Build #26: JER mini
Ever since I first saw them, I thought the keycaps in the GMK Triumph Adler 90 numpad kit were great. Just that simple touch of a circle around the legends gives them a unique character (pun intended). When I saw the JER mini, a 65% with an additional four-key column on the left side of the board, I knew I had found the perfect pairing of keyset and kit.
Or so I thought…
Let’s begin with a look at the PCB included with the JER mini kit. You’ll notice the four-key column on the left, but another interesting feature of this board is the indicator placement for the number lock, caps lock, and scroll lock keys, in the lower right of the PCB on a tab that juts out from the board.
Flipping it over.
Clipped and lubed GMK screw-in stabs, in place on the PCB.
If you didn’t already notice, the bag of insulating washers in the first couple of photos wasn’t there for show. For this board, at least one of the stab mount holes looked like a prime candidate for a screw short, so I preemptively placed washers on all of the stab screws. Here’s a closer look at the problem child.
The plate for the JER mini is attractively finished. So attractive, in fact, that I didn’t notice one issue until later in the build. Do you see the problem? We’ll come back to it later.
Here, I’ve mounted three LEDs in the NCS pads just prior to soldering.
A few Tealios have been mounted in the plate and soldered in place for good fit, prior to doing the whole board.
Aside from getting the PCB and plate properly fitted, installing switches in stages also allows me to do the Sorbothane strut mod for sound dampening. Here’s a closer look at one such strut, placed just to the left of the left arrow keyswitch.
All soldered and ready for installation. See the problem with the plate yet? Don’t worry, we’ll get there.
Moving onward, the base of the JER mini has an offset brass weight, screw-mounted into the base. It’s an interesting design choice, and I like it. I’m less fond of the obvious machining marks around the interior corners of the board, but there were no marks of this sort on the exterior, so life goes on.
Smudged brass. It arrived like this. I still haven’t decided if I’m going to try to clean it, or just let it get smudged and develop a “lovely patina”. While we’re here, notice the unusual design of the base, consisting of a front base and back base with different slopes. A little odd, but in an appealingly quirky way. Also endearing: “March 2019.A.D”. Assisting future keyboard historians trying to place this board’s millennium of origin, perhaps?
The user-facing edge of the top of the JER mini also has a quirky look to it. The front edge has a bar which is screw-mounted to the top of the case; this bar covers the NCS indicator LEDs, and provides a slightly rounded bezel on the front of the keyboard. That’s all good and well, but less good and well is what I found when I attempted to attach the bar. I wasn’t getting good fit initially, and upon unscrewing, what do we have here? A helicoil.
The other screw hole had one as well, but at least that one didn’t come out with the screw. Still, disappointing that this was the choice that was made, instead of properly tapping screw holes. This is the first time I’ve seen this in a kit.
I was originally planning to use Sorbothane for PCB/case interaction, but I didn’t have Sorbothane thin enough. I switched to 2mm craft foam, which was about the right thickness - at least given the next issue.
All screwed together, but not well; the screws meant to attach the plate to the case weren’t long enough, so initially, this ended up being a sandwich mount board, look Ma, no screws. Also note that the bar in front is a little out of alignment, due to the screw problems I mentioned earlier, but we’ll correct that later.
Did you figure out the problem with the plate? The cutouts for the plate are on the wrong side for the Enter key’s stabilizer wire, so on the upstroke, the left wire makes contact with the plate. A little bit of additional lube on the underside of the plate above the wire mitigates this to an extent. Still, starting to form an impression of the QC for the design and the kit…
The next issue is all on me, though. Finally, let’s put those beautiful TA90 numpad caps on that southpaw function column, the really cool lettered ones. Nice! Nice? Wait a minute.
When I mentioned the TA90 numpad kit, I left out an important piece of information. This isn’t just a numpad kit - it’s an R0 numpad kit. Somehow, I failed to understand the significance of that until I put the caps on the board. Welcome to the Great Wall of Keycaps. Look at the height of those monsters relative to their partners across the aisle.
So, we switch back to regular R1 function keys. Sigh.
What about the other problems with the board?
- I switched to M2x6 screws to attach the plate to the case, where I discovered that these holes weren’t tapped much better than the helicoiled ones, but I was able to get enough screws in to hold things together. With the plate mounted as intended, I had to take the craft foam out of the case as there wasn’t enough room for that any longer, but fortunately, the sound wasn’t negatively impacted.
- I installed a new helicoil in the screw hole for the front bar - not great, but enough to hold things together. I suspect I will need to return to this hole and redo the helicoil or retap the hole.
After all of that, how is the board? Despite all of the QC failures, I like it. The external aesthetics match expectations, and the sound is crisp and clacky without being irritatingly clacky. I suspect the tight tolerance between the PCB and the case has something to do with this clean sound.
- This was my first time using bootmapper. A quirky interface, to be sure, but it was able to define macros interactively, something I’ve found missing from other interactive programming tools. All things being equal, however, I’d still prefer QMK if that was an option here. What can I say, I like to code.
- This was my second time using 205g0 on switches, and I think my application this time around was right on the edge of being too much - I wouldn’t say the switches are sluggish, but they do feel just a tiny bit slower. I’m a bit surprised by this, though, as I only lubed the stems, not the housings, in an attempt to keep exactly this problem from happening, under the theory that switch use would be enough to spread the lube where it needed to go. Clearly more experience is needed.
- When it comes to fit, finish, and QC, not all kits are created equal. I’m afraid I’ve been spoiled by the level of QC in several other kits I’ve built, and I understand the risks involved in group buys. On the other hand, at the price points of typical custom kits, unmet expectations sting. Welcome to the hobby! In fairness, I note that the GB runner is attempting to address issues that have come up with this GB.
- I’d really like to understand the effect of internal case volume on sound. I need more boards with tight PCB-to-case tolerances.
- R0. Gotta pay attention to row profiles! With that said, I see why R0 would be appealing on larger boards and/or boards with wide bezels.
How about GMK TA 90 R2 with an R2-4 numpad?
case: JER mini (light blue)
case dampening: n/a
PCB: JER mini (Bootmapper)
plate: gray aluminum
plate/PCB dampening: Sorbothane strut mod (50 Duro)
stabilizers: GMK screw-in (1x6.25u, 3x2u)
stabilizer mods: clipped, lubed with SuperLube, installed with fiber washers
switches: V2 Tealios
- springs tub-lubed with Krytox GPL 104
- stems hand-lubed with Krytox GPL 205g0
- no housing lube
keycaps: GMK Triumph Adler 90
HxWxD (without caps or feet): 1.13" x 13.38" x 4.56"
HxWxD (without caps): 1.19" x 13.38" x 4.56"
HxWxD: 1.38" x 13.38" x 4.56"
assembled weight: 3.64 lb