Cooler Upgrades

I should've wrote about this first since it happened before the other stuff, but that didn't happen obviously.

Anyway, I decided to upgrade Melty and Mei-chan with all-in-one liquid coolers and utilise Mei-chan's Zalman 9900 Max for Lie-chan, since Lie-chan's "theme" is blue (given it's not very visible).

For Melty, I went with Corsair's Hydro H100i v2, and for Mei-chan, I went with Corsair's Hydro H80i v2. We'll just say that the H80i v2 was interesting without spoiling... Anyway, I also decided on Fractal Design's Ventury high pressure fans in place of Corsair's stock radiator fans. I'll just say that this was a costly upgrade, but it'll eventually be worth it for many reasons.

Anyway, I first thought about what I needed to do, because Lie-chan had Melty-'s stock AM3 bracket, and Lie-chan's stock AM2+ bracket was laying around. So first I would need to remove the Zalman 9700 from Lie-chan and Zalman 9900 Max from Mei-chan before I can even install anything.

After doing so, I installed the AM2 bracket onto Mei-chan's motherboard and reinstalled her into her case (forgot to say that I carried her to the room where Lie-chan is to make it easier to do this all). Installing the H80i v2 in Mei-chan was interesting to say the least... It's something that was definitely possible, but also something that's kinda one of those things where it spawns some "I dunno if that's a great idea..." or something like that I think.

According to Corsair, you're supposed to install the radiator first and then the water block, but that wasn't happening because of Mei-chan's case. Radiator orientation was another thing, since there were only two possible orientations I could install it that would fit within the case. I also redid the cable management in her since I had to rerun the EPS cable since the original route would be obstructed by the cooler.

Redoing the cable management was decently easy since the cables have been "broken in" and I was able to run it kind of how I originally wanted to when I transferred Mei-chan to that case.

I also added a USB 3.0 PCIe card and a 3.5 inch 2-port USB 3.0 front bay panel, and I had to get a bit creative to route the PCIe power to the 6770.

Anyway, I'll shut up for a bit and show some pictures.

Motherboard view.

Hidden portion.

Since I was able to, I took some inside pictures, though I used the closed panel on the open side instead of the fan panel... I'll probably make more sense with pictures, since I'm failing to describe well at the moment.

Taking pictures as I creeped my phone into the crevice.

There was really limited space for the cable for Corsair Link USB cable, but it was pointless to install since there wasn't an "off-brand" Corsair Link program for Linux. Anyway, the northbridge heatsink was the limiting factor on the space, and as much as I wanted to change the colour for the Corsair logo on the water block, I didn't want to hack the Corsair cable to get Triela up in Windows and download and install the software to do a quick change. Eventually I found and remembered that I have some right-angle mini B adapters that I can just connect a regular A to mini B cable to connect to Triela. I'll eventually do it, but that's eventually...

Anyway, I also put one of the Fractal Design's high pressure fans in the front of Mei-chan's case and swapped the NZXT white fan that was on the side for Fractal Design's high flow fan. I ended up adapting the high pressure fan in front since it was a bit noisier at full speed than the NZXT fan that I had originally put there, but I didn't take pictures of it. I left the one on the side alone since the sound profile wasn't as bad as it was before the upgrade.

Anyway, pictures.

Can't see the fins any more since it's black and not white.

(I think I forgot to say that I moved Mei-chan back to her respective location after getting her back together.)

I "peeled" the filter back to try to show the water block and clearance of the radiator.

I think I was hoping to put another high flow fan in the top, but that wasn't going to happen once I saw the clearance. Anyway, I booted Mei-chan to look at the temps and such and to try to take a picture of the white Corsair logo on the water block.

It's fuzzy, but you can barely make out the text... Maybe... (I didn't "peel" the filter back for this shot because I still had Melty to do.)

Satisfied enough with the results, I quickly focused on Melty because I was going to install her "in-place" instead of pulling her out from under the "desk" while there was still enough daylight to do it. I obviously removed the Zalman 9900 from her and noticed that it was much easier than Mei-chan's (I'll rant about this in a bit). And once that was done, I reinstalled her bracket and cleaned the thermal paste (maybe not in that order). I also forgot to reground myself at some point while cleaning her CPU, but got lucky.

I played around with the cooler seeing which way to run the tubes, and I decided to run it at the back of the case. I also swapped the stock rear fan with another Fractal Design high flow fan. Installing the radiator was interesting because the screws were actually too long and I had to snag some of the washers from the fan side to make up for it. After mounting the water block, I kinda did a quick tuck/run of the cable because I was going to eventually pull her out to redo the cable and install the drilled out front panel for the soundcard switchboard.

Anyway, a quick picture I took before doing a "quick" test and to play with Corsair link before dinner.

It's hard to tell from this angle, but the tubes don't touch anything.

I made the logo red, and followed something I read on Corsair's forums or whatever to set it at 254 instead of 255 so that it properly save it to the water block's firmware memory thingy. It's not really red like the Sapphire logo or the LEDs from the Sound Blaster Z card and sort of a pink, but there's probably not much I can do about that and I don't care enough (given my K95 is set to a pink).

After dinner I focused on getting Mei-chan's 9900 Max onto Lie-chan. I found that the screws were better made from the 9900 Max I pulled off of Melty, and then eventually found that the front brackets were also different (the back brackets and nuts were the same though). I don't remember what I did, but I used the screws from Melty's 9900 Max (and probably the front bracket too?) and got Mei-chan's 9900 Max on Lie-chan. Once Lie-chan was back together, I booted her up for a quick test.

I was quite happy, since she was much quieter and much more consistent in fan speed and noise. With the 9700 (which is non-PWM), she would boot quiet, and then after a reboot, she would boot loud (obviously another reboot she would be quiet again). No longer, since the 9900 Max is PWM.

I think I also did some cable management.

I also swapped out the Rosewill rear fan for Mei-chan's Cooler Master rear fan but had to add it to the rail that went to the motherboard's weird "extra PCIe x16" Molex power header. I think I had at least more slack than the Rosewill fan though. I think I was wanting to use one of the fan headers on the motherboard, but it wasn't going to reach with the graphics card and PCI slot fan.

I was wanting to also do a H80i v2 for Lie-chan, but either I would have to make some sort of sturdy adapter or transfer Lie-chan to a mid-tower case, with the latter option probably being the best option. I might in the future, but not now.

I think I was finished about seven in the evening, starting from about seven in the morning, and went to relax for the night before heading to sleep.

I did have another picture of Mei-chan after installing the H800i v2, but I deleted it off of my phone for whatever weird reason...

Before reinstalling the other add-on cards and cable management.

Also forgot to mention that the H80i v2 pulls some of the exhaust heat from the 6770 through the radiator, but without a blower-style GPU, there's not much I can really do.

Anyway, I felt pretty satisfied to have cleaned the cable management on Lie and Mei-chan from what they were and everything is better than it was. Melty's a bit quieter, and it's nice, but it might take a bit of time to get used to... Since I can't hear if she's on or not with the H100i v2 and the new rear fan.

I think I covered everything and I've spent a ton of time typing today. Hopefully I'll post about the soundcard switchboard soon before my probable radio silence until I do more with the Raspberry Pi clock project thing that I hopefully touched on at some point.

Anyway, ciao!

P.S. I think I might've used the 6-32 x 3/16 PPHMS screws I got from Fastenal for mounting the H100i v2, but I can't remember and I'm not at home.

Soldering Station and Protoboard Power 4

Now that everything is made and assembled, and that the pictures are uploaded, I'll now go about it.

I got the PCBs first, but I set them aside since I didn't have any of the parts from Mouser. I think I tried to do the soundcard switchboard first and then switched to this, but I can't remember at the moment.

I got the plates back from Xometry and their bead blast finish wasn't really what I was expecting - I kinda wished I didn't spend an extra 3 USD for it.

Hard to tell the coarseness of the sandblast, but I tried doing what I could with my phone.

I took them to work to beadblast myself and once I figured out how to get the booth set up, I ended up with a much nicer finish, though with a bit of scarring from Xometry's bead blasting.

Again, did what I could with my phone.

Once I received the parts, I began assembling the PCB. I apologise ahead of time that I literally took a picture after soldering each item.

The usual 3-up, top, and bottom pictures taken of the main board.

The usual 3-up, top, and bottom pictures taken of the fan and light board.

I started with the main board, since I wanted to get past the static and/or moisture-sensitive components first (the SMT LEDs that I ordered are moisture-senstive). I first started with the red and green SMT LEDs since those were going to a bit more difficult to position than the red SMT LEDs.

3-up of the red/green SMT LED soldered, single-board view, and somewhat close-up view.

Next was the red SMT LEDs (again, moisture sensitivity).

3-up of the red SMT LEDs soldered, single-board view, and somewhat close-up view.

After I put the other two boards away, I soldered the remaining static-sensitive component, the SOT--23 dual-diode device (BAT54CLT1G).

I used a couple of the fan/light boards to prop the board so that I wouldn't put pressure directly on the SMT LEDs.

Next was the SMT resistors for the LEDs.

Again, using the fan/light boards to avoid pressure to the SOT-23 diodes.

Next was the fuse indicator relays (G6L-1P-DC12).

Made it somewhat neat, though it's not going to be readily visible.

Now to focus on the main components, starting with the shortest component, the SMT fuses.

Blank blue ones are 2A fuses (hence the "2" next to them), and the ones with the "M" on them are 4A fuses (hence the "4").

Next in height was the Hirose headers for the blowers and the output to the breadboard.

Nothing really striking to note here.

And then the Hirose header for the output to the fan/light board, which was a bit surprising to me that it was shorter than the Molex header for the input.

Big, but not as big as some of the other components.

The Molex input header was next.

Flange thing to prevent someone plugging in a 24-position ATX connector directly to the board, though someone could plug in a 20-position ATX connector, though it'd probably be obvious since the header is 22-position. I think the clip cutout might prevent the 20-position ATX connector from being plugged in.. I've not checked.

Last on this side is the relays, which were much larger than I thought they were.

Relays kinda dwarf the rest of the components.

Last item to be soldered were the switches, which were a bit tricky to get the bushing parallel to the Z axis, but I ended up doing a partial assembly to get them lined up with their respective holes just enough.

I left the standoffs attached since I was going to be needing them soon.

I realised that the anti-torque washer and the internal tooth lock washer weren'tt going to fit because I only calculated for the widest point that the nuts can take up. While it wasn't a big deal, it was definitely an "oops" moment.

I pushed the light pipes into the panel, which was a bit tougher than I was used to, since the ones at work are a lot easier (probably the holes in those panels are probably bigger than the manufacturer recommends). While I felt like it would be pointless to use the retention rings, I installed them anyway. I used some blue Loctite for the bottom nut and finalised the board install to the panel.

I think I had the "Main power" label on the panel before I started working with the board.

Next was to finish the board with the switch labels.

Not professional-looking, but decently discernable what each switch does.

A couple quick shots between the board and the panel.

I set the assembly aside in a anti-static bag and began with the fan/light board, starting with the short Hirose headers for the lights.

A lot of spots for the fan headers, but just wait.

Next was the fan headers, and I did the corner ones first before utilising another fan/light board to help keep the other 26 of them in place.

Yep... Wow.

Last is the other Hirose connector.

Meh. Seriously, meh.

And now I needed to work on the cable between the main board and the fan/light board. I had worked a bit of it at work from some cable scraps, so I essentially just had to crimp the connectors on and load the plugs correctly.

Connectors crimped, plug loaded, both ends.

I then covered the individual cables with heatshrink and then shrank the heatshrink.

Before and after.

And then I covered the somewhat ugly cable ends with another tube of heatshrink.

Before and both ends after srhinking.

Next was to make the harness, which would be a lot easier, since I wouldn't have to deal with the fairly stiff cable of the fan/light cable. I crimped and loaded the mating receptacles for PCIe, EPS, and ATX 24.

EPS, ATX 24, and PCIe portions.

I then loaded the harness plug.

Loaded, or so I thought...

Now it was time to hook everything up and give it a go.

Harness plugged into the board; unit attached to the PSU with cable and fan/light board; fan attached, fan and PSU oriented.

I hit the power switch on the PSU and the light didn't come on, after a minute of trying to figure out what was wrong, I shut it off and looked at everything - there was also a slight smell, but I'll get to that in a bit. If you go back to the previous post, and take a good look at the pictures here, you'll see what I did... Which was flopping the wire order in the harness plug around.

In this orientation, the 12 volt wires (yellow) are supposed to be at the bottom of the plug.

Luckily I had the Molex deplugging tool, so fixing the problem was fairly quick and easy.

Harness now in the proper order.

And then I set everything back up and attempted the test again.

Second attempt.

Tried it again, and no dice. I took the broken-ish multimeter I had (the one in one of the other pictures) to measure the voltage output for PS-ON and +5VSB, which both gave no readings, the PSU died from whatever shorted. I luckily had another PSU around though it was a 20-position ATX and not a 24-position. With how the harness is made, it doesn't really matter that much.

Setup, oriented, PSU switch on, power switch on, fan switch on, blower and light switches on. I also had another picture with the blower and light switches n the down position, but I accidentally deleted them from my phone since I couldn't tell what it was.

Third time was a charm and I was just as excited as Frankenstein when his monster came alive. Now it was time to set up the unit at the solder station.

Overall view, and close up of the fan/light board.

The fan/light board is in a temporary location, and I'll eventually make something so it's not dangling there. I have actually put a spacer between the board and the fan so the screw is straighter, and have spliced a fan connector to the fan, but I didn't take pictures because I forgot. Anyway, I decided to have a bit of fun before I went to bed.

White fans connected to fan/light board, light and blue fan connected to other power supply via breadboard.

At some point I got the Engineer Inc. PA-20 crimper and was able to properly crimp the tiny Hirose connectors, and I did the connectors for the blowers and lights.

Wires stripped for crimping, crimped connectors, loaded plug.

If I remember, I'll take some more pictures. After getting my Iwiss ratcheting crimper, I decided to test it on the cable for the breadboard power, and after tinning the wires, I decided to dress the panel a bit with a OSH Park sticker that I received with the second version of the soundcard switchboard (more on this in the respective post).

Now it's not as plain.

The Iwiss crimper worked a lot better than the Engineer Inc. crimper, and I loaded the breadboard power strips accordingly, though I had to substitute yellow for green (12V).

I'll eventually clean the breadboard...

The station is complete (at the cost of a PSU), and I'll eventually take some pictures of everything. I would've been a bit more careful and double-checked the harness, but I was short on time.

Eventually I'll remember to switch the switch on the PSU on before switching the power switch, but it might take a bit of time (it also doesn't help that the switch is under the power cord).

As expensive as this project was, it was definitely worth it. Eventually I'll add more fans and give myself a larger space to work, but it is what it is for now.