The Z Build 2.0

Aaand, we almost immediately had a leak.

A surprisingly slow one too, considering at some point the o-ring on this Aquacomputers temperature sensor fell completely out, and I never noticed...

IMG_20200326_220143.jpg

Digging through the water cooling box for spare o-rings that look like they will fit well...
 
Hmm.

I'm wondering if I got a bad pump.

The pump I was already using in my old build was nice and quiet. You can barely hear it.

The new pump, holy shit, the pitched hum is enough to wake the dead...

See below, I first plug in my old one, then plug in the new one, and then disconnect the old one.



Any thoughts? I'm wondering if I should request an RMA.

They are both EK D5 PWM G2 pumps.

My old nice and quiet one was manufactured in 2018, the new loud one was manufactured in 2019.

edit:
Apparently hearing th edifference depends very much on what speakers you are using. On the built in speakers on my laptop I can't hear either of the pumps, just the creaking floor from the footsteps upstairs.
 
Zarathustra[H] said:
Hmm.

I'm wondering if I got a bad pump.

The pump I was already using in my old build was nice and quiet. You can barely hear it.

The new pump, holy shit, the pitched hum is enough to wake the dead...

See below, I first plug in my old one, then plug in the new one, and then disconnect the old one.



Any thoughts? I'm wondering if I should request an RMA.

They are both EK D5 PWM G2 pumps.

My old nice and quiet one was manufactured in 2018, the new loud one was manufactured in 2019.

edit:
Apparently hearing th edifference depends very much on what speakers you are using. On the built in speakers on my laptop I can't hear either of the pumps, just the creaking floor from the footsteps upstairs.
Click to expand...

I just had a similar problem but even more severe. I am using a singularity d5 pump top and I thought I killed two pumps somehow as I would hit the switch and nothing. Put in a 5 year old swiftech and it pumped water but was incredibly loud. Turns out it was the Oring. The singularity supplied Oring was ever so slightly to thin and when compressed the pump impeller(am i using that right?) would jam or grind against the bottom of the pump top. Dropped in an old black Oring and that swiftech was perfectly quiet, as were the other two pumps I thought had failed.
 
Smoked Brisket said:
I just had a similar problem but even more severe. I am using a singularity d5 pump top and I thought I killed two pumps somehow as I would hit the switch and nothing. Put in a 5 year old swiftech and it pumped water but was incredibly loud. Turns out it was the Oring. The singularity supplied Oring was ever so slightly to thin and when compressed the pump impeller(am i using that right?) would jam or grind against the bottom of the pump top. Dropped in an old black Oring and that swiftech was perfectly quiet, as were the other two pumps I thought had failed.
Click to expand...

Hmm.

Interesting.

I've never taking my pumps apart. How difficult is it to source these D5 O-Rings?
 
Smoked Brisket said:
I just had a similar problem but even more severe. I am using a singularity d5 pump top and I thought I killed two pumps somehow as I would hit the switch and nothing. Put in a 5 year old swiftech and it pumped water but was incredibly loud. Turns out it was the Oring. The singularity supplied Oring was ever so slightly to thin and when compressed the pump impeller(am i using that right?) would jam or grind against the bottom of the pump top. Dropped in an old black Oring and that swiftech was perfectly quiet, as were the other two pumps I thought had failed.
Click to expand...
Zarathustra[H] said:
Hmm.

Interesting.

I've never taking my pumps apart. How difficult is it to source these D5 O-Rings?
Click to expand...

Wait a minute... Are you talking about the one that sits between the D5 pump and the top? The big one, maybe 3-4" in diameter?
 
As an eBay Associate, HardForum may earn from qualifying purchases.
Alright,

So first, a note about how much coolant this loop takes:

IMG_20200328_180525.jpg

That's a gallon bottle which was full before filling the loop. Just by guesstimation I'm calling that 2/3 of a gallon. Maybe a little bit more.

Now about the noisy pump.

Smoked Brisket said:
I just had a similar problem but even more severe. I am using a singularity d5 pump top and I thought I killed two pumps somehow as I would hit the switch and nothing. Put in a 5 year old swiftech and it pumped water but was incredibly loud. Turns out it was the Oring. The singularity supplied Oring was ever so slightly to thin and when compressed the pump impeller(am i using that right?) would jam or grind against the bottom of the pump top. Dropped in an old black Oring and that swiftech was perfectly quiet, as were the other two pumps I thought had failed.
Click to expand...

Smoked Brisket said:
Yes, the ring between the pump and top.
Click to expand...

This seemed like a very reasonable root cause to me. I also considered that maybe the fact that ym mounting holes weren't perfectly straight might be putting undue forces on the system that could be causing noise.

Since it is a pain to do just one, refill and test I decided to do them all at once and see if it fixed it.

So, first, I widened the mounting holes. Originally they were drilled with a 5mm bit. I widened them with an 8mm bit, and then took my Dremel to them to even out the edges:

IMG_20200328_184402.jpg

I was a little concerned that the rubber vibration dampening washers might pull through, so I poked around my tool box and found an old bag of #8 washers, which I added on the bottom.

IMG_20200328_184414.jpg

Then with the pump out, I pulled the noisy side off to examine the o-ring. I also pulled the o-ring out of my old XSPC Photon reservoir for comparison:

IMG_20200328_184746.jpg

It's a much thicker O-ring, and it is slightly smaller in diameter. I was able to streatch it and make it fit in the EK housing though, but it took some serious hand strength to tighten down the fixation ring. I was a little concerned it might leak, but I decided to try it and see what would happen.

Also, while the loop was drained , I decided to do some cleaning.

I'm slightly regretting using the Sysprep in my loop. It seems like it loosened a lot of gunk in the new radiators, and then proceeded to deposit it at the top of the resrvoir, right on the water line.

IMG_20200328_181853.jpg

I took the glass cylinder out and washed it. A little bit of a pain in the ass. Most reservoirs I've used are threaded on either end and easy to take apart. This one uses a rod down the center, which holds everything together, so you need more than two hands to hold everything in place when aligning everything and tightening it back together. Lets just say I prefer the threaded ends.

Anyway, whatever this was I wiped off looks grey on the glass, but on the paper I used to clean it looks green. I'm guessing it is copper oxide, and that the Sysprep is a mild acid. I'm not entirely thrilled. Hopefully a few flushes will get everything out.

While I had the reservoir apart, I removed the vortex inhibitor in the bottom, and installed a tube on the port that drains to the pump. I had noticed that a lot of air bubbles returning from the loop were being sucked right back into the pump. I was thinking a extension tube there would limit the air bubbles being sucked back in, and might also help all of the water in the reservoir be used rather than just sit there while the same water keeps being recirculated.

It was partially successful. Air bubbles still get sucked in, but not as much as before:



So I reassembled everything, and the second pump is quieter than it was, but still much louder than the first one.

I am torn on what to do. I might RMA it, but I am not sure. At idle it would be obnoxious, but I could set the aquaero up such that at idle it is only using the quiet pump, and only turns on pump 2 when fluid temp goes high enough that fans would be spooling up anyway, and probably drown out the pump noise. Going to have to think about this one.
 
A few more notes.

While flushing the loop, I decided to power up the Aquaero, and hook it up to my laptop using an internal USB header to USB-A plug adapter

I was intrigued by some of the flow meter readings.

One pump at 100% (doesn't matter which one) results in 1.1GPM through the loop.

Two pumps at 10.0% results in 1.6GPM

These are much higher flow rates than I was anticipating. I'm happy.

There is always a challenge though.

My temperature sensors are of the fancy "Calitemp" Aquabus variety. They are supposed to be more accurate, but who the hell knows. In order to get all four to work, I have been using a Aquabus x4 adapter. In my last case I used a short 4-pin fan extension to avoid mounting the x4 board straight to the aquabus. This configuration required a longer extension, which of course caused the aquabus temperature sensors to not be detected by the unit. I have confirmed that they work by plugging them straight in.

So, now I have to disconnect the Aquaero again, and install the X4 board directly to the unit, and figure out how to route the temperature sensor cables to their destination in this huge case...
 
Alright. A few more issues, one of which leaves me absolutely dead in the water. :(

I'll tackle the big one first.

My Aquaero works perfectly with the six 140mm Noctua fans I moved over from my old build. They are speed controlled as expected.

It does not control my 16 new 120mm Noctua fans well at all.

When I try to control them the output looks (approximately) something like this:
0% PWM -> 0% Fan Speed
10% PWM -> 0% Fan Speed
20% PWM -> 0% Fan Speed
30% PWM -> 0% Fan Speed
40% PWM -> 0% Fan Speed
50% PWM -> 0% Fan Speed
60% PWM -> 0% Fan Speed
70% PWM -> 0% Fan Speed
80% PWM -> 0% Fan Speed
85% PWM -> 10% Fan Speed
90% PWM -> 50% Fan Speed
95% PWM -> 75% Fan Speed
100% PWM -> 100% Fan Speed

I could contine like this, controlling the fans with limited granularity, but it is also not reliable and sometimes will not work at all.

I spent hours troubleshooting. Things work better with fewer fans per port, but still not the way it is supposed to.

I suspected splitter problems, cable problems, that my Aquaero was dead (and I may have damaed it during my troubleshooting, because now PWM port 4 doesn't seem to work at all) but in the end, it was none of these things.

A visit to the Aquacomputer community forums reveals that the problem is Noctua. On their newer revision of their fans they decided to disregard the Intel PWM spec from 2005 and do things their own way. EK has apparently done the same with their fans with similar results.

The result? They work marginally as single fans, and absolutely will not work well at all controlled as large groups of fans.

I'm hoping it is not too late to get Amazon to take all of my fans back. I spent a shit ton of money on these things. And if they do, now I need different fans. I'm going to have to disassemble the loop to get the fans out, re-route all the goddamned cabling.

I am absolutely fuming right now. I thought Noctua were the good guys...

Anyone know if there is a Noctua hardware rep on these forums?
 
Thunderdolt said:
WTF. That would make my blood boil.

Seems like the kind of thing that Steve @ GN would rip the fan makers on if he knew about it.
Click to expand...

Yeah, I am pissed. I checked. More than 30 days have passed, so I can no longer return these to Amazon. I'm trying to speak to an actual person to see if they will make an exception, but I can't seem to figure out how to get past th ebullshit menus to speak with an actual person. Maybe I am too tired. I will try again tomorrow.

I have sent an email to Noctua's support. I will see what happens.

I am not that frequent of a visitor of GN, but I will see if I can bring it to his attention, thanks for that tip. I think I may have linked to some of his stories back in 2017 when I was working the HardOCP front page news, but I can't remember.
 
Well, I'm hoping Noctua will take responsibility for this and do something. Otherwise I am out almost $500 in fans and Chromax rubber edges for those fans.
 
Yikes, here's to hoping Noctua corrects that for you. They need to update their product pages (as does EK) that their fans do not work with traditional PWM solutions.
 
Thunderdolt said:
WTF. That would make my blood boil.

Seems like the kind of thing that Steve @ GN would rip the fan makers on if he knew about it.
Click to expand...

I emailed Steve at GN. Have not heard back yet, but thank you for that suggestion.

I have also posted a HardForum News post here.

Hopefully this will save someone else from being in the same situation I am in.
 
That sucks. If the problem is that Noctua expects a pull-up or driver from the controller side then it sounds like you can add your own pullup with some resistors. Create a voltage divider with two resistors between +12V and GND on the 4-pin. You can use a 2 kohm and a 3 kohm or 3.3 kohm. Connect the middle to the PWM signal to pull-up the line to ~5V. So +12V ----- 3 kohm ----- PWM signal ----- 2 kohm ------ GND. You would need to do this for each fan.
 
I'm still hoping Noctua comes through, but I am unwilling to wait. Their support line says they will respond within 7 days.

In the mean time, some users on the Aquacomputer forums have expressed excellent results with Arctic P12 and P14 fans.

They do not have as high max flow and static pressure numbers as the Noctuas, but they do provide a decent amount, and reportedly do so in near silence up to a surprisingly high rpm.

I like silence. The 140mm iPPC Noctua's I ahve been using are inaudible below ~475rpm, very quiet up to about 650rpm, OK at 800rpm, but obnoxiously loud at above 1000rpm. At their full 2000rpm speed they can wake the dead. My reason for going with so much radiator capacity in this build was that I never want to have to run them at high speeds because of the noise, so I will unlikely need the higher capacity of the Noctuas anyway.

As an added bonus they come with some sort of wire that allows daisy chaining up to 5 fans in a row, which ought to make my wiring situation much more manageable.

Estimated arrival date is April 6th through 10th.

We will see when they get here...
 
Out of curiosity, have you run the Noctua fans (with the splitter) to one of the PWM headers on your motherboard?
 
gigatexal said:
Dang dude. I’m seriously impressed.
Click to expand...

You know, my original plan was to go with quad 480's. Two in the front, two in the top. That is what Corsair claims this thing will fit.

Turns out it won't though unless you use really thin radiators and fans, as the top and front will interfere in the top corner

Then I priced out dual 360's up top and dual 480's in the front. I balked at that, in large part due to the fan prices adding up.

So instead, I kept my existing 420mm rad for up top, and added dual 480s in the front.

The irony is that now that I ahve selected cheaper fans in place of the Noctuas that don't work, I actually wouldn't have had an issue with the price, and would have saved myself a ton of time and effort making the custom bracket I needed in order to make everything fit.

This is stage 1. Chances are I'll add a set of 360's up front in the future :p
 
I don't think you're missing much by having only the one radiator up top. The second radiator would be pretty choked on the fan exhaust due to the shelf.
 
Thunderdolt said:
I don't think you're missing much by having only the one radiator up top. The second radiator would be pretty choked on the fan exhaust due to the shelf.
Click to expand...

There is actually a surprising amount of space there, I don't think it would be chocked at all.

Here is a picture from when I first got the case.

1585707012659.png

Here, with the radiator installed, you can see we have about 70mm, almost 3"

1585707120157.png

Subtract out 25mm for a layer of fans for the push pull, and maybe another 15mm if I use a 60mm thick radiator instead of the 45mm thick one that is in there now, and that leaves me with 30mm clearance. That's more than an inch. Should be plenty to get some airflow going.

Some air will go through the cable passthrough holes as well to the other side of the case, and can vent to the outside via the gaps around the back glass panel.

I don't think it would be bad at all.
 
Removing the grommets would certainly help with flow there. The thing is, the shelf there would probably still hamper flow if those fans were exhausts. If those fans were intakes, then the front rads would need to be exhausts or else you'd run into the lack of exhaust problem that I may be seeing on mine. I actually never thought that would be a thing, but I guess normal cases don't have 22-28 intake fans.

I'm thinking it might be time to cut some clearance into the back of mine. Have you seen how obstructed the exhaust fans are?
Image from iOS (9).jpg
 
Zarathustra[H] said:
I'll put a layer of paint on the top of the table to prevent it from rusting again...
Click to expand...

Please don't paint the table. It's too beautiful.
I can't recall the brand, or type, but there is a dry lubricant spray, either silicone or graphite, I don't remember. But it's used on table saws and other wood tools. Prevents rust and keeps it shiny smooth.


Love the attention to detail! And pics
 
Interesting note:

My 24 fans, while shipping from Amazon don't get the normal prime shipping right now, but arrive In a few days, with no issues or warnings.

Aquaero parts - however - have a warning about other items being higher priority and taking longer to ship, and won't arrive for more than two weeks.

One would think that both of these were equally non-essential?
 
Last edited:
Zarathustra[H] said:
Interesting note:

My 24 fans, while not shipping from Amazon don't get the normal prime shipping right now, but arrive Inna few days, with no issues or warnings.

Aquaero parts - however - have a warning about other items being higher priority and taking longer to ship, and won't arrive for more than two weeks.

One would think that both of these were equally non-essential?
Click to expand...

My hard line couplers were deemed essential but my soft tube couplers were not.
 
Most likely due to the fact that Amazon warehouses are understaffed and have to fulfill a variety of needs, while third party suppliers don't have to prioritize essential items. The delay is in the warehouse, not the shipping service.
 
You should put this beast up on million dollar pc
 
Zarathustra[H] said:
Alright. A few more issues, one of which leaves me absolutely dead in the water. :(

I'll tackle the big one first.

My Aquaero works perfectly with the six 140mm Noctua fans I moved over from my old build. They are speed controlled as expected.

It does not control my 16 new 120mm Noctua fans well at all.

When I try to control them the output looks (approximately) something like this:
0% PWM -> 0% Fan Speed
10% PWM -> 0% Fan Speed
20% PWM -> 0% Fan Speed
30% PWM -> 0% Fan Speed
40% PWM -> 0% Fan Speed
50% PWM -> 0% Fan Speed
60% PWM -> 0% Fan Speed
70% PWM -> 0% Fan Speed
80% PWM -> 0% Fan Speed
85% PWM -> 10% Fan Speed
90% PWM -> 50% Fan Speed
95% PWM -> 75% Fan Speed
100% PWM -> 100% Fan Speed

I could contine like this, controlling the fans with limited granularity, but it is also not reliable and sometimes will not work at all.

I spent hours troubleshooting. Things work better with fewer fans per port, but still not the way it is supposed to.

I suspected splitter problems, cable problems, that my Aquaero was dead (and I may have damaed it during my troubleshooting, because now PWM port 4 doesn't seem to work at all) but in the end, it was none of these things.

A visit to the Aquacomputer community forums reveals that the problem is Noctua. On their newer revision of their fans they decided to disregard the Intel PWM spec from 2005 and do things their own way. EK has apparently done the same with their fans with similar results.

The result? They work marginally as single fans, and absolutely will not work well at all controlled as large groups of fans.

I'm hoping it is not too late to get Amazon to take all of my fans back. I spent a shit ton of money on these things. And if they do, now I need different fans. I'm going to have to disassemble the loop to get the fans out, re-route all the goddamned cabling.

I am absolutely fuming right now. I thought Noctua were the good guys...

Anyone know if there is a Noctua hardware rep on these forums?
Click to expand...

Man sorry to hear that. Going to post a reddit link to a guy with a similar fan setup. Hes a good dude and will probably answer any questions you have. Take a look at his connectivity setup:


On my build I got the fans working, but using double sided LL120s/Silent Wings 3s (2 Corsair fan/led controllers, commander pro, and 2xphantek fan controllers, splitters everywhere).
 
capnstabn said:
Man sorry to hear that. Going to post a reddit link to a guy with a similar fan setup. Hes a good dude and will probably answer any questions you have. Take a look at his connectivity setup:


On my build I got the fans working, but using double sided LL120s/Silent Wings 3s (2 Corsair fan/led controllers, commander pro, and 2xphantek fan controllers, splitters everywhere).
Click to expand...


Yeah, so I have gotten through with Noctua Support. They promised me they would find a solution and requested the lot codes and information from my fans. I provided it end of last week, but haven't heard back yet. (Not complaining, it has been a weekend after all)

In the interim I have also ordered a full complement of Arctic P12 and P14 fans which are reported to work like the traditional PWM fans should, without degradation by adding more fans per channel.

When I ordered them they claimed they were not on Amazon's "non-essential" list, but they have not shipped yet, and are now estimating an arrival date of April 15th, which is kind of a bummer.

At this point I guess I will know what Noctua decides to do by then, so I'll know whether or not to return the Arctic fans.

It's a little disappointing to be cooped up in my house due tot he virus and not even have my PC for some occasional distraction, but I do understand that there are more important things to ship people than computer parts right now...
 
Zarathustra[H] said:
Yeah, so I have gotten through with Noctua Support. They promised me they would find a solution and requested the lot codes and information from my fans. I provided it end of last week, but haven't heard back yet. (Not complaining, it has been a weekend after all)

In the interim I have also ordered a full complement of Arctic P12 and P14 fans which are reported to work like the traditional PWM fans should, without degradation by adding more fans per channel.

When I ordered them they claimed they were not on Amazon's "non-essential" list, but they have not shipped yet, and are now estimating an arrival date of April 15th, which is kind of a bummer.

At this point I guess I will know what Noctua decides to do by then, so I'll know whether or not to return the Arctic fans.

It's a little disappointing to be cooped up in my house due tot he virus and not even have my PC for some occasional distraction, but I do understand that there are more important things to ship people than computer parts right now...
Click to expand...

QQ are you only running one Aquero? I think you are going to need at least 2 for 24 fans.*EDIT* saw that you are using the unit that has the screen. I wonder if you just need another one of the simple aqueros that runs into one of the fan channels on it, and the simpler aquero can provide extra power delivery but still control/output data on everything.
 
Last edited:
capnstabn said:
QQ are you only running one Aquero? I think you are going to need at least 2 for 24 fans.*EDIT* saw that you are using the unit that has the screen. I wonder if you just need another one of the simple aqueros that runs into one of the fan channels on it, and the simpler aquero can provide extra power delivery but still control/output data on everything.
Click to expand...

I don't need one fan channel per fan.

That's only if you use volt control for for some reason want to obsessively read the rpm.of every fan.

If you use PWM you only need one fan channel per fan speed you want.

All versions of the Aquaero 6 have 4 PWM channels.

My plan is to use one PWM channel for the 6 140mm fans up top, one PWM channel for the 16 120mm fans up front, and the remaining two for Pump1 and Pump2.

I was originally going to share one PWM channel for both pumps, but since one pump is so much louder than the other, I plan on keeping the quiet one on all the time, and only have the loud one kick in during high load.

Maybe I'll change this if I am successful in RMA:ing the loud pump with EK.
 
You must log in or register to reply here.
Back
Top