GH60 prototype betatesting [Call for layouts, GUI testing]

[SmallFry]

Very cool. I can't wait! Thank you once more Komar!

[komar007]

Very cool. I can't wait! Thank you once more Komar!

There's nothing to thank for so far:D

[WhiteFireDragon]

I got a chance to try komar's firmware linked in the OP. It still has a few bugs in the modifier keys. I also flashed it with hasu's hex file and the first layer worked perfectly fine for a poker layout. I don't have switches soldered on yet so I haven't tried the function layer yet.

Also one of the PCBs might have a broken trace, or possible the atmega chip is a dud. The switches that corresponds to keys 8, L, K, and M does not respond. The corresponding diodes for these are D41, D42, D43, and D44. I can't visually see any bad traces though. Still trying to see what's wrong with this area.

Nice write up Whitefiredragon, I only used my iron and did find it quite difficult.  I think the pads on the X1 should be enlarged to allow for soldering stations.

How long did it take you for iron only? When I tried to solder those X1 pads, I had to desolder a few times to reseat it. I eventually got it, but I wasn't 100% sure all the pads made contact, so I ended up just using hot air to be safe.

[komar007]

I got a chance to try komar's firmware linked in the OP. It still has a few bugs in the modifier keys. I also flashed it with hasu's hex file and the first layer worked perfectly fine for a poker layout. I don't have switches soldered on yet so I haven't tried the function layer yet.

Also one of the PCBs might have a broken trace, or possible the atmega chip is a dud. The switches that corresponds to keys 8, L, K, and M does not respond. The corresponding diodes for these are D41, D42, D43, and D44. I can't visually see any bad traces though. Still trying to see what's wrong with this area.

There's probably a problem with column 9 - but the keys that shouldn't work are: 8, I, K, M (I instead of L...).
Please check the trace from pin 8 of the microcontroller. Maybe it's broken or there's a cold joint on it.

[WhiteFireDragon]

Ah sorry that was a typo. L is fine, it's 8, I, K, M that are not responding. I reflow everything and even replaced one of the diode to check. It must be a trace or the atmega chip. I don't see anything wrong with the traces connecting these switches, so it's probably near the atmega pin. Which one is pin 8 of the atmega?

Btw I have 4 built, and only 1 of them have this problem.

[JBert]

I wonder if there's anyone going to beta-test the ISO layout... Or will komar007 build one of each anyway?

[komar007]

I wonder if there's anyone going to beta-test the ISO layout... Or will komar007 build one of each anyway?

I'll check if the switches fit where they should, but won't build one. I hate ISO:)


EDIT:

Ah sorry that was a typo. L is fine, it's 8, I, K, M that are not responding. I reflow everything and even replaced one of the diode to check. It must be a trace or the atmega chip. I don't see anything wrong with the traces connecting these switches, so it's probably near the atmega pin. Which one is pin 8 of the atmega?

Please see the ATMega32U4 documentation (http://www.atmel.com/Images/7766S.pdf, page 3). Check continuity with a multimeter between pin 8 and right sides of all diodes on this net. Maybe the via near D44 is broken for example.

[mkawa]

wfd, are you local to socal?

also can you resend your PM to this account?

[WhiteFireDragon]

Ah sorry that was a typo. L is fine, it's 8, I, K, M that are not responding. I reflow everything and even replaced one of the diode to check. It must be a trace or the atmega chip. I don't see anything wrong with the traces connecting these switches, so it's probably near the atmega pin. Which one is pin 8 of the atmega?

Check continuity with a multimeter between pin 8 and right sides of all diodes on this net. Maybe the via near D44 is broken for example.

Good news is that there is continuity between pin 8 and those 4 diodes. So at least the PCB is ok because that’s not replaceable. I’ll desolder the atmega controller off this one and try a chip from another PCB. If it works, I’ll just buy a replacement chip for the other PCB. No visible cold joints or broken traces anywhere that I can see, so it must be a bad controller chip:




I'm going to change out the USB mini B connector for a micro B. I think all mico B connectors only has 2 "arms" to solder onto the PCB pads, but the PCB currently has 4. The positions might be different so it might not work. Can you verify that these parts will be ok?

USB micro B (SMT, bottom mounted, receptacles)
Atmega32U4-AU

wfd, are you local to socal?

also can you resend your PM to this account?

For some reason, something triggered me to send the PM to your evil twin, mkawa_ (with the underscore).  I just sent you the PM again to the correct account.

[hasu]

I guess you have bridge between pin8 and 9(PB0 and VBus). This will always keeps PB0 high and means switches off,  it makes sense to me.

[WhiteFireDragon]

Yeah but from the pic, there is a very clear separation between pin 8 and pin 9. There is a small chance of a solder bridge behind the pins that the camera didn't pick up, but I don't think I'm going to diagnose this any further. I'll just put a new chip on to see if that was the problem.

[hasu]

Why not use just multimeter
But that pin may be already lactchuped unless you have luck.

[komar007]

I'm going to change out the USB mini B connector for a micro B. I think all mico B connectors only has 2 "arms" to solder onto the PCB pads, but the PCB currently has 4. The positions might be different so it might not work. Can you verify that these parts will be ok?

USB micro B (SMT, bottom mounted, receptacles)
Atmega32U4-AU

I'm afraid all USB micros have the same pin pitch, which is different than the one in USB minis...
I don't think USB micro can be done on this board without redesigning it.


The controller part is ok.
If the chip is indeed broken, that's weird.
Please test it for shorts with other pins, especially VCC, before desoldering.

[SmallFry]

I'm going to change out the USB mini B connector for a micro B. I think all mico B connectors only has 2 "arms" to solder onto the PCB pads, but the PCB currently has 4. The positions might be different so it might not work. Can you verify that these parts will be ok?

USB micro B (SMT, bottom mounted, receptacles)
Atmega32U4-AU

I'm afraid all USB micros have the same pin pitch, which is different than the one in USB minis...
I don't think USB micro can be done on this board without redesigning it.

Noooooooooooo! Heh, just kidding. I don't mind, but USB micro would have been nice.

[JBert]

I wonder if there's anyone going to beta-test the ISO layout... Or will komar007 build one of each anyway?

I'll check if the switches fit where they should, but won't build one. I hate ISO:)

Never mind, Dirge has actually posted an ISO GH60 on Deskthority - http://deskthority.net/photos-videos-f8/komar-60-board-t4697.html - so it's been tested.

[jdcarpe]

Yes, he did, but you wouldn't be able to find it by searching for "GH60." He never calls it by name, only "komar 60% board." I guess GH is so universally hated at DT you can't even say GH60 there.

[WhiteFireDragon]

I shipped out 5 prototype boards today. Each one takes me 2 hours to do, including programming, cleaning the flux, and testing all connection with a multimeter to make sure it all works. I flashed them with hasu's firmware, which will work for the default poker layout (didn't test function layer) so they can be used immediately.

[komar007]

I could have soldered half of those, but I was sure most people wanted them in parts, shame. I also needn't have packed the parts in separate bags.
I hope it doesn't make a problem...

[mkawa]

i was going to assemble, but when i realized both of you had access to hot air i just facepalmed and had wfd do it. btw, got my board today. thanks dudes. unfortunately it's a little low on the todo right now, but i should be able to sit down and get the switches on next weekend so i can hit the codebase and start testing

[jdcarpe]

Received my GH60 in the mail today and promptly assembled it.

Thoughts:

1. I need to reprogram the firmware (hasu's) that WFD loaded on it. I assembled a 1.5 mod layout, and the R_ALT key acts as R_WIN.
2. If using a 7x spacebar with Cherry stabilizers, you might want to install a stronger spring/switch for the SPACE. I used all browns, and the spacebar has a lazy reset.

Other than that, everything seems to work great. Once the rest of the beta testers assemble theirs and test them, I think we are ready to proceed with a group buy.

Great job, komar! And thanks for the SMD assembly, WFD!

[__red__]

Okay - so I have my keyboard.

Who has a suggested first keycap set they can recommend to me.  I want it *fast* :-)  Fast and perty.  Who / what should I buy?

Also, Pi / Prop laptop screen #1 has now been acquired: http://www.ebay.com/itm/271102649447?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649

[TheProfosist]

go wasd for caps quick and easy pertaing he has all the keycap sizes you want

[Sai]

Received my GH60 in the mail today and promptly assembled it.

Thoughts:

1. I need to reprogram the firmware (hasu's) that WFD loaded on it. I assembled a 1.5 mod layout, and the R_ALT key acts as R_WIN.
2. If using a 7x spacebar with Cherry stabilizers, you might want to install a stronger spring/switch for the SPACE. I used all browns, and the spacebar has a lazy reset.

Other than that, everything seems to work great. Once the rest of the beta testers assemble theirs and test them, I think we are ready to proceed with a group buy.

Great job, komar! And thanks for the SMD assembly, WFD!

Show Image

jdcarpe , it is so beautiful.  :'( :'( :'(

[JPG]

 /drool

[WhiteFireDragon]

2. If using a 7x spacebar with Cherry stabilizers, you might want to install a stronger spring/switch for the SPACE. I used all browns, and the spacebar has a lazy reset.

I'm assuming that you're describing a generic "issue" with a 7x spacebar and cherry stabilizers? It doesn't make sense for this to be specific to only this PCB unless I'm missing something here.

[SmallFry]

Not gonna lie, the space bar looks giant on the small board. I like it. As for WFD's comment, I've never had an issue with my Ergo clears and Cherry Stabs, so I don't know if JDCarpe's problem is isolated or not.

[jdcarpe]

Yeah, basically a generic issue, not anything specifically to do with the GH60. Except that it's the only 60% board I own with a 7x spacebar.

I swapped out the stem and spring for a clear and it's better now. I think many Cherry Corp boards with 7x spacebars use a stronger switch for the spacebar, probably for the same reason.

[__red__]

Non-responsive to USB.  Grrr... working it'

[dirge]

Yes, he did, but you wouldn't be able to find it by searching for "GH60." He never calls it by name, only "komar 60% board." I guess GH is so universally hated at DT you can't even say GH60 there.

I'm just giving credit to the person who's done all the work so far.  I don't hate geekhack, but some of the people are difficult to get on with.  I don't think geekhack is universally hated at deskthority either.  I'm sure if someone posts a thread on there to offer the gh60 it would be very well received.

(original post edited due to me calming down...)

[komar007]

Great to see assembled boards!
I've been working on the programming GUI till 4 a.m. today, but the major part is done.
I've decided to go with tkinter/python. Sorry to all who think it's ugly, but this choice has the best dependencies to implementation simplicity ratio for me.
I also have a working communication library in python/pyusb which is capable of sending the layout to the keyboard in run-time.
If someone wants to play with it before I post the GUI or try to run it in windows, please send me a PM.

[IvanIvanovich]

I don't care if it's pretty as long as it makes sense and is easy enough to use.

[__red__]

Does the ATMEGA32U come with the DFU firmware already on it like the instructions infer?

I'm unable to get the IC to enumerate as a DFU device, I get "Unknown Device" and "Device is malfunctioning".

I've gone over every single connection and discrete and even hooked up the oscillator to an oscilloscope to check that that was running and everything checks out.

Any suggestions as to where to go next?

Thanks


Red

[hasu]

Did you install FLIP and its driver on the device?

I assume you are on Windows.

At first time you start DFU bootloader on Chip 'Found New Hardware Wizard' will come up on Windows.
If you install device driver properly you can find chip name like 'ATmega32U4' under 'LibUSB-Win32 Devices' tree on 'Device Manager'. If not you shall need to update its driver on 'Device Manager'.
You will find the driver in FLIP install directory like: C:\Program Files (x86)\Atmel\Flip 3.4.5\usb\

[WhiteFireDragon]

Maybe check for solder bridges behind the atmega's legs and verify everything with a DMM.

The problem that I whined about in the previous page turned out to be a solder bridge between 2 legs, which was impossible to see by eye or camera because it was behind and underneath the two legs. You can see that blown up pic that I posted earlier didn't even show that bridge, but a DMM verified that short. I just flooded that area with flux and reflow.
 

[komar007]

Does the ATMEGA32U come with the DFU firmware already on it like the instructions infer?

I'm unable to get the IC to enumerate as a DFU device, I get "Unknown Device" and "Device is malfunctioning".

I've gone over every single connection and discrete and even hooked up the oscillator to an oscilloscope to check that that was running and everything checks out.

Any suggestions as to where to go next?

Thanks


Red

I'm sorry to read that.
I once had a similar problem, and changing the usb port helped. I don't know what's wrong with one of my usb ports, but the keyboard didn't work on one of them. Strangely, it worked fine on both in linux.

If you're on windows, check what USBlyzer has to say about it.
If you run out of ideas, try to connect to SPI with an AVR programmer if you have one (or make one, like USBasp) and try if you can communicate with the device.

Electrically, the first thing to check would be the D+ and D- resistors R1 and R2 and the crystal. Since the crystal works, check that VCC is 5V and stable.

You can also hook up stabilized 5V (without the usb connector) to VCC and GND pins under the logo. The one under "o" in ".org" is GND and the one under "pr" in "programmable" is VCC. Check that the device drives about 26.5mA.
Then, while still supplying the voltage, press the programming button. The current consumption should drop to about 10mA and then slowly rise and settle at about 16mA.
I hope that if the chip is broken we can see it by observing the currents.

[WhiteFireDragon]

^ That kind of detailed diagnostic procedure is something I would never be able to provide.

[__red__]

Maybe check for solder bridges behind the atmega's legs and verify everything with a DMM.

Yeah, I walked the pins in pairs (including the corners) and I have no shorts there.

If you're on windows, check what USBlyzer has to say about it.

Connection Status: Enumeration of device failed.

IE, The negotiation hasn't even reached the point where the USBID has been declared so driver installation never happens.

If you run out of ideas, try to connect to SPI with an AVR programmer if you have one (or make one, like USBasp) and try if you can communicate with the device.

Electrically, the first thing to check would be the D+ and D- resistors R1 and R2 and the crystal. Since the crystal works, check that VCC is 5V and stable.

So, funny thing.  Looking at it in more detail the crystal appears to be providing a clean sine wave but at something like 3.2V to 3.4V peak to peak.  I'm suspicious but I don't know what the peak to peak should be and neither the ATMEGA32U or crystal datasheet give me any indication of what I should see there.

That being out of expected range would give my symptoms and I would need to suspect either the oscillator, discretes or the IC itself.

I'm ordering a fresh set of components today as measuring the RC values in circuit is obviously pointless and I don't have the patience to re-use 1c components and attempt to clean solder off the ends :-)

You can also hook up stabilized 5V (without the usb connector) to VCC and GND pins under the logo. The one under "o" in ".org" is GND and the one under "pr" in "programmable" is VCC. Check that the device drives about 26.5mA.
Then, while still supplying the voltage, press the programming button. The current consumption should drop to about 10mA and then slowly rise and settle at about 16mA.
I hope that if the chip is broken we can see it by observing the currents.

I'll try that when I get back from work this evening!

Thanks!


red

[komar007]

I forgot that you don't have the firmware on the chip, and that 26.5mA was when the firmware was working, so I think you should be getting something between 10mA to 16mA more likely.

Here's a screenshot of xtal1:

Note that this is a 60MHz scope and the signal is 16MHz, so it should be a bit higher in reality.
I don't know if this helps, but maybe.

[__red__]

Yeah, I'm pretty sure my p2p range is very different which may very well point to a problem.

Is that pin 16 or pin 17 you measured that from?

Also - nice scope, what is it?

Also, I'm about to re-purchase the BoM.  Any changes from the latest in github that I need to know about?

Thanks for taking the time!

EDIT: Where did you source your ASX3F?  I'm having issues finding a supplier.



Red

[komar007]

Yeah, I'm pretty sure my p2p range is very different which may very well point to a problem.

Is that pin 16 or pin 17 you measured that from?

It's pin 17. Pin 16 looks similar, but p2p is a bit smaller.

Also - nice scope, what is it?

Voltcraft DSO-3062C (aka Hantek DSO5062M, Tekway DST1062B)

Also, I'm about to re-purchase the BoM.  Any changes from the latest in github that I need to know about?

Thanks for taking the time!

I've just pushed, but what was before seems correct.

EDIT: Where did you source your ASX3F?  I'm having issues finding a supplier.

Unfortunately in a local shop.
Please try to find anything of similar size and footprint.
I'm attaching the datasheet:  ASX3F.pdf (221.59 kB - downloaded 341 times.).


EDIT: I think these should be fine.

[SmallFry]

You never stated if you got FLIP and drivers installed, but I would assume you have from the steps you've taken already Red.

[komar007]

You never stated if you got FLIP and drivers installed, but I would assume you have from the steps you've taken already Red.

It shouldn't really matter, the device doesn't enumerate at all:(

[SmallFry]

You never stated if you got FLIP and drivers installed, but I would assume you have from the steps you've taken already Red.

It shouldn't really matter, the device doesn't enumerate at all:(

Oh...whoops. That's a shame...

[__red__]

Some laser-porn...  Prototyping in plywood as it smells like campfires instead of ass:

[Glod]

Some laser-porn...  Prototyping in plywood as it smells like campfires instead of ass:

Awesome!

[WhiteFireDragon]

I recorded two separate prototype soldering, one with hot air only and with iron only.

[alaricljs]

Some laser-porn...  Prototyping in plywood as it smells like campfires instead of ass:

So is that a weird camera angle or is your gantry aligned to a vertical plane?

[__red__]

Weird camera angle.  The camera also appears to be upside down.  Reason for that is that we try to avoid being in front of the laser while it's cutting plywood or MDF since the light is as bright as an arc.

Sure, the polycarbonate screen filters out the laser, but watching it cut or the light reflected from it tires the eyes so we choose not to risk it.

[WhiteFireDragon]

_red_,  you're mounting your prototype board on that "plate" made of wood? Post back how the sound and bottom-out feel compared to a steel plate.

[Glod]

I recorded two separate prototype soldering, one with hot air only and with iron only.

watching your video(s) right now while i'm working