It’s been a super long journey but I think it’s finally over. The watch works, the packaging and booklet are amazing thanks to my wife, the limited run is just about fully sold off. I think I’m a happy man. Thank you to everyone who’s read this blog or even just asked about the watch in passing. Thank you to the customers who were interested in the limited run of a pet project I started 4 years ago. I’m proud that people have agreed to give me their hard earned cash for something I’ve made with my own hands.
I don’t currently have any plans to make more of these. The current design would need to be tweaked in many ways to make the whole thing more manufacturable. There are some wonderful people who are helping to take product photos of the watch to perhaps showcase it on Fratello’s online magazine. If that garners a lot of attention and demand is high (and life permits) then I’ll come back with a version 2 of the wireless nixie watch.
Again, can’t thank everyone involved enough, if I’ve talked to you about the watch then you had a hand in making it become a reality.
Emin Vartanians
P.S. Below is a full dump of my project source files. PCBs, code, CAD, everything (sketchup source files were too big so I only included the final step files here) I hope this inspires some other curious person to tinker. Blogs like this were pivotal for me in realizing what is possible and igniting the passion that made this 4 year saga possible. I've worked off of this Google Sheet as a traveler for these watches, documenting everything they went through, you can see quality pictures on there probably. I’ve also listed all the things that made this version of the watch harder to produce. The source engineering has mistakes, this list is the redlines and future revision ideas...
Again, can’t thank everyone involved enough, if I’ve talked to you about the watch then you had a hand in making it become a reality.
Emin Vartanians
P.S. Below is a full dump of my project source files. PCBs, code, CAD, everything (sketchup source files were too big so I only included the final step files here) I hope this inspires some other curious person to tinker. Blogs like this were pivotal for me in realizing what is possible and igniting the passion that made this 4 year saga possible. I've worked off of this Google Sheet as a traveler for these watches, documenting everything they went through, you can see quality pictures on there probably. I’ve also listed all the things that made this version of the watch harder to produce. The source engineering has mistakes, this list is the redlines and future revision ideas...
- Have you tried running two tubes at once using this regulator? Would need a whole new matrixing scheme on the MCU
- Change the tube mounting to connectors. Samtecs are juuuust to tall which is worth it. Shore up the top strip of power board
- Don’t use the foremost anode on tubes since the pin is far fro the rest of the digits. Just use the anode between the 9 and 0 cathodes so you don’t have to pay the cost of a longer bend radius of that front anode into the connector or board
- Inter board headers should be spaced as 2 equal 4-conductor headers with one contact missing. This way all you have to do is take out a contact as opposed to the cutting and glueing that had to happen this time
- Power board battery cut out should be fixed It’s too short right now.
- Charging chip TS should be tied low through 10k
- Find a way to squeeze another pair of pins from board to board using the miso/mosi lines act as CHRG! And anything else so you don’t have to have double duty signals. This double duty cleverness leaders to some chargers not liking the watch. Make sure when the miso/mosi are being use bog stuff doesn’t happen (like 180v enable).
- Fix Q10 problem, its pads are missing on the board!
- Accel is discontinued. Look at FXLN8361QR1
- See about moving accel to bottom in case it needs to be reworked
- Board files have inconsistent track widths… what happened?
- Case faceplate feet need to be thinner to account for tube nipples at the top
- Power PCB sits higher than previously anticipated due to coil + acrylic being thicker so bring up the platform and cut the hold down posts a bit.
- Check battery level as soon as you come out of reset so you don’t end up waiting a minute to turn off display for dead batteries
- Better battery management could be to only wake up once every second if battery is <20% or maybe once a minute below 10% just to keep the time. Gotta change wake up prescalers AND interrupt handel math if that happens
- This whole tuning per SN thing for battery (3v3 ref changes per board) and accelerometer (each chip’s gain and offset seems to be wildly different is annoying. Figure out a way where it just works, or is at least easier to automatically tune instead of custom loads per watch.
- Battery percentage calculation is pretty costly with the polynomial. Maybe split up into linear sections.
final_engineering_files.zip |