The Z Build 2.0

A brief update.

I am dead in the water on the fan front.

Noctua:

Contacted Noctua Support on March 31st. Heard back from them on April 3rd, when they requested batch numbers from all the fans. I provided those batch numbers on the same day. Have heard nothing yet. It is unclear to me if Noctua support is even open during the pandemic, or if they are just down to limited staff, slowing things down.

Arctic:
So, as I mentioned above, as a backup in case I am not successful with Noctua I ordered a full complement of Arctic P12 (120mm) and P14 (140mm) fans. They are reported to still function per the original Intel PWM spec in the Aquacomputer forums. I need 7x 140mm fans and 16x 120mm fans. All 7x 140mm fans have arrived. One of the 120mm fans have arrived. The remaining 15 are coming in three 5 fan bulk packs which were ordered on March 31st, and still have not shipped...

1586367913778.png


The arrival estimate said April 3rd to April 16th when I ordered them, so I upgraded to two day shipping and assumed, since I did not get the warning that they were going to be delayed by more essential orders, that they would go out quickly and be here sooner than that, but here we are 8 days later and they still have not shipped.

I guess this project is just destined to never get done.
 
Alright.

I received the fans and the replacement Aquaero, and have done some work.

I have lots of updates and pictures to post, and I will, but first I want to talk about a weird issue I have been having.

When I first test filled and leak tested the loop with distilled water, I was seeing great flow rates. With one pump at 100%, I was at 1.1 GPM. With both pumps at 100% it went up to 1.6GPM.

As readers of this thread may recall, then I ran into my Noctua fan problem, which had me troubleshooting, disconnecting and reconnecting fans a lot. At some point there the 4th PWM port on my Aquaero stopped working. I'm guessing during my troubleshooting it was ESD damage. So I ordered the replacement Aquaero and the fans and waited.

Last night I finally got everything together, and filled the loop up with my real coolant, EK Cryofuel.

I hooked up the Aquaero to start configuring fan and pump speed curves and noticed something really strange. Instead of the flow rates I previously had of one pump at 1.1 and two pumps at 1.6, with both pumps at fulls speed I was only seeing 0.9-1.1 GPM flow rates. Leaving the system on for a while these flow rates slowly decline until they drop to 0.6GPM and stay there. Funnily enough, cutting power and starting it again, starts everything back at 0.9-1.0 GPM again, but then it gradually declines again.

According to the RPM signal readouts, both the pumps are running at their specified max speeds of ~4800rpom, so the pumps seem OK.

Also, during the fill, everything seemed normal, with coolant levels dropping and air bubbles coming out at about the same rate I am used to during filling.

Everything looked normal.

I know propylene glycol based coolants like the Cryofuel are a little bit more viscous than distilled water, and could theoretically result in reduced flow, but they are still 90% distilled water, so that impact seems like it would be minor if even measurable at all.

Based on this, I am guessing this is a flow meter readout issue, not an actual flow issue, but I have no good way of verifying that fact.

Since my old Aquaero still works fine except for the dead 4th PWM port, I decided to test if the new Aquaero was to blame, by hooking it up to the flowmeter and looking at the measurements. The flow rate measures exactly the same between the two Aquaeros.

Based on this, my remaining theories are:
- Whatever damaged my old aquaero (ESD?) also damaged my flowmeter, and it is no longer providing accurate readings.
- An air bubble is stuck in the flowmeter causing it to provide inaccurate readings (Seems unlikely that this would happen with the coolant with a surfactant, and not with the distilled water)
- Air is stuck elsewhere in the loop causing actual flow problems. (Again, seems unlikely that this would happen with the coolant with a surfactant, and not with the distilled water)
- Something got in my loop, got stuck in the flow meter, and is causing it to read inaccurately.
- Something got in my loop and is causing an obstruction actually reducing flow.

I am getting really tired of draining, and troubleshooting this thing. I am pretty close to just hooking up the rest of the wiring and firing it up, looking at my temperature deltas between loop temp and GPU temp at full load (they were between 4C and 8C depending on title before at my measured 1.1GPH in my old loop) and using that data to decide if I have a real flow issue or just a flowmeter issue, and then if it is just a flowmeter issue. live with it, order a new flow meter, and install it the next time I break into the loop.

I could drain it (AGAIN!) and take out and disassemble and clean the flowmeter to see if there is anything wrong with it too.

Not sure what to do next. I am starting to get really tired and frustrated with everything going wrong with this build. I have been without my main machine since late January at this point...

I'm also completely out of distilled water. I had just enough premixed coolant to fill the loop, with only about an ounce or two to spare. If I spill any during the drain and fill I won't have enough to fill it again. I have two bottles of coolant concentrate I could use to supplement it, but I am completely out of distilled water to to mix them with. At target I've been buying it for under a dollar a gallon. Everything on Amazon is $18 or above per Gallon.

I can't leave the house right now, at least not until next Monday, as my stepson is currently here, and he is set to go back to his dads house next Monday, where they have a high risk individual living, so we have been vountarily self isolating, not leaving the house at all, so I can't pop over to Target and pick up distilled water...

You know, I don't mind not seeing people or going places, but having my projects grind to a halt due to long wait times or inability to get needed parts and supplies is driving me absolutely insane.

I'd appreciate any thoughts, or suggestions!
 
Last edited:
Could it have to do with that pump top? You have noticed some unusual noise coming from the second d5. I have had some experience with the single version of that EK/d5 clamp system. Try to loosen the screw locking in that d5 with the rubber ring. I had to back mine up a hair in the last build I used that EK plastic clamp/rubber ring system. When I first locked in the pump and switched on it made crazy grinding noises until i backed up a turn or 2 on the retention screw.
 
Last edited:
Could it have to do with that pump top? You have noticed some unusual noise coming from the second d5. I have had some experience with the single version of that EK/d5 clamp system. Try to loosen the screw locking in that d5 with the rubber ring. I had to back mine up a hair in the last build I used that EK plastic clamp/rubber ring system. When I first locked in the pump and switched on it made crazy grinding noises until i backed up a turn or 2 on the retention screw.


Good suggestions, I appreciate it.

I was hoping the air would be the case. I had it running overnight thus far with no improvement through. It is my experience that even plain distilled water without a surfactant would have worked out troublesome trapped air by now, and that the coolant with a surfactant definitely should have, but I'm not going to rule it out, and will continue running it hoping for improvement.

I don't think the pump top is the cause. I do have a much louder than expected humming sound, but it does not sound like anything is grinding or scraping, and the RPM of both pumps is exactly where one would expect based on the pump specs (just north of 4800rpm). If it were grinding against the inside of the top, I would expect that to slow down the rotational speed due to the friction. Also, it would explain why flow from both pumps would be affected.

My theory on the humming sound is shifting away from the pump and pumptop and towards it maybe just having to do with the layout of the case, sadly. I am going to program the aquaero such that the loud pump only kicks in during high load when the fans are spinning up, so it will give me max performance at load, and won't be a nuisance at idle.

Right now, I am working on cable routing and wiring up all of the internals, so I can do a GPU load test and see what the delta T between the coolant temp and the core temp is like. I know my old loop sat at about 1.0 to 1.1 GPM at full speed on one pump, and that resulted in the GPU generally running between 5 and 8C above water temp.

I'll run my GPU load test with only one pump running, which gave me ~1.1GPM in previous flow tests. If my GPU to coolant temp delta is greater than the 5-8C I am used to, I'll know the flow is actually low. If it is about the same, I'll know the flowmeter reading is the problem.
 
Alrighty.

I did some follow up testing with the system assembled and running, and I believe I have confirmed that it is the flowmeter readout that is incorrect, not the actual flow that is diminished.

I loaded up the system with a loop of heaven benchmark at 4k, maximum quality settings, vsync off, and let it run until the system reached a steady state.

The system was maintaining a temperature delta between the measured coolant temperature at the GPU and the core temperature of the GPU of only between 2 and 4C. If the flow meter readouts were accurate, this should not be possible.

I wonder if the flowmeters can be opened up and maintained...
 
So, I decided that the flowmeter is the culprit, but that I am not going to drain and refill until after the pandemic is over, so I lugged the thing up from the basement to my office...

...and it damned near killed me. Fully loaded with radiators, blocks, components and 3L of coolant, this thing weighs a bloody ton. At 6'3" and 240lb, I'm no small guy, but I could barely budge it at times.

In the end I probably wouldn't recommend this case. it is spectacular, beautiful to look at, huge and spacious, will fit more radiators than you can shake a stick at, will fit two goddamned computers, weighs a bloody ton, and has no handles. There is literally no good place to grip this thing to move it. And that snot to mention its other annoying design flaws.

Anyway, I'll post some more pics of the final wrapup over the next couple of days. For now I am going to rest. Moving this thing literally took everything out if me. I hope I didn't hurt my back! (I'm being serious!)
 
Last edited:
It's amazing how heavy these things get. I only have 2x 240s with a 750ml res and 4 spinner HDD and I swear the damn thing weighs like 40lbs.
 
So, I decided that the flowmeter is the culprit, but that I am not going to drain and refill until after the pandemic is over, so I lugged the thing up from the basement to my office...

...and it damned near killed me. Fully loaded with radiators, blocks, components and 3L of coolant, this thing weighs a bloody ton. At 6'3" and 240lb, I'm no small guy, but I could barely budge it at times.

In the end I probably wouldn't recommend this case. it is spectacular, beautiful to look at, huge and spacious, will fit more radiators than you can shake a stick at, will fit two goddamned computers, weighs a bloody ton, and has no handles. There is literally no good place to grip this thing to move it. And that snot to mention its other annoying design flaws.

Anyway, I'll post some more pics of the final wrapup over the next couple of days. For now I am going to rest. Moving this thing literally took everything out if me. I hope I didn't hurt my back! (I'm being serious!)

Yeah I had to put my Enthoo Elite upstairs to have more room to work. I think it gained about 15 pounds since it went up there, and it was already really heavy. Coming down is going to be scary.
 
Corsair lists the empty case at 65lbs. Once complete, these builds can start to get surprisingly close to 100lbs.

The flow rate is odd but, honestly, that's about the same flow rate that all four of my flow sensors (2 MPS400, 2 virtual from the D5 Next pumps) come up with on my loops.

Which meter are you using? The booklet from the MPS sensors warns that they're calibrated for a specific fluid. I believe they're a thermal sensor like a MAF in a car, so the thermal properties of the fluid matter. At any given flow rate, a fluid with better themal properties would read higher in this case.

I've noticed issues in my loop with either bubbles or flow stall in my MPS sensors. They can become quite divergent from the sensors in my pumps if I don't have the flow rate high enough. A vertical installation might resolve this if it's just bubbles getting trapped.

There is also a calibration method for the MPS sensors in the manual which I don't recall offhand very precisely. I recall it being pretty straightforward though, with you essentially just halting flow for 30 seconds or so while it performs its own calibration.
 
Alright.

Here's a note about how easy it is to lose a couple of hours or more with a build like this.

Before I took all the Noctua fans out, I found that a couple of them (furthest down on the bottom) were not spinning. Something had obviously become disconnected, and of course it was in the worst possible place.... In there:
IMG_20200329_185828.jpg

Had to take the top fan off to even get access. Now I had to get in there, but only two fingers fit at a time.

And there it is in the far back, one of the fan plugs has popped out of its extension, and is just about unreachable.

1587605851469.png

Not wanting to drain, and break down th eloop, and dismount the radiator in order to reach it, the two fingered fumbling began.


IMG_20200329_191408.jpg

Almost, but not quite. Just can't get it in!

And in the end, SUCCESS!

IMG_20200329_191532.jpg

I was able to get in there with a needle nose plier and hold the receptacle with it, until I could shove the fan plug in.

This little detour took what seemed like forever. I don't know how long it was, but it was definitely an hour or more of cursing.


And everything was back where it belonged again... For now...

IMG_20200330_134434.jpg
 
Next a side note about loop contamination.

I keep having this white stuff get stuck on the sides of my reservoir.

1587606309088.jpeg

It's only on the top where the waterline sits at one time or another during filling.

It's not even visible when water or coolant is in the loop, but as soon as I drain it, it is back at the top, so I disassemble the res and wash it again.

Flush with DI, drain, wash, flush with DI, drain, wash, literally rinse and repeat.

It has diminished with each flush, but very very slowly

The thing is, I thought I took extreme measures to wash and flush out my new radiators. I mean, I hooked them up to the hot water spout in the laundry singk, and just blasted hot water through them for ages. Then I took them up to the kitchen. Did a fill and shake with dish soap, followed by a nother sink rinse, and then a fill and shake with a vinegar solution, followed by one more final long sink rinse, and still I get all this gunk?

I remember this happening when I built my last loop back in 2016 as well.

Two things are the same.

1.) I had new Alphacool radiators, and they have a bit of a reputation of being a bit gunky... (but seriously, THAT gunky?)

2.) I flushed my loop with the "Sysprep" bottle that came with my Primochill tubing.

With sysprep, my thought process was "it can't hurt, might as well try it", but maybe it can hurt after all?
 
While down there I looked over at the workbench. My dust mask I used when grinding all that plastic early on in the project was there.

IMG_20200412_171428.jpg

What do you know. I had a freaking N95 mask all this time...
 
So, Moving along to the new fans.

I wound up buying Arctic P12 (16x for the front) and P14 fans (6x for the top, and 1x for the rear vent) based on good words from another user on the Aquacomputer forums.

I have to say, I am impressed with these things.

They are cheap (like $8-9 per fan, if you buy the 5x bulk packs, compared to Noctua's $26 to $30 per fan)

They work really well.

They have these little pigtails that allow you to conveniently string up to 5 fans in a row:


IMG_20200412_174203.jpg

These pigtails are a great idea. They make large builds like this with many fans in a row so much easier to assemble. I just with the pigtails were slightly longer. Like, exactly the length of the side of the fan. That way everything would line up perfectly, and not bunch up.

Instead we wind up with an awkward bundle like this:

IMG_20200412_174122.jpg

It is still way easier than with individual fan cables though. Hopefully I won't have to go through that ordeal with the fan extension cable in the unreachable spot again.

Also, here is where I give away a free business idea to any premium fan manufacturer who wants it.

1.) Make a fan. A good fan. Static pressure, low noise all of that.
2.) Give it a detachable modular fan cable.
3.) Put little golden fingers on all four 8 sides of the fan, connected by wires internally to the motor.
4.) Manufacture little jumpers, you can just press down, joining two adjacent fans electrically.
5.) Voila. Now you only need one fan wire to the fan closest to the source. The rest are joined to eachother using jumper clips.
(be sure to include either two different sizes, or universal joiner clips so they fit the two different spacing standards for radiators)

Feel free to steal this idea. I'm sure as hell never going to use it.


Anyway, back to the Arctic fans.

So, the hard black plastic comes across as a little cheap, especially compared to the fancy. But you know what. THEY ARE cheap. And they are pretty great too. Yes, their max static pressure and airflow numbers are below those of the Noctua iPPC fans I was initially building with, but they are also quieter at max speed. Id' wager that if you ran the Noctuas at 75% speed, they'd both produce similar airflow, static pressure and noise levels as the Arctics.

Since my main rationale for building this monster build is that I never want to run my fans at full noise levels again, do I even care about this?

At this point my only reason for wanting to go back to my Noctuas (which I do) is that they are so damned nice looking.

Noctua got back to me, btw. Have offered to replace my fans with ones that follow the PWM spec, so eventually I will get the replacements, and when I do, I'll eventually get around to swapping them in.
 
Last edited:
Also,

I forgot to mention.

After I sent them this video EK OK:ed my noisy pump replacement:



They originally thought I was running the pumps at different speeds. Once I assured them that both were actually running without the PWM connectors plugged in, and thus at full speed (4800rpm) they agreed to replace the noisy one.

Then I got the replacement. The pump was a little loose in the package banging around. It came out not exactly looking pretty. I couldn't remember what my previous ones looked like, but those scratches just didn't look right:

IMG_20200417_211415.jpg

I have since been assured that this is in fact how most D5 pumps look, and it is in fact normal.

Either way, installed the new pump, and it is still as noisy as the one that was replaced. I don't know what the problem is. I have tried with different O-rings to see if that solves it as previously recommended in this thread. That did nothing. Maybe it is the Dual Pump top that is somehow uneven or out of tolerance? I have no idea, and I am done fighting with it.

I'm just going to set up pump control such that during idle, the quiet pump is the only one running. The noisy one will only kick in during high load, at which point I'm hoping the fans will drown it out.
 
Your images of loop disassembly to hunt for wires made me realize something: This case is so thick that you could put QDC bulkheads on the motherboard/component firewall and then plug your radiators directly into there and do the loop routing back there. Pumps could be hidden inside of the HD box. That way, you'd get super clean routing AND nice easy maintenance.

Nuts. I might have to build a second one of these.
 
And there it is.

It is done.

When I finish a big project I normally like taking a glamor shot for posterity and posting it here. This time I'm just exhausted. I don't want to deal with it anymore.

You get phone picture of it sitting on my workbench. Please pardon the fingerprints. It's been a long project, I forgot to wipe them down, and the flash made them a hell of a lot more visible than they were before.

1000D.jpg


It took 29 years after I built my first PC as an 11 year old of steadily building larger and larger computers (if we ignore that brief stint with a Shuttle SFF build in ~2009) but I finally did it. I built a PC I can barely carry.

I did eventually get it up the stairs to my office last night though. My back hurts a little today, but it is there.

For a second I was worried it wouldn't fit under my desk. I had measured it - of course - but you know how that goes. Turns out it just sneaks under the desk, but that also became a problem.

My old Corsair 750D case shipped with terrible little rubber feet which came off after no time at all. That can definitely not be said for this behemoth. There is a massive metal foot plate on the bottom, and the rubber pads on the bottom of it are solid and sturdy. I liked this at first, but it made it damned near impossible to slide it on my rug under my desk. Standing with those rubber feet on the rug, neither god nor man could slide that case.

For a while there I thought I may have to move the desk., lift the case in, and then place the desk down on top of it.

Then I did some thinking. I found an old stack of health insurance cards. You know those thick laminated paper-like things. I stuck one under each rubber pad, and what do you know. With the rubber separated from the rug, I could slide the thing like a sled under the desk.

Here it is:


IMG_20200421_223756.jpg


(Please ignore the Yoga block speaker stand. They were supposed to be a temporary thing until I built my own wooden stands. That was 3 years ago... I swear, I'll still get around to it!)

After some tweaking of pump and fan speeds, this thing is great. Amazingly silent at idle, and very mild mannered even at full load. I'm very happy with the cooling performance.

I forgot the cardinal rule of noise though. You never get silence. Once you remove the loudest noise, something else becomes the loudest noise, and your ears adjust, and now that noise starts to annoy you.

For me right now that is the fan on my power supply. It's the only audible thing at idle, and where before I couldn't even hear it, now it is bothering me. I guess I'll add "PSU Fan Mod" to the list of future projects.

Anyway, I'm calling this one done, which is good, because I am just about done. And I need to move. And fix my car. This thing has already consumed way too much of my life :p
 
Last edited:
That's awesome! Congrats on the completed build!





Dut, you know, if you don't want to hear that PSU any more, there are a lot of PSU these days that will operate fanless during web surfing and Office loads. There are some, like the overkill AX1600i, which will even operate fansless under gaming.

Just sayin' ;-)
 
That's awesome! Congrats on the completed build!





Dut, you know, if you don't want to hear that PSU any more, there are a lot of PSU these days that will operate fanless during web surfing and Office loads. There are some, like the overkill AX1600i, which will even operate fansless under gaming.

Just sayin' ;-)

My Seasonic Prime Platinum 1200 has a "hybrid mode" which is supposed to allow it to operate fanless at lowmloads, but it does not appear to work. Maybe I am just drawing too mucb power at idle with all of these pumps, fans, 24 cores etc.
 
Congrats on finishing your build.

I haven't updated my system in 6-7 years. I think the last time I changed my coolant was 4 or 5 years ago, only topped off since then. Want to tear it down and upgrade eventually, but haven't found a compelling reason to due to primarily playing indie games.
 
Alright.

Test mounting radiators is done, and this case has displayed some more stupid annoyances.

The sliding fan/radiator mount brackets that come out for easy access are an amazing feature. I love them. But Corsair put their holes in really stupid places...

The screw holes in the 3x 140mm mount up top are oblong so you can slide things back and forth to give yourself some leeway. Well. At least the front and the rear fan holes are. The one in the middle? Not so much. So, in order to give me as much space in the front as possible, I slid the radiator as far back in those oblong holes as possible. This means the middle fan holes don't line up anymore...

View attachment 223847 View attachment 223848

In the grand scheme of things, it's not the hugest of deals, because with all the other screws in place, even with the center fan not being screwed in at all, it does not budge. It is sandwiched in there and will not move. So I guess it will work anyway.


I was really worried I wouldn't be able to fit the radiators both in front and in the back, but I was able to make it work. The top and the bottom need to slide in and out together now though, as the only way to make it work was for the bulge on the top radiator to fit in the gap between the two on the bottom.


View attachment 223849 View attachment 223850 View attachment 223851

It's going to be tricky to link them together up top. Probably going to require a 90 degree swivel bend on each of the ports, but it should work. Hope the tight bends won't hurt flow too much...

So, push-pull both top and front are not going to work at the same time:

View attachment 223852

Maybe I can use some sort of 140mm to 120mm fan adapter on that last slot up top and make it fit. Any suggestions?

If I only had another half an inch it would work, which the next part really frustrating.

Why on earth do I need so much clearance in the back?


View attachment 223853 View attachment 223854

Two whole inches wasted in the back.

This is a recurring theme with this case. Such an awesome case with so many awesome features only to be held back by some stupid design decisions.

I wish I were better with sheet metal. I'd fabricate my own mounting tray up top, moving everything back a bit.


Next issue. The motherboard tray rises up from the base of the case causing a little bit of a space restriction. In push-pull with a 60mm radiator this only gives me a quarter of an inch of clearance for airflow.


View attachment 223855 View attachment 223856 View attachment 223857 View attachment 223858

So, that drive cage is definitely going to have to come out. The back third of the rear radiator is going to have flow constriction due to only having a quarter of an inch of clearance, but that I'm going to have to live with.


It's not as bad as it looks through, as there are passthrough holes behind the motherboard tray, so some air can flow through there.

View attachment 223861 View attachment 223862

The back of the case also has some vent holes, so some air should be able to flow through there.


This is a glorious sight though. I love it:

View attachment 223859
...you don't realize how much one can spend in fans until... you buy them all!
 
My Seasonic Prime Platinum 1200 has a "hybrid mode" which is supposed to allow it to operate fanless at lowmloads, but it does not appear to work. Maybe I am just drawing too mucb power at idle with all of these pumps, fans, 24 cores etc.

Might there be a software setting you need in order to enable that? Or just a switch you need to hold? With non-streaming web surfing, my build seems to pull only 125w or so, and that's with 30w coming from the extra GPUs. Compare that to a laptop power brick at 80-90W which doesn't even have a vent nevermind a fan. I'd think any desktop PSU these days should be able to handle that fanless.
 
I wound up buying Arctic P12 (16x for the front) and P14 fans (6x for the top, and 1x for the rear vent) based on good words from another user on the Aquacomputer forums.

I have to say, I am impressed with these things.

They are cheap (like $8-9 per fan, if you buy the 5x bulk packs, compared to Noctua's $26 to $30 per fan)
ive been really happy with mine too. great bang for buck!
 
WTF! Go to start up my desktop this morning and it has no power. Just like last time.

It took months of RMA's and rebuild to get to the point of getting it up again, and it lasts a few days and then it is down again.

Same as last time, it was working perfectly, shut it down at night. Won't power on next day. No apparent leaks or other problems. I'm pissed.

Not quite sure what to do. Last time, after mistakenly RMA:ing the PSU and motherboard, it turned out to be a dead CPU.

I just don't have the time to do more troubleshooting right now.

What are the chances of getting two bad Threadrippers in a row, one an RMA for the last?
 
How dead is dead? My build usually takes 4-5 attempts to get going from a cold start. Not really sure why, but my money is on some type of PSU circuit protection kicking in as a result of the inrush.
 
How dead is dead? My build usually takes 4-5 attempts to get going from a cold start. Not really sure why, but my money is on some type of PSU circuit protection kicking in as a result of the inrush.

It went from easily powering up on first try for teh week and a half since the build completed, to not powering up at all now... I've tried many times, tried pulling power from the wall, and plugging back in, nothing.
 
How dead is dead? My build usually takes 4-5 attempts to get going from a cold start. Not really sure why, but my money is on some type of PSU circuit protection kicking in as a result of the inrush.

I did some googling to see if anyone else was having issues with their Threadrippers, and found a few posts of people suggesting their default voltages were WAAAY too high. Like 1.48v.

These symptoms of running fine for a little while but dying over time line perfectly up with a CPU that has been overvolted.

I can't confirm this theory though. Before, I never looked at the voltage. I usually never do, just trusting the default BIOS settings, unless I plan on overclocking, which I had no plans on doing with this CPU. And now it is too late to check...

Running theory right now is that this is due to BIOS default voltage settings being much too high.
 
Do you get fans twitching and do any relays click? If so, that's how my build is. It didn't happen when brand new but it started to happen over time. I'd hit the power button, I'd hear the relay(s) click, motherboard RGB would turn on, fans would twitch, and then the machine would turn off. The whole process takes maybe half a second. It's like the in-rush is too big so the PSU shuts down. By the 4th or 5th try, it usually boots.

Of course, now that I think about it, this hasn't happened since I put the machine on the new 20A circuit.
 
That is a lot of radiator porn, did you get it all working in the end?

I really like the black tubing, it has a nice industrial look.

Alright, so believe it or not I am still working on it.

I moved this year, and it was way more work than I expected, delaying work on the build significantly.

Lots of updates to come in the near future.

I'm still cursing this damn case though. It is pretty, but I got it because it is large, and I figured a large case would be easy to work on, but I still have to squeeze my hands in constricted spaces where I can't see and try to tighten down fittings.

It's a little bit like changing the rear plugs on a V8 Chevy Monza, or a late 90's DOHC V6 Ford Taurus.

A lot of this shit:

PXL_20201222_232942829.jpg


I know they say being tall with big hands is supposed to be a huge benefit as a man, but more and more I wish I had been born with genes that blessed me with 5 feet in height and tiny little hands.

Would be so much easier to do these things, and man would I be happier on planes.
 
Last edited:
For instance, these little fuckers on the bottom of the radiators are damned near impossible to get in.

PXL_20201223_034514346.jpg
 
Some updates.

1.) I really hate the Corsair 1000D case. If I were to do it all over again, I'd never buy it. It is every sort of bleeding knuckles, tight fits without visibility awful. I feel like I am stuck with it for the time being due to the radiators I selected going with.

2.) I finally got the water loop reassembled, and all the wiring connected to a new loop configuration (which I will show off when I have everything running)

3.) My fan issues are resolved. Well, sortof. Noctua replaced all of my fans claiming that the ones I would be receiving would be from a new batch, using the old design before they messed up PWM compatibility, but these behave the same as the old ones did. Fans stay at 0 rpm from 0% to 97%.
Fans start spinning at high rpm at 98% and a little faster at 99% for the max speed of the fans. I am using a workaround for this. The Aquaero allows you to set a minimum and maximum fan speed and it re-scales the range to th erange that you set. When I set it to a minimum of 98.3% and a maximum of 99.3% it re-scales the fan range to work within those two values. I expected this to be awful, since I'd be compressing the entire speed range of the fan down to just one percentage unit, but it actually works fairly well.
Now when the fan gets a signal of 0% they spin at about 400rpm which is dead silent to my ears, and they scale up pretty smoothly until 61% when they hit max speed of about 1850rpm. I would compress the range even further to be able to utilize the full range of values, but unfortunately, the software does not allow using less than 1 percentage unit between the min and max, so this will have to be it.

In testing I had an issue where random fans would turn off. Turns out this was because I was using too small of a PSU for testing. My old 150w Antec PSU's from some old cases I gutted, just couldn't provide enough on the 12V rails to power all the fans consistently. Once powering them from my 1200W PSU, all the fans spin as expected.

4.) My new loop which I will show off later, is working well, but now I have the dreaded, press power, nothing happens, problem. I have checked and double checked my connections. I have also unplugged the PSU from the motherboard and GPU and powered it up by bridging the "green wire" to ground so I know it works and powers up.

At this point I have a brand new CPU from AMD, my third, I have tested with other RAM, I have tested with all of my expansion cards unplugged in the past. Despite having sent the damn motherboard to Gigabyte and have it returned with "no problem found" I'm blaming the motherboard. It's the only thing it could be.

I'm going to do a little more troubleshooting today, but if I can't get it to work, I guess I'll just have to order another one, and eat my $500, as they are never going to give me a refund a year later...

I'm probably never buying anything from Gigabyte ever again though. That is for sure.

It's outright depressing to think about how much my hardware has depreciated in the time I haven't had the system up and running...
 
Last edited:
It's outright depressing to think about how much my hardware has depreciated in the time I haven't had the system up and running..
lol my cpu went from $330 to if i were to sell it today in the forums (120-150max) dropped half its value in only year or so time:) I wonder how your current cpu compares to the new 5950x? I really want to get one of the 5x series now
 
lol my cpu went from $330 to if i were to sell it today in the forums (120-150max) dropped half its value in only year or so time:) I wonder how your current cpu compares to the new 5950x? I really want to get one of the 5x series now

Yeah, well, at least you got to use it during that time, right? :p

Alrighty.

I did some more troubleshooting, checking all connections and removing all other components. That motherboard and CPU are a complete brick together. Not any indication of power at all. Not even a test LED.

Short the power on pin to ground on the power supply and the power supply powers up though, so I know it isn't the PSU.

Since I have tested without everything else connected, and the CPU is a brand new RMA from AMD, the only conclusion I can make is that the motherboard, despite Gigabyte's RMA hvaing tested it and found it to be functioning normally, and sent it back is not working.

I went to order a replacement, but everything on Amazon has really long delivery dates, so I went to Microcenter today and bought the only TRX40 motherboard they had on hand. An open box (Asus Certified with Full Warranty) Asus RROG Zenith II Alpha. (why does this make it sound like this motherboard is not even in Beta? :p Curious choice of name there Asus...)

I'm a little concerned about the open box nature of the thing, but it was the ONLY ONE they had, and I can't lose any more days to shipping. I need to get this done during my days off from work, so it is what it is.

The Asus ROG board gets high ratings in user reviews. It's a little pricy, and has way more Christmas tree lights (and an OLED screen? Lolwut?) than I would have liked, but I've gone way too long without my main rig due to moving in the middle of this mess, and just not having the time to troubleshoot.

I'm doing the motherboard translplant now. Wish me luck!
 
Last edited:
Well, there they are side by side.

Surprisingly similar looking. Is here a reference design for motherboards?

PXL_20201227_012215097~2.jpg


Will have to read the manual to figure out what the hell that weird 9th dimm slot is for...
 
Well, at least the nerve wracking, potentially pin bending part is over....


PXL_20201227_015213251.jpg


This motherboard is weird though.

Two eight pin and one six pin and one four pin molex connectors on the board...

And then there is that weird extra RAM slot just labeled DIMM2.

Also looks like it came with a ROG screwdriver, USB Stick (?) and some sort of coaster (???)

PXL_20201227_015610409.jpg


Going to have to read the manual before proceeding
 
Last edited:
Solved another mystery by reading.


The "DIMM 2" slot is where the proprietary format m.2 explanation card goes if you wany to add more than three m.2 drives.

PXL_20201227_021512166.jpg

Cool solution, but really dumb name.

Probably won't be needing that. At least not initially.

Additionally one of the three m.2 slots is in the back of he board, so if you want to install something there, the entire motherboard has to come out...

Brilliant design ASUS.

Makes me wonder if I should install my second m.2 drive in slot three on the back first, so that if I want to add a third drive in the future I don't have to take the board out...
 
Last edited:
Well, I am dead in the water again.

The beauty of buying a open box motherboard is that the little blister pack with all the specialty M.2 risers and screws is missing...

The ones from the Gigabyte motherboard won't work, because they were some sort of specialty riser/screw combo which also attached the heatsink.

MicroCenter has a few generic ones (which hopefully will be the right size) in stock. Will have to go there (again) tomorrow.
 
Alright.

Ready for an open air test using an old Quadro I had kicking around.

If this doesn't work I am all out of ideas.

Fingers crossed!
PXL_20201227_030711800.jpg
 
Back
Top