Rapid prototyping
There is a particular kind of optimism that arrives around 7:30 on a Sunday evening in the workshop.
You convince yourself that you are going to systematically test variables, carefully document outcomes, and emerge several hours later with meaningful conclusions and a cleaner process.
And then somewhere around dial number three, I mentally noted that I left my notebook upstairs and rather than pause to go get it, the instant gratification instinct took over completely.
Last night’s session was focused on rapid prototyping. The goal was straightforward enough: improve edge quality on dial cutouts while refining the date window positioning for the NH35 movement.
Over the course of the evening I produced ten dial variants, changing power settings, speed, interval spacing, etc. and dimensions in LightBurn while trying to understand why measurements that looked correct on screen were not translating perfectly into brass.
The encouraging part was that the proportionality largely worked. Marker placement, visual balance, and overall geometry felt coherent. The less encouraging part was realizing how quickly multiple simultaneous variable changes make troubleshooting almost impossible.
Change power and speed at the same time? Then alter kerf compensation? Then nudge the date window dimensions?
By the fifth iteration it becomes very difficult to know what actually improved the result and what simply changed it. The 1.95mm center hole is repeatable and seems to be working well with the hand post. The date window is still a bit off. Spoiler alert, a perfect NH35 dial blank was not achieved tonight.
This is probably obvious to anyone with manufacturing experience, but part of this project is learning those lessons firsthand rather than theoretically.
A few takeaways from the session:
Work one variable at a time.
Slow down enough to document settings properly.
Trust iteration, not instinct.
Measurements inside software are only the beginning of the process.
The desire for immediate progress is incredibly strong.
Keep the stoke high. Slow the process down.
Also, in a cruel and frankly unacceptable development, the old Sonos system stopped connecting midway through the session, leaving the workshop completely without music. Troubleshooting LightBurn settings in silence while the laser repeatedly reminds you that brass is less forgiving than expected is less fun.
Still, ten more dials exist now than did yesterday morning. The start of a good stockpile for painting and pad printing R&D :)
