F104+SSK+122+62+77+50+Ergo orders now open! New Kishsaver+Industrial Model F Keyboards

[pandrew]

@Runslikeapenguin

The 0s in signal level monitor mean that you have a row shorted to ground.

The fact that it's intermittent means that most likely the ribbon cable solder joints are poking through the tape that's on them, and shorting to the back case.

Open up the keyboard, and put extra isolation on the row of 30 solder joints on the back where the controller is soldered to the ribbon cable.

[Runslikeapenguin]

@Runslikeapenguin

The 0s in signal level monitor mean that you have a row shorted to ground.

The fact that it's intermittent means that most likely the ribbon cable solder joints are poking through the tape that's on them, and shorting to the back case.

Open up the keyboard, and put extra isolation on the row of 30 solder joints on the back where the controller is soldered to the ribbon cable.

Just a piece of Kapton or super 33 tape?

[pandrew]

Just a piece of Kapton or super 33 tape?

I guess in this case it doesn't really matter, any insulating tape will do. And you don't really need the high temperature resistance that kapton tape is famous for. I'd use kapton tape because that's what it's already supposed to come with. Also kapton tape leaves less adhesive mess than other tapes I have. In case I need to remove it in the future. I'm not familiar with super33.
Or you can choose whichever feels like it's more resistant to mechanical piercing to make sure this doesn't happen again.

You could also try something like putting a thin sheet of plastic inside the back case, to prevent shorts.

[Runslikeapenguin]

Just a piece of Kapton or super 33 tape?

I guess in this case it doesn't really matter, any insulating tape will do. And you don't really need the high temperature resistance that kapton tape is famous for. I'd use kapton tape because that's what it's already supposed to come with. Also kapton tape leaves less adhesive mess than other tapes I have. In case I need to remove it in the future. I'm not familiar with super33.
Or you can choose whichever feels like it's more resistant to mechanical piercing to make sure this doesn't happen again.

You could also try something like putting a thin sheet of plastic inside the back case, to prevent shorts.

Pandrew. I type this reply to you from the model F. a single piece of Super 33 seems to have done the trick. many thanks!

[Ellipse]

Thanks again pandrew for your help.

It seems that as a general guideline, if the row or column reads 0 in the pandrew diagnostic tool's Signal Level Monitor, then there is possibly a short to ground and it is best to add a couple layers of insulating tape (polyimide/Kapton, electrical, Super 33, etc.) to the bottom of the controller where the ribbon cables are. And if the row or column reads a value higher than 0 (example: 110) but does not respond to key presses, then maybe the row or column needs more solder applied to the through hole, on the controller and/or on the capacitive PCB.

I have updated the manual - troubleshooting section.

[pandrew]

Here's an incomplete list of some things I can think of right now:

• If a single row reads all zeros in signal level monitor, and it's intermittent, or it is one of the "inside" rows in the order of pins, then it most likely just needs extra insulation.
• If it's one of the outside rows, and it's not intermittent, it could still be an insulation issue, but it could also be a short to one of the neighboring GND pins.
• If you only have rare issues, where you can't identify anything weird in signal level monitor, it could still be insulation issues. Imagine the solder joints have poked through the existing insulation, and when you're typing on the keyboard, you're putting more pressure on it, and only while typing you're getting a short. So I guess it would be a good idea to always start by recommending to add more insulation if there are any issues at all, especially with keyboard types that are known to have ribbon cable pins very close to a metal case.
• You will almost never see only a column showing 0. I can't imagine any naturally occurring situation in which case this would occur.
• If an entire row is showing 0, but it's showing 1023 on a single key(or 4095 on a single key on TH-based controllers), then a column is shorting to a row. I've never seen this, and it's unlikely to happen, but I guess it's worth mentioning, as it could technically happen in some custom built keyboards.
• If you see a column that doesn't react to any keypresses, but has a 100-ish signal level (or 400-ish on TH-based controllers), it could be that column is broken somewhere, or shorted to a power rail. In this case you will likely also see this column slightly lower than the unpressed state of other working keys, cause the other keys have more opportunity to interfere column to row on the board. ( like you can see in markvanlan's post here: viewtopic.php?p=499107#p499107 ) Problems with columns can be fairly easy to debug with column tester and a multimeter. Set the state of the column in column tester, and measure on the two ends of the ribbon cable. (voltage to ground). In the end you'll probably resolve it with reflowing the solder joints.
• If you see a row that doesn't react to any keypresses but has a 100-ish signal level (or 400-ish on TH-based controllers), then it could be a broken row connection. You'd likely also see this row have a slightly lower signal level than the unpressed state of other rows. This can be debugged with a wire, and manually shorting the row to gnd/+5V. You expect to see 0 when shorting to GND, and 1023/4095 when shorting to +5V. You would try seeing for example if it works at the controller pins, and then moving to the sense pcb pins, if it doesn't work on the sense PCB but works on the controller, then clearly the problem is somewhere in-between. In the end you'll probably resolve it with re-flowing the solder joints.
• If some keys on a row/column work, and some don't, then it could be caused by flipper/spring issues, or it could be a broken trace on the PCB. Identifying which keys are bad can help with some debugging to figure out which one it most likely is.
• If two keys are equally affected by a single key-press, it could be two columns or two rows shorted. You can figure out by listing which pairs of keys are behaving weird. (by equally affected I mean looking at signal level monitor you will see similar signal levels. I don't necessarily mean they will press-release at the same time, cause different calibrated thresholds could make the pairs of keys look different)
• The following are possible controller electrical issues:
  • If it doesn't show up on the PC, or it's showing a USB error message:
    • Try replacing the USB cable. It happens very often that there's an issue with the USB cable. Beware also of unpowered USB hubs, USB extenders, and front ports on desktop PCs, which are really extended from the motherboard, and quality can vary. (I'm not saying don't use these as long as it works, I'm just saying that if you're having an issue, confirm with a non-hub backside port, that the issue is not caused by the port you're using)
    • Try applying some compressive pressure on the usb port on the controller, in case there was a soldering issue with it.
    • Components that could be at fault: the crystal, the reset resistor, the usb data line series resistors, the microcontroller, the USB port, the caps attached to the crystals, there could be shorts between closeby pins of these. Most of these you could check with a multimeter. The crystal you could check with an oscilloscope.
    • Some of the USB-C controllers might not work with smart USB-C-to-USB-C cables, so try a USB Type A to Type C.
    • If you programmed the wrong firmware into the keyboard, especially for example if you wrote an atmega32u4-based firmware into an atmega32u2-based device, or vice versa, then it won't be able to enumerate on USB. Follow the steps to put the keyboard into bootloader mode with the manual shorting method, and flash the correct firmware with QMK Toolbox, or avrdude, etc...
  • If it shows up qmk toolbox as if you would have sent the keyboard into bootloader mode always, then there could be a problem with the HWB pin being shorted to ground.
  • All keys showing 0 signal level, or 1023 signal level (or 4095 signal level on the TH-based controllers) Means that there is something wrong with either:
    • The DAC, or communication with the DAC. Maybe some communication signals are shorted to each other or to ground, or they could be broken.
    • The resistors on the output of the DAC generating the reference voltage.
    • The biasing voltage divider
    • This situation can be further debugged with a multimeter
  • All keys show kinda okay unpressed state, like 100-ish (or 400-ish on the TH-based controllers), but no keys react to any keypresses, or half of the columns don't react to any keypresses (or 1/3 or 2/3 of the columns in case of beamsprings): In this case there could be a problem communicating with the shift registers. Can be further debugged with a multimeter, and the column tester in util.exe
  • If a particular row is always 0 or always 1023, or always 4095, or maybe just showing random numbers, there could be a problem with a particular row's analog circuitry. Something could be shorted to some power rail, or could be left floating. Can be further debugged with a multimeter.

[Ellipse]

As a small update - as part of the current batch of keyboard parts in the container ship I ordered some USB-C braided cables in red and gray - now available in the store. These are only available as additional cables, not as replacement cables for the ones included with the keyboard (the included ones are all black; all cables are 3 meters in length). https://www.modelfkeyboards.com/product ... ard-cable/

[markvanlan]

I really hope I didn't ruin my board.. I was trying to flash the default layout back onto the board and I used the [qmk configurator](http://35.164.28.200:5000/) site to compile and download the hex. I used the "universal" file rather than the "wcass" recommended in the manual (damnittttt). After flashing with that firmware, the keyboard doesn't work at all and isn't recognized by pandrew or qmk toolbox.

Is the best course of action to follow this guide?? https://imgur.com/a/kth6dHt#O0IArSE

I'm really shocked flashing that firmware completely took this board offline...

[Ellipse]

That post is out of date - I just removed it. Please follow the manual on the project web site. The easy fix is to short the prog pads with the procedure shown in the manual if the utility does not work. Check what the device name is in Device Manager when it is plugged in to the computer - if it says ATMEGA32 DFU then it is in bootloader mode and you can load new firmware with Atmel Flip.

[markvanlan]

awesome, thank you Ellipse!

[markvanlan]

I continually tried shorting the connection between the pads, plugging in the USB cable, and immediately/very shortly after removing the short. Not seeing the device in the Device Manager (not 100% where to look in here), but also Atmel Flip doesn't seem to see a controller. I'm getting slightly worried over heree!

[pandrew]

Don't worry. If all you did is flash the wrong firmware, then you didn't destroy anything. You need to short the PROG pads while plugging the device in. You need to find something low resistance to short the PROG pads with. You might think that a screwdriver is gonna be enough, but due to surface oxidation it is likely not gonna be good enough. I like to use sharp metal tweezers, because they can dig into the pad a little bit, breaking up the surface oxidation, and I can actually be sure it's shorting. Apply constant strong pressure on the pads while plugging in the keyboard to make sure they're perfectly shorted. If you hear the windows USB plug-in sound (assuming the sound effect is enabled), then you know you have succeeded.

I've even seen people tell me they tried everything to short the pads with, and then I asked them to solder the pads together and it suddenly works. There's something about the size of those pads, or about the fact that they need to be shorted in the perfect moment when the chip leaves reset, that make people think they are perfectly shorted and yet they aren't. This affects me too, sometimes it takes me up to 3 tries to get it into bootloader mode.

Another thing to double check that you are indeed shorting the PROG pad and not the reset pad. I believe in some versions of the wcass controller there is a reset pad on the opposite side of the board. So don't be like 'I see some pads, those must be it', make sure to be shorting the correct ones. (I lost a couple of debugging hours in the past to someone shorting the wrong pads, that's why I'm accentuating this point too.)

When flashing beware that I think atmel and qmk toolbox drivers might not be compatible (not 100% sure about this, since I don't use windows, but often it's like that, the open source tool often needs its custom driver), so depending on which one is picked up by windows, only one of the two tools will work, and you might need to update the driver in device manager to switch to the other tool. So if you used atmel flip to flash the last time, and since then you have not newly installed qmk toolbox, (giving it a chance to overwrite the driver), then keep using atmel flip.

[pandrew]

A couple more notes:

-On the wcass PCB the PROG pads are on the same side on which the components are. If you don't see chips while shorting pads, then you're shorting the wrong pads.

-The universal controllers use a different incompatible chip (atmega32u4 vs the atmega32u2 of the wcass). Atmel has not developed a way to encode the chip the .hex files are for in the .hex file, so there is no way for the flashing process to autodetect this. Also, other tools like qmk toolbox can't really innovate much on this front, cause then they'd become incompatible with existing .hex files. Unfortunately the situation is that you can flash anything onto you atmega that is in intel hex format, doesn't even have to be AVR code.

[Ellipse]

Thanks pandrew - this is good to know. Often it is tricky for me to short prog so your advice will be handy.

I wanted to share a nice sound test video of the new Model F with and without the solenoid, posted by someone on Twitter - I think this video captures the sound well.

https://twitter.com/i/status/1467896098861817858

[ironicmoustache]

So I swapped a couple of springs to get rid of buzz (on K and M) and, in the process of making sure I got it right and that everything worked (it did), noticed that both of those keys feel a bit heavier to press/have slightly more travel before the springs buckle. And then I noticed that there's a small "area" around there that has ever-so-slightly heavier springs/keys.

Not significantly heavier, mind you; I forget it when I'm blazing through an article, but my stupid anxiety-curiosity does bring it back to the forefront of my thinking from time to time, haha.

I tried installing another spring (this time with the tweezer method) to see if I messed up the spring install and it felt exactly the same. I've had no such issues with a couple of other spring swaps/flips in other areas of the board, so I'm wondering if it's just a... thing, for lack of a better term, that there can be a small island of slightly heavier keys/extra travel?

Trying to get enough coins together to test more "scientifically", so I'll report back once I do. But thought I'd just put this out there first.

[robo]

I wanted to share a nice sound test video of the new Model F with and without the solenoid, posted by someone on Twitter - I think this video captures the sound well.

https://twitter.com/i/status/1467896098861817858

I'm sure I've said it before but I'll say it again - people putting solenoids into their keyboards are completely nuts
How about an electronically controlled air horn that blasts with every key press? You heard it here first!

[Sheepless]

I wanted to share a nice sound test video of the new Model F with and without the solenoid, posted by someone on Twitter - I think this video captures the sound well.

https://twitter.com/i/status/1467896098861817858

I'm sure I've said it before but I'll say it again - people putting solenoids into their keyboards are completely nuts
How about an electronically controlled air horn that blasts with every key press? You heard it here first!

You know, most people would say that mechanical keyboard collectors are completely nuts. So, if you're already crazy, why not go the extra mile and make one of your keyboards sound like a typewriter?

[I0IParzival]

It is a good way of making sure you are not asked to fix the code over a videocall with the client, just unmute yourself while typing and you do not have to worry of interruptions or people making comments about it

[Bschuster]

I wanted to share a nice sound test video of the new Model F with and without the solenoid, posted by someone on Twitter - I think this video captures the sound well.

https://twitter.com/i/status/1467896098861817858

I'm sure I've said it before but I'll say it again - people putting solenoids into their keyboards are completely nuts
How about an electronically controlled air horn that blasts with every key press? You heard it here first!

Yooooo, someone found my video.

So, when I bought this keyboard my goal was to build the loudest, heaviest keyboard I possibly could. Solenoid was an obvious choice.

And yes, it’s so loud it’s actually distracting. If I’m on a livestream and start typing with it on my comments will blow up. I also can’t watch tutorials and code with the solenoid on.

However, for solo coding or writing, there’s nothing like having the hammer of fucking Thor striking down defiantly as every keystroke rumbles into my computer.

[dddyrseth]

How many Watts of power does the F77 need?

I'm going to install a rack of AA- or AAA batteries inside the case.
Also a Bluetooth transmitter connected to the Keyboards logic board(maybe soldered to the USB connections.

Someone probably posted this here already, but I'm just going straight at it.
Any ideas? Any help? Anyone who tried- failed or succeded, send me the link to their post please

Thanku TYPERS

[Bjerrk]

I'm going to install a rack of AA- or AAA batteries inside the case.
Also a Bluetooth transmitter connected to the Keyboards logic board(maybe soldered to the USB connections.

Just curious: what's the need for Bluetooth in a board that is essentially too heavy to move around?

[Muirium]

Switching between multiple hosts with a keystroke can be very handy, *if* you can pull it off.

[dddyrseth]

I'm going to install a rack of AA- or AAA batteries inside the case.
Also a Bluetooth transmitter connected to the Keyboards logic board(maybe soldered to the USB connections.

Just curious: what's the need for Bluetooth in a board that is essentially too heavy to move around?

If you want to have the keyboard on a desktop-rack in bed for example, or on your lap for a chance, or the fact that it's less cables.

[dddyrseth]

How many Wattages does the F77 pull, good sir?

[mmm]

Just curious: what's the need for Bluetooth in a board that is essentially too heavy to move around?

I find it nice not having to bother with cables on my desk (in theory, I manage to keep at least a few unused cables at my desk anyway) for a cleaner work environment. Also it's quite nice not having to deal with any cables or adapters, when just wanting to type something out from another location than my desk.

"But the model F uses a USB cable, you don't need any adapter!" he yelled out in excitement.

Not all of us can afford the luxury of full-sized USB ports on our laptops. These are trying times

[robo]

However, for solo coding or writing, there’s nothing like having the hammer of fucking Thor striking down defiantly as every keystroke rumbles into my computer.

I get that this is 100% subjective and what follows is purely my own opinion, so I just post it here for conversation, but to me there's a line between the natural sound made by a key mechanism (including how it reverberates in the body of the keyboard), and connecting an entirely separate electronically controlled noise-making device. One is inherently part of the keyboard and its feel, the other is... a sort of weird kind of auditory skeuomorphism. The solenoid in the case is supposed to fake the sound of an IBM typewriter mechanism, and I get that there's a historical significance to it since IBM really did install these in computer keyboards to reassure typewriter users in the early days, but it's still weird to me to want to have that.

There have always been little desktop computer programs that could make typewriter noises from your speakers as you typed, and they are all basically annoying after a minute or two. The solenoid or other noisemaker seems exactly the same to me (aside from probably better latency).

^ 100% my own subjective opinion that is not meant to invalidate anyone else's fun ^

[Ellipse]

I wonder for the solenoid users if it can help increase typing speed/accuracy in a noisier environment? You would hear when each key signal is sent in a noisier environment.

[jsheradin]

How many Wattages does the F77 pull, good sir?

Out of curiosity I broke out a current probe and recorded 60 seconds worth of typing.

Spoiler:

IMG_20220225_212440.jpg (490.3 KiB) Viewed 6858 times

DS1Z_QuickPrint2.png (49.56 KiB) Viewed 6858 times

Looks to be about 1.75W with no real difference between idle or use. My measurement gear's not really good for the mA range so not sure how accurate this is.

[dddyrseth]

How many Wattages does the F77 pull, good sir?

Out of curiosity I broke out a current probe and recorded 60 seconds worth of typing.

Spoiler:

IMG_20220225_212440.jpg
DS1Z_QuickPrint2.png

Looks to be about 1.75W with no real difference between idle or use. My measurement gear's not really good for the mA range so not sure how accurate this is.

Thanks!

[dddyrseth]

Does anyone know of any F77 compatible Keyboard controllers with Bluetooth and USB-C?