GH60 prototype betatesting [Call for layouts, GUI testing]

[komar007]

Thanks, how thick is the PCB and how thick is the PCB w/ the USB?

Standard FR4 1.6mm, 5.6mm with USB. The connector is subject to change, because it is not available in China + we may switch to micro.
Anyway, it will rather get smaller than larger.

[WhiteFireDragon]

komar, a few things to consider changing in the final production board. I know I just designed a switch hole to accommodate switch rotation, but keycaps cannot be rotated. The "+" shape from the switch stems don't have the same thickness. The vertical part is thinner than the horizontal part. If you try to install keycaps, it's very stiff, although you can still force it on. However, once you do this, you'll loosen the female part of the keycap stem. If you install these keycaps on other switches, it might be too loose. This won't be a problem if you buy a set specifically for this keyboard, but if you transfer keycap sets between keyboards a lot, you'll have a few ruined caps.

Now it's a question of whether you value you keycaps more or want more layout options for the PCB.





For the USB, it shouldn't be too hard to change 5 pads for the pins and the 4 large pads for micro USB support.

[The_Beast]

What caps have to be rotated for what layouts?

[WhiteFireDragon]

Almost every key that is not 1x size, and a few of the 1x keys too.

[The_Beast]

well that sucks giant dongs

[IvanIvanovich]

It's not really an issue with keycaps. It dosen't make them so loose they will fall off by themselves if you turn the board upside down, or at least i never had that problem and I have had some boards with 90 rotated switch. The wider crossbar is the important one and that is not the one that will get looser.
If you remove all the rotated switches you end up with many layout not possible anymore.
It might be a consideration to some, but redesign the whole thing and lose layout choice is too drastic for the sake of keycap might get a little looser fit.

[komar007]

Yes, I'm starting to see this is a bigger problem.
The only thing is when we put a few switch footprints near each other, we loose PCB mounting support.
Maybe we can fix that by adding two pads where the diode normally goes and have the user put a wire there to mount the switch more firmly.

[WhiteFireDragon]

Some overlap location is possible but some are not because then you can't even mount the switches properly. Personally, I'd rather keep all the layout possibilities and sacrifice the caps. Those loose caps won't totally but useless, it'll still mount without falling off although it'll just be looser. It might be a problem for people using rare vintage sets or expensive sets, or for resale value. SP sets are known to have loose modifier to begin with, so all the bottom modifiers have rotated switches so that'll make it worst.

Workaround is to just put a thin piece of tape in the female stem of the caps, although that's not as elegant. Personally I don't care for this "problem", but just pointing this out to be considered in case you decide to change anything.

[komar007]

Well, I don't know. Seems to be the next thing to make a poll for. The first one would be the mini vs micro usb.

[TheSoulhunter]

Humm, I'm strongly against the rotated switches...
To me such "workarounds" are against "an engineers pride" n producing a quality product, if that makes sense.
But then again, I'm from Germany and my profession is to build high speed trains, so my thinking might be biased

[The_Beast]

Humm, I'm strongly against the rotated switches...
To me such "workarounds" are against "an engineers pride" n producing a quality product, if that makes sense.
But then again, I'm from Germany and my profession is to build high speed trains, so my thinking might be biased

I feel kinda the same way

[komar007]

Humm, I'm strongly against the rotated switches...
To me such "workarounds" are against "an engineers pride" n producing a quality product, if that makes sense.
But then again, I'm from Germany and my profession is to build high speed trains, so my thinking might be biased

Well this is a decision to be made between rotated switches and overlapping drill holes, which means soldering to an incomplete, not fully metalized pad.
The question remains: what is worse, rotated switches and tight keycaps or less durable, harder to solder joints + lack of one PCB support pin on most switches.
I honestly don't know...

[alaricljs]

Kind of boils down to how many people are wanting to use this without a plate.

[jdcarpe]

Kind of boils down to how many people are wanting to use this without a plate.

I personally won't be using it without a plate. The PCB seems a bit flimsier than the Poker PCB and has quite a bit more flex, even with just normal use.

Unless we choose a beefier PCB, anyway.

Also, I am in the process of plate mounting all my Pokers, if that gives any indication if how I feel about it.

[IvanIvanovich]

I don't like plates, but if it is decided to go with redesign so plate required, I can just be happy with my prototype and not get anymore.

[TheSoulhunter]

Humm, I'm strongly against the rotated switches...
To me such "workarounds" are against "an engineers pride" n producing a quality product, if that makes sense.
But then again, I'm from Germany and my profession is to build high speed trains, so my thinking might be biased

Well this is a decision to be made between rotated switches and overlapping drill holes, which means soldering to an incomplete, not fully metalized pad.
The question remains: what is worse, rotated switches and tight keycaps or less durable, harder to solder joints + lack of one PCB support pin on most switches.
I honestly don't know...

Option 3, no compromises in quality but less possible layouts!?
But yeah, we should have a poll...

[relcc]

I would also go for plate mount. Personally I think trying to get a 'PCB mount multi-use board' is madness. Turning the switches 'the right way' would also solve the discussion in the other thread about rotated multi-use holes in the plate which may or may not work. And the stabilizer holes that do not fit the PCB size currently is also solved as stabilizers are mounted one the plate.

Another thing you would get; the possibility of running traces on the PCB for LEDs as switches can be mounted with LEDs again. I'm not a LED-person myselve, unless someone designs a board with per key programmable LEDs.

[komar007]

Let's get one thing straight: rotating the switches back to normal positions will not prevent plateless design. If we decide to do so, it's only a matter of the lack of one pin, which isn't such a big deal, and I'm going to add pads which will allow to mount the switch firmly with a piece of wire.
It's only that more pads will be incomplete and the joints may be more likely to break.

[relcc]

Maybe you can use thicker copper on the PCB to prevent breaking.

[WhiteFireDragon]

I don't like plates, but if it is decided to go with redesign so plate required, I can just be happy with my prototype and not get anymore.

I'm pretty happy with the prototype PCBs as is, so I probably won't be getting a final production board either.

But I'll see to it that these make it to the final revision and production before I go completely MIA from this GH60 project. I already have my fully functional GH60 board 

[Tranquilite]

I intend on using mine without a plate, and I don't particularly care about rotated switches, but on the other hand I don't mind doing a bit of finagling to get things working "right" either. My only concern would be if these design changes caused the prototyping phase to last longer (though if you do have to run more prototypes, I already have all of the switches/stabs/keycaps/etc needed to build one of these *wink wink*)

[TheProfosist]

You guys do realise that lots of keyboards use rotated switches already right?

[komar007]

Which ones?
I've only seen 180 degree rotations.

[TheProfosist]

Which ones?
I've only seen 180 degree rotations.

I believe my WASD has some in the right side mods i know for sure that the SteelSeriees 7G i used to have had that as well


ive had caps off too many keyboards to count and good number of them had at random it seems 1-3 switches rotated im sorry i dont have exact models but I havnt touched a non custom keyboard in a long time.

[asura]

0.15mm is well within the elastic deformation of ABS and should be within elastic deformation of PBT - depending on the technical quality of the original casting.

[IvanIvanovich]

For ones with 90... almost every commerical MX numpad. Also... after making an error with WFD in regards to what to solder for me... (some confusion on my part led to the smd diodes not solder and I sure can't do that tiny stuff...) and you mentioning changing it to use a tie down wire. Instead, you could go old school and use in switch diode instead of surface mount so it has a dual purpose. Maybe that will even be cheaper with the fab since they have to do less?

[jdcarpe]

Yeah, the in-switch jumper wire idea would be great. Then I wouldn't have to remove the jumper wire from all these PCB mount switches.

[komar007]

Yeah, the in-switch jumper wire idea would be great. Then I wouldn't have to remove the jumper wire from all these PCB mount switches.

But that would be just for modifiers.
So many extra holes will increase the cost and decrease rigidity.

[jdcarpe]

Still a good idea, even if it's just for modifiers.

[The_Beast]

Can someone tell me about how far the USB port will hang off the PCB?

[komar007]

Currently 2.8mm.
What are you making, The_Beast?

[The_Beast]

Custom case brofessor ^__^

[komar007]

If it's made of wood, count me interested:D

[The_Beast]

If it's made of wood, count me interested:D

Alum, wood maybe later

[TheProfosist]

If it's made of wood, count me interested:D

Alum, wood maybe later

beast can you make a stainless case?

[WhiteFireDragon]

Can someone tell me about how far the USB port will hang off the PCB?

Depends on how you solder it. The ones for the prototpe boards are fully surface-mounted, and the pads allow for quite a bit of variability. It can be mounted to stick out anywhere from 1.80mm to 4.00mm and still work fine. All the ones I soldered and sent out stick out about 2.15mm. For reference, both Poker and Pure stick out 2.60mm. They're through-hole, so that's not variable.

Interested in what case design you come up with, although I think I have an idea. My CAD file is pretty much done too, so maybe we can converge the design.

[regack]

wood! wood! wood! wood! wood!

...who am I kidding, I'll take whatever comes, because I can't CAD to save my life D:

[__red__]

My USB socket is reassessed 2mm.  I didn't want it protruding past the pcb.

[The_Beast]

Can someone tell me about how far the USB port will hang off the PCB?

Depends on how you solder it. The ones for the prototpe boards are fully surface-mounted, and the pads allow for quite a bit of variability. It can be mounted to stick out anywhere from 1.80mm to 4.00mm and still work fine. All the ones I soldered and sent out stick out about 2.15mm. For reference, both Poker and Pure stick out 2.60mm. They're through-hole, so that's not variable.

Interested in what case design you come up with, although I think I have an idea. My CAD file is pretty much done too, so maybe we can converge the design.

I can open AutoCAD, Solidworks or inventor part files. I designed mine in AutoCAD since all the shops I talked to perfered that over solidworks. Right now I'm having someone check over my design. If you want to send me yours, just PM me

[kmiller8]

Soo ve fnally gtten arudn to uttng m bar together wth swtcches. Everythngseems t be worng fne, ,ust the swtches ive use seem to e nicredbl finsey andd unreiable. I nee t get some newswtches and test those on the oaor....

On another keyboard: So I've finally gotten around to putting my board together with switches. Everything seems to be working fine, just the switches I've used seem to be incredibly finickey and unreliable. I need to get some new switches and test those on the board.

I know the FW works good, because if I manually short the pins on the back of the board, the controller picks them up instantly.

Assembly was a snap, no problems whatsoever with the PCB mounted switches I installed, and updating the firmware worked just fine. I really like the PCB mount option for stabilizers, and the rotated switches weren't a problem with the POM caps I was using.

Ok, that's about all I've got.

Thanks komar for the board, and WFD for the SMD work

[TheProfosist]

lol kmiller I though that first paragraph might have been german at first

[mkawa]

sigh, sorry folks, this is on my kbd todo list right after i make the next two ssk protos

[SmallFry]

lol kmiller I though that first paragraph might have been german at first

I thought it was written in Prof.

[TheProfosist]

lol kmiller I though that first paragraph might have been german at first

I thought it was written in Prof.

i swear its the keyboard not me

[WhiteFireDragon]

No one brought this up yet, but all the holes for the standoffs are slightly offset to the left. komar, can you move them over by a fraction of a millimeter? For example, the hole for the between the "G  -  H" keys and "]  -  \" keys should be dead center. The other ones are a little harder to reference, but I can give you the exact coordinates for those. I know you have oval holes to account for any offset, but if you fix the holes to exact locations, you won't need a "looser" hole. Also, the USB port is slightly offset too. It should be moved slightly to the right (looking at the top of the PCB).

Any updates on what you want to do with rotated switches or mini vs micro USB? Probably should get a poll up. You can write it or I can write it.

[komar007]

I'll look into the holes' positions problem more thoroughly.

How do you make a poll on geekhack? Could you do it?
I'm currently fighting with my firmware trying to get it to work on atmega32u2, for the upcoming GHPad;)

[WhiteFireDragon]

Ok, I'll take care of the poll.

[kmiller8]

so I re-built my GH60, I'll take some Keyboard GORE pics later of it. I may not be able to help design custom layouts or do some fancy firmware, but when I say I tortured this poor PCB, I tortured it...

BUT I'm typing on it right now, this PCB is amazingly high quality, after tearing apart a "practically filco" QFR for blue switches, I was HIGHLY disappointed with the PCB quality, but that's just a rant for another day.

So yes, pics soonish, and I'll continue to follow this thread watching for new firmwares to test

Once again Thanks Komar for the board, and WFD for the SMD <3

[alaricljs]

after tearing apart a "practically filco" QFR for blue switches, I was HIGHLY disappointed with the PCB quality, but that's just a rant for another day.

Actually the QFR PCB is nearly identical to the Majestouch 1 Filco, the 2 is where they went dual-layer.  Carter says CM chose single layer PCB on purpose after testing a bunch of variants.

[kmiller8]

Actually the QFR PCB is nearly identical to the Majestouch 1 Filco, the 2 is where they went dual-layer.  Carter says CM chose single layer PCB on purpose after testing a bunch of variants.

really... hmm. Yeah it was a one sided PCB I tore apart (ripped about 10 traces :/) I hadn't know they specifically chose the one-sided. Like I mentioned in IRC, I destroyed this QFR PCB beyond future use, but the 30 year old Cherry PCB I tore apart at first is still perfectly fine (except for being broken into 5 pieces )

Thanks for the infos.