So it begins! (3960x + custom liquid cooling)

[101m4n]

I last replaced my rig in 2015. It was a water-cooled ncase-m1 build with a 4790k and a gtx970. She was a beautiful machine, but she's getting a little long in the tooth now.
Anyway, about a week ago, I found a good deal on a used 3960x and trx40 motherboard. I decided to place a bid, and if I got the parts at a decent price I'd go forward with the build.

I'm here so you can probably guess what happened. Ended up getting the board and CPU for £1100!

 

The planned specs are as follows:

• 3960x
• Gigabyte trx40 aorus master
• 64GB 3600mhz memory
• 500GB sabrent rocket pcie4.0 ssd
• Fractal design s2 meshify (because I did another build here on the forum in a fractal case and I thought it was pretty solid)
• MSI gtx 1080 gaming-X (as I already have it, and I don't really need anything faster)
• Custom loop! (because building a PC without one is no fun at all)

The plan is to update the build as it unfolds over the next few days, adding posts to the thread and updating the OP as things progress.

 

[Sunday 20/09/2020]

I have some of the parts in hand already and the rest should arrive tomorrow.

I don't actually have any ddr4 on hand (all my systems are too old for that) so I won't know until tomorrow if the secondhand CPU/Motherboard I have been sold are any good! 

 

[Later that same evening...]

Well that didn't take long.

The first bit of drama I've run into is that the CPU came without the orange carrier tray or shmansy AMD torque wrench. This is all I got:

Quick google reveals that it's a t20 driver with a torque of 1.5 newton-metres. It looks like the carrier frame isn't actually load-bearing, and is just there to make inserting/removing the processor safer (less likely to accidentally break the socket). Also the CPU has notches in it, like any other processor, which engage with the socket when it's inserted. So it looks like the carrier frame isn't responsible for aligning the pins either.

I managed to get it installed with this thing, which I originally bought for installing disk breaks on my mountain bike:

It's pretty crude, but at least I know all the screws should be roughly the same tightness (and at least somewhere close to 1.5NM).

The board and CPU after a little cleaning:

On another note, the motherboard doesn't seem to have been very well taken care of. There are some missing screws, three for the backplate and another for the first m.2 slot. Hopefully I can give it a better home that it had previously, tsk tsk.

That being said, I'm even more apprehensive about turning the thing on now...

 

[Monday 21/09/2020]

Did a post test (without cooling), and all was well. A weight off my mind for sure.
Time to get this show on the road:

 

P.S. Thanks ocuk, I love it when my harribo comes with free watercooling parts 

 

So the radiators came with two sets of longs screws, some too long for the fans+panel, and some that were too short.

I don't think these radiators are intended to be installed like this (mounted to a panel through the fans). But I liked the way it looks, and I managed to make it work in the end:

This also gives the top fans a little more clearance, which can only be a good thing.

The current status of the build:

 

[Tuesday 21/09/2020]

I was up working up until about 3 in the morning last night working on the build, this is how it looks now:

Unfortunately, it doesn't hold air. Worse still, it's a relatively slow leak which means I'm going to have to hunt for it a bit...

First things to check as usual, rotary fittings.

 

I had to use different tubing for the fill port connection as the black stuff wouldn't bend enough:

Oh well, still looks fairly nice

 

[Tuesday evening]

So I found the leak, it was the right angle fitting at the bottom of the pump. I replaced it with a clear tube and put the ball valve at the end:

And on the back:

The tube is old and a little manky, but it's out of sight so I'm not too bothered. There aren't any clamps, but these barbs are actually meant for larger tubes, so they'll hold. Certainly took some doing to get them on there.

Time for a leak test:

Wish me luck 

 

[A bit later]

Disaster has struck.

 

The filling of the loop went off without a hitch, no problems there. But while I was bleeding it, I got the bright idea to connect the top fans to the external power supply I was using to power the pump. No particular reason, just wanted to see them spin, maybe get a feel for the amount of airflow. The fan controller though was connected to the motherboard with a 12v and ground via a system fan header, so I can only assume the board saw 12v and a path to ground and promptly tried to start up. There was a pop, some magic smoke escaped from somewhere near the system fan header, and that was that.

I have been building systems for almost ten years and I have made plenty of mistakes, but this is the first time I've killed a motherboard . As this is my fault, I consider myself morally obliged to eat the cost of a replacement board.

All in all, a bad day.

I have tested as many other components as I can, PSU, memory, SSD, and they are all fine which is a good sign. The GPU wasn't installed at the time either so that's fine too. But the big ticket item is the CPU. I have no way to test it, and if it's dead, then my day will get a lot worse.

 

Moral of the story:
Be very careful when using an external power supply to power components in your PC. Make sure that any components you connect it to are completely disconnected from the rest of the system. If they are not, you are in for a bad time.

 

[Wednesday 23/09/2020]

I have taken it out of the system now and found the borked chip:

The serial number is 3947s934gb, made by nuvoton. There are several of them on the board, they appear just to be fan controllers.

If that's the only damage, it might be repairable 

 

[Thursday 24/09/2020]

New board has arrived. This evening I'll get it installed, hopefully I can polish off the build without killing this one.

Even at the cost of a new motherboard, I'm actually still £150 below retail price on the motherboard and processor. Fingers crossed the CPU isn't damaged.

 

[Later that evening]

The system works now, but there are more issues. 

 

The first is that the fans have very little clearance on their front face and scrape against the metal on the radiator and case at higher RPMs. So I'm going to have to remount them with the spacers on the other side. I have done this for the front radiator, but the top radiator will be a little more difficult. I may have to drill the threads out of the nuts I was using as spacers.

 

The other issue is the GPU. The machine simply refuses to post with it installed.

I have tried other GPUs (an rx550, and a gtx680) and they work fine.

I have tried the 1080 in other systems, and it works fine. I even tried it in an am2+ board from 2009 and it worked perfectly!

I have measured the pcie power cables with a voltmeter to check that the card is getting power. It is.

I have tried updating the motherboard bios.

I tried the GPU with every past bios that has ever existed for it. No luck.

This is actually a problem that I have had before with this card. Another board a couple years ago, an (also gigabyte) gaz68xp-ud3p also refused to post with the GPU installed. I'm beginning to think this is some sort of weird bios bug that only presents itself with MSI GPUs and gigabyte motherboards. It could also explain why there are so many MSI gaming X cards from this generation on ebay. And why they seem to be cheaper than the others.

Whatever it is, it must be something pretty fundamental to have spanned 9 years of hardware. To top it all off, the post code I get with the card installed (0D) isn't even in the manual! Just fantastic.

 

[Friday 25/09/2020]

So I picked up a replacement card on ebay and it seems to be working properly. The plan is to sell the old one to cover the cost.

The evga card works.

 

I have seen many reports of 10 series cards not posting in various motherboards and a disproportionate number of those issues seem to be with MSI cards. I pretty sure they just guffed up the bios on these cards somehow. Oh well, I suppose I'll just include a note and a return postage label with the card when I sell it.

 

Here's the system now:

Not without some more issues though (why not?).

One of the things I wanted was fan control via coolant temperature. Fortunately the board supports this with a couple of headers for 10k thermistors. For reasons that defy all logic and sense though, certain fans can only be controlled by certain sensors. The sensor that suited my needs however, was sandwiched between the edge of a usb2 header and the bottom-most PCIe slot.

I ended up having to solder some new leads on that would fit in the gap:

Can't say I was thrilled about soldering this close to by far the most valuable computer I have ever owned, but hey ho. It had to be done. I'm sure the cardboard will protect it from the 250 centigrade molten metal .

Time to get an OS installed :

[nick name]

Cool rug.  Where'd you get that rug?

[101m4n]

Just now, nick name said:

Cool rug.  Where'd you get that rug?

Here you go 

[101m4n]

Well that didn't take long.

The first bit of drama I've run into is that the CPU came without the orange carrier tray or shmansy AMD torque wrench. This is all I got:

Quick google reveals that it's a t20 driver with a torque of 1.5 newton-metres. It looks like the carrier frame isn't actually load-bearing, and is just there to make inserting/removing the processor safer (less likely to accidentally break the socket). Also the CPU has notches in it, like any other processor, which engage with the socket when it's inserted. So it looks like the carrier frame isn't responsible for aligning the pins either.

I managed to get it installed with this thing, which I originally bought for installing disk breaks on my mountain bike:

It's pretty crude, but at least I know all the screws should be roughly the same tightness (and at least somewhere close to 1.5NM).

The board and CPU after a little cleaning:

On another note, the motherboard doesn't seem to have been very well taken care of. There are some missing screws, three for the backplate and another for the first m.2 slot. Hopefully I can give it a better home that it had previously, tsk tsk.

That being said, I'm even more apprehensive about turning the thing on now...

[101m4n]

Did a post test (without cooling), and all was well. A weight off my mind for sure.
Time to get this show on the road:

 

P.S. Thanks ocuk, I love it when my harribo comes with free watercooling parts  

[101m4n]

So the radiators came with two sets of longs screws, some too long for the fans+panel, and some that were too short.

I don't think these radiators are intended to be installed like this (mounted to a panel through the fans). But I liked the way it looks, and I managed to make it work in the end:

This also gives the top fans a little more clearance, which can only be a good thing.

The current status of the build:

 

[101m4n]

I was up working up until about 3 in the morning last night working on the build, this is how it looks now:

Unfortunately, it doesn't hold air. Worse still, it's a relatively slow leak which means I'm going to have to hunt for it a bit...

First things to check as usual, rotary fittings.

 

I also had to use different tubing for the fill port connection as the black stuff wouldn't bend enough:

Oh well, still looks fairly nice 

[101m4n]

So I found the leak, it was the right angle fitting at the bottom of the pump. I replaced it with a clear tube and put the ball valve at the end:

And on the back:

The tube is old and a little manky, but it's out of sight so I'm not too bothered. There aren't any clamps, but these barbs are actually meant for larger tubes, so they'll hold. Certainly took some doing to get them on there.

Time for a leak test:

Wish me luck 

[101m4n]

Disaster has struck.

 

The filling of the loop went off without a hitch, no problems there. But while I was bleeding it, I got the bright idea to connect the top fans to the external power supply I was using to power the pump. No particular reason, just wanted to see them spin, maybe get a feel for the amount of airflow. The fan controller though was connected to the motherboard with a 12v and ground via a system fan header, so I can only assume the board saw 12v and a path to ground and promptly tried to start up. There was a pop, some magic smoke escaped from somewhere near the system fan header, and that was that.

I have been building systems for almost ten years and I have made plenty of mistakes, but this is the first time I've killed a motherboard . As this is my fault, I consider myself morally obliged to eat the cost of a replacement board.

All in all, a bad day.

I have tested as many other components as I can, PSU, memory, SSD, and they are all fine which is a good sign. The GPU wasn't installed at the time either so that's fine too. But the big ticket item is the CPU. I have no way to test it, and if it's dead, then my day will get a lot worse.

 

Moral of the story:
Be very careful when using an external power supply to power components in your PC. Make sure that any components you connect it to are completely disconnected from the rest of the system. If they are not, you are in for a bad time.

[HM-2]

Best of luck with the new motherboard. Processors tend to be pretty robust so hopefully there's no damage there.

 

Love the use of EK-ZMT too, my ongoing build is going the same way. Something very industrial about it.

[101m4n]

31 minutes ago, HM-2 said:

Love the use of EK-ZMT too, my ongoing build is going the same way. Something very industrial about it.

Yup, it's solid stuff.

On an unrelated note, I actually distilled the coolant myself while I was waiting for parts:

Certainly not very practical, and I burned myself a few times, but good fun

[For Science!]

1 hour ago, 101m4n said:

Yup, it's solid stuff.

On an unrelated note, I actually distilled the coolant myself while I was waiting for parts:

Certainly not very practical, and I burned myself a few times, but good fun

Surely its more expensive than just buying it: https://www.overclockers.co.uk/mayhems-ultra-pure-h2o-watercooling-cleaning-fluid-1-litre-wc-12r-mh.html

 

I guess English tap water is also so hard that it should leave a nasty residue back in the pot  Maybe you can confirm with a TDS probe how distilled the water actually is.

 

The whole setup looks quite nice, reminds me of a the giant hanging incense in China.

[101m4n]

3 hours ago, For Science! said:

Maybe you can confirm with a TDS probe how distilled the water actually is.

Unfortunately I don't have one. I did attempt to measure conductivity with a plain multi-meter with probes set a fixed distance apart. The resistance is about double tap water, ish. I say ish because the measurement isn't very stable. It's certainly higher than industrially de-ionised water, but I wouldn't expect it to be as low anyway because the condenser was made of copper. It should be clean enough for my purposes though.

3 hours ago, For Science! said:

Surely its more expensive than just buying it: https://www.overclockers.co.uk/mayhems-ultra-pure-h2o-watercooling-cleaning-fluid-1-litre-wc-12r-mh.html

 

I guess English tap water is also so hard that it should leave a nasty residue back in the pot

Yup:

This was towards the beginning as well, it was much worse than this at the end.

3 hours ago, For Science! said:

Surely its more expensive than just buying it

We'll see when the gas bill comes in! I distilled about three and a half litres, it took about 8 hours total.

[101m4n]

36 minutes ago, Grumpy Old Man said:

C₂H₅OH

Afraid not

[Analog]

It's sad for the motherboard mate. I understand that you killed it and you feel obliged to bite the bullet and pay retail price for another motherboard, but IMO you should still try to get an RMA for it. You know stuff just happens sometimes and who knows the motherboard wasn't DOA in the first place. I have killed a  graphics card in the past from flashing a custom BIOS on it and pushed the clocks so high the VRM gave way. I RMAed it and it was honored. Didn't have to spend anything additionally. 

[101m4n]

Just now, Analog said:

It's sad for the motherboard mate. I understand that you killed it and you feel obliged to bite the bullet and pay retail price for another motherboard, but IMO you should still try to get an RMA for it. You know stuff just happens sometimes and who knows the motherboard wasn't DOA in the first place. I have killed a  graphics card in the past from flashing a custom BIOS on it and pushed the clocks so high the VRM gave way. I RMAed it and it was honored. Didn't have to spend anything additionally. 

Can't RMA. Gigabyte only does RMA through authorised retailers. It's not done by serial number like for example ASUS does. Also, as it's second hand I don't have proof of purchase.

I have taken it out of the system now and found the borked chip:

The serial number is 3947s934gb, made by nuvoton. There are several of them on the board, they appear just to be fan controllers.

If that's the only damage, it might be repairable 

[Analog]

2 minutes ago, 101m4n said:

Can't RMA. Gigabyte only does RMA through authorised retailers. It's not done by serial number like for example ASUS does. Also, as it's second hand I don't have proof of purchase.

I have taken it out of the system now and found the borked chip:

 

The serial number is 3947s934gb, made by nuvoton. There are several of them on the board, they appear just to be fan controllers.

If that's the only damage, it might be repairable 

I get it now. I was under the impression the board was new. yeah, if it is second hand... not much that can be done at this point, especially without proof of purchase. 

 

That chip you took a photo of might be reparable, but at least as far as I know a blown up fan controller at the most results in the fan plugged in that connector no longer running or running at max speed. It shouldn't case a no POST issue. 

[101m4n]

4 minutes ago, Analog said:

I get it now. I was under the impression the board was new. yeah, if it is second hand... not much that can be done at this point, especially without proof of purchase. 

 

That chip you took a photo of might be reparable, but at least as far as I know a blown up fan controller at the most results in the fan plugged in that connector no longer running or running at max speed. It shouldn't case a no POST issue. 

That's what I was thinking as well.

There could be a firmware aspect to this too though. I can imagine some microcontroller running off 5vsb somewhere doing pre-power-on checks, initialising devices on the board etc, then getting stuck at the dead one. Maybe that's just wishful thinking though.

 

edit* it's also possible the dead chip is now a short and that's what's keeping the system from powering on.

[101m4n]

New board has arrived. This evening I'll get it installed, hopefully I can polish off the build without killing this one.

Even at the cost of a new motherboard, I'm actually still £150 below retail price on the motherboard and processor. Fingers crossed the CPU isn't damaged.

[Analog]

1 hour ago, 101m4n said:

New board has arrived. This evening I'll get it installed, hopefully I can polish off the build without killing this one.

 

Even at the cost of a new motherboard, I'm actually still £150 below retail price on the motherboard and processor. Fingers crossed the CPU isn't damaged.

 

I just checked your other build with the copper piping and it looks sick af! Are you planning on doing the same style with this one or are you over the copper looks? 

[101m4n]

27 minutes ago, Analog said:

I just checked your other build with the copper piping and it looks sick af! Are you planning on doing the same style with this one or are you over the copper looks? 

Thanks!

 

No I'm going for an all black build with ek zmt this time around, there are some pictures earlier in the thread.

Did consider doing copper again, but I wanted something more practical (and lower effort ).

Just as well, with the flexible tubes I didn't have to empty the loop to remove the motherboard!

I'm also not watercooling the GPU, as it's an old card from another machine and I wouldn't mind a ray-tracing capable one at some point in the future. If I get a new GPU for it, then I will likely watercool that (which the soft tubes will also make easier). I'll be waiting for AMDs answer to rtx3000 before I make any decisions though.

 

Incidentally, the copper build actually belonged to a friend, he moved out of the country at the beginning of the year and disassembled that system. I bought some of the components from him and they are actually being used here, namely; the pump-res, power supply and a few right angle fittings.

[101m4n]

The system works now, but there are more issues. 

 

The first is that the fans have very little clearance on their front face and scrape against the metal on the radiator and case at higher RPMs. So I'm going to have to remount them with the spacers on the other side. I have done this for the front radiator, but the top radiator will be a little more difficult. I may have to drill the threads out of the nuts I was using as spacers.

 

The other issue is the GPU. The machine simply refuses to post with it installed.

I have tried other GPUs (an rx550, and a gtx680) and they work fine.

I have tried the 1080 in other systems, and it works fine. I even tried it in an am2+ board from 2009 and it worked perfectly!

I have measured the pcie power cables with a voltmeter to check that the card is getting power. It is.

I have tried updating the motherboard bios.

I tried the GPU with every past bios that has ever existed for it. No luck.

This is actually a problem that I have had before with this card. Another board a couple years ago, an (also gigabyte) gaz68xp-ud3p also refused to post with the GPU installed. I'm beginning to think this is some sort of weird bios bug that only presents itself with MSI GPUs and gigabyte motherboards. It could also explain why there are so many MSI gaming X cards from this generation on ebay. And why they seem to be cheaper than the others.

Whatever it is, it must be something pretty fundamental to have spanned 9 years of hardware. To top it all off, the post code I get with the card installed (0D) isn't even in the manual! Just fantastic.

[For Science!]

1 hour ago, 101m4n said:

The system works now, but there are more issues. 

 

The first is that the fans have very little clearance on their front face and scrape against the metal on the radiator and case at higher RPMs. So I'm going to have to remount them with the spacers on the other side. I have done this for the front radiator, but the top radiator will be a little more difficult. I may have to drill the threads out of the nuts I was using as spacers.

 

The other issue is the GPU. The machine simply refuses to post with it installed.

I have tried other GPUs (an rx550, and a gtx680) and they work fine.

I have tried the 1080 in other systems, and it works fine. I even tried it in an am2+ board from 2009 and it worked perfectly!

I have measured the pcie power cables with a voltmeter to check that the card is getting power. It is.

I have tried updating the motherboard bios.

I tried the GPU with every past bios that has ever existed for it. No luck.

This is actually a problem that I have had before with this card. Another board a couple years ago, an (also gigabyte) gaz68xp-ud3p also refused to post with the GPU installed. I'm beginning to think this is some sort of weird bios bug that only presents itself with MSI GPUs and gigabyte motherboards. It could also explain why there are so many MSI gaming X cards from this generation on ebay. And why they seem to be cheaper than the others.

Whatever it is, it must be something pretty fundamental to have spanned 9 years of hardware. To top it all off, the post code I get with the card installed (0D) isn't even in the manual! Just fantastic.

With TR and 10-series GPU I had this funky issue that it didn't like the particular display port cable I was using (DP-to-miniDP) getting a full fat DP-to-DP or a HDMI allowed the machine to post properly. I also remember another person who had this issue on this forum that I helped out in the past. Have a go with a different monitor cable.

 

Edit: found it

 

[101m4n]

34 minutes ago, For Science! said:

With TR and 10-series GPU I had this funky issue that it didn't like the particular display port cable I was using.

Hmm.

I have tried several displays, a vga via an active vga to dvi adapter and a "full fat" display port to my displayport monitor. As for hdmi, I don't actually have an hdmi monitor or cable! I suppose I could go down the street and get one, then try plugging it into the TV... But the GPU would still be useless in normal use. I also don't think this will solve the problem. The system doesn't detect the card at all. In the bios, the PCIE slot registers as empty.

I am currently hunting for other 1080s in the area that I can go pick up this evening. Then I can sell this card on ebay to recoup some of the cost. It's a shame, I really like this card. Too bad both customer service for MSI and gigabyte are a bag of dicks.

[For Science!]

3 minutes ago, 101m4n said:

The system doesn't detect the card at all. In the bios, the PCIE slot registers as empty.

This sounds to me like you are booting with another card in the primary PCIe slot with the 1080 in a secondary PCIe slot? I wouldn't be surprised if it doesn't register in the BIOS. But if you can get into the BIOS, try disable CSM (compatibility support module). And after doing that, put hte 1080 back into the primary slot to see if anything changes.