboard and gpu, as i do this i touch the cpu block. its roasting hot, nearly enough to burn me. i get out my cat s61 phone with the thermal and measure the temperature of the block... 72 degrees even after being off for 5 mins. some semblance of the cause occurs to me... the pump mustve unplugged or stopped working causing the cpu to overheat, thermal throttle then shut down but why the leak? was it simple thermal expansion in the scorching hot fittings or something worse? turns out worse was an understatement.

10 minuets after disaster: first up i tried just tightening the fitting it was a little loose, but it didnt help. i finally managed to drain the loop, i remove the inlet fitting to find it horribly constricted, clouded and deformed. same goes for the inlet but to a lesser degree... the fluid in the tubes had gotten hot enough to surpass the glass transition temperature of my PETG tubing (80c) causing it to deform, and deform quickly under the pressure of the compression fitting. luckily the tube had not bowed outwards (due to lack of pump pressure) so i only needed to remove approximately 2cm from the inlet and outlet tubes and i had enough slack to pull this off without clashing with a bend. i then plumb it back in. i dig around in the hard drive bay tofind my pump cable. which had, as i suspected become unplugged. this is a design issue with the EK pump/res combos... they have stupendously short cables... like i have my outlets pointing towards the glass side so the cable comes out towards the back. i want to run the cable through the drive bay at the front bottom. the included cable is barely long enough to cross the width of the unit and reach into the hole. this means that the connector is close to the frame of the case and that any tension on the extension cable will pull the cable from the device plug.... this was the root cause... either way i plug it in and despite having trimmed the pipes fluid still comes out, though not enough to overwhelm the surface tension of the top of the EK block and i mop it up before it goes over my VRMs. i investigate further... i make a little closed loop with one 120mm rad, the two tubes and the cpu block in it, i seal one end and blow in the other and get instant bubbles from around the fitting. i remove the tube and repeat, i get the same. i then remove the fitting from the block and put my thumb over the hole and success! i have a seal, i look into the block and can see a definite dent around the hole from the oring... this means the temperature of the acetal top plate had exceeded 105 degrees... the pressure of the fitting causing the acetal to dimple and let fluid out... so i then re "plump" the oring on the fitting, reinstall and righted carefully with grips and a cloth and retest... and thank fuck... a seal. i rebuild and refil the pc total ordeal after discovering the fault and flicking the power off? about an hour... however all issues are not solved yet... since the event my thermals are through the roof... ive had to remove my 5.1ghz oc on my 9900k and back off thevoltage as a synthetic load was ramming it straight up to 100c. but even now its in the 90s under load. i suspect the thermal paste was dmaged by the extreme heat and so ill be replacing that as soon as some more comes (i used myu last installing this 9900k) anyways, apologies for the wall of text, i hope you enjoy my missfortune and always check your pumps y'all

[Mick Naughty]

I always have a second monitor up that displays system info. So that would have been a dead giveaway. Can also put safety measures in place. If the cpu was that hot all of the fans should have went to 100% if using the correct software, even audible warnings. 

 

[Lemlurker]

1 minute ago, Mick Naughty said:

I always have a second monitor up that displays system info. So that would have been a dead giveaway. Can also put safety measures in place. If the cpu was that hot all of the fans should have went to 100% if using the correct software, even audible warnings. 

 

to prevent fans ramping and dropping and to make up for my lack of fan headers on my motherbaord all my fans are plugged into a commander pro which is temperature sensing from a priobe on the outflow tube of the cpu (kicks up the fans when outflow exceeds about 35c (allows also for zero rpm running when browsing t'internet at night with some tweaking of the fan curves and lights. however i might in stall a probe on the block itself. however there is no way for the commander pro to warn me of the pump not being plugged into it or a poarticular probve getting hot. ideally id set a certain header to turn off the pc if it readz zerom rpm but that isnt a feature and i couldnt get into windows long enough to bring up any hardware info

 

[Mick Naughty]

10 minutes ago, Lemlurker said:

to prevent fans ramping and dropping and to make up for my lack of fan headers on my motherbaord all my fans are plugged into a commander pro which is temperature sensing from a priobe on the outflow tube of the cpu (kicks up the fans when outflow exceeds about 35c (allows also for zero rpm running when browsing t'internet at night with some tweaking of the fan curves and lights. however i might in stall a probe on the block itself. however there is no way for the commander pro to warn me of the pump not being plugged into it or a poarticular probve getting hot. ideally id set a certain header to turn off the pc if it readz zerom rpm but that isnt a feature and i couldnt get into windows long enough to bring up any hardware info

 

Guess that’s a fault there. My fans don’t ramp up or down. Constant 30%. But I simply check the box for them to go to 100% at 75c on the cpu. Which is impossible for it to hit. But a nice feature to include on the software. 
Mobo software should still have a way to display alerts. 

[Lemlurker]

5 minutes ago, Mick Naughty said:

Guess that’s a fault there. My fans don’t ramp up or down. Constant 30%. But I simply check the box for them to go to 100% at 75c on the cpu. Which is impossible for it to hit. But a nice feature to include on the software. 
Mobo software should still have a way to display alerts. 

used to be when things werent connected to cpu fan header it shat the bed but since aios and similar that featuee increasingly died... id like corsair to have emergency elert settings should a marked fan read zero rpm when told to be higher and ideally some what of the box to error code or prevent posting should it detect no rpm

 

[Blai5e]

2 hours ago, Lemlurker said:

used to be when things werent connected to cpu fan header it shat the bed but since aios and similar that featuee increasingly died... id like corsair to have emergency elert settings should a marked fan read zero rpm when told to be higher and ideally some what of the box to error code or prevent posting should it detect no rpm

 

Well, you said the coolant temperature climbed sky high. You can use ICUE's "Notifications" section for your Commander Pro to add an entry to run all your fans to 100% when a certain temperature threshold is reached. I assume you're probe is capable of reading your coolant so... 45°C or 50°C - set fans to 100% (and/or a choice of 3 other actions).

[Lemlurker]

3 minutes ago, Blai5e said:

Well, you said the coolant temperature climbed sky high. You can use ICUE's "Notifications" section for your Commander Pro to add an entry to run all your fans to 100% when a certain temperature threshold is reached. I assume you're probe is capable of reading your coolant so... 45°C or 50°C - set fans to 100% (and/or a choice of 3 other actions).

requires coolant flow to be able to read it... wasnt hot more than 3 cm down the line. my probe is about 15cm from the block, itd never read as hot

its moot anyways as nmost of the time it barely got into windows 

 

[NotSoEpicMods]

I hate to say it but I hope this serves as a lesson to double check everything after messing around in a pc case. Even if it's just to replace a hard drive or fan you never know what you might nudge or unplug. Anyways maybe take a look at the Phobya flow through temp sensors(https://www.dazmode.com/shop/watercooling/fittings/specialized-fittings/sensor-fittings/temperature-sensor-mf-g14-matte-black-sleeved/). I'm using 2 of them in my Cosmos Sv2 build and one of them is sitting on the outlet of my EK block so basically no distance to the block and they work with the commander pro.