Oh yes. More of this, please.
Spectacular! This would be great even if the switches did nothing at all.
Wouldn’t this be a good application for a touch panel, as much as I dislike them?
…after looking closer, wouldn’t. The switches provide the tactile experience no fondleslab could ever compete with.
For the laser etching, he did it somewhat wrong. The right way that worked for me is to spraypaint the acrylic (like he did), cut the panel outline and the holes (like he did) - from the nonpainted side, then flip the panel and engrave through the paint. That way the paint is on the opposite side of the acrylic and is protected against chipping and wear.
Bonus points for the Weyland-Yutani logo.
Thought for the control system. If the MQTT pubsub system does not have too much of a delay in the messages, it could be implemented for control of stuff - every button press or switch toggle would send an event to the broker so the other devices that are attached could take action. A generator of USB HID events could be attached to the broker as a listener, which would allow mapping panel events to HID events - on several different machines.
WS2812 RGB LEDs (“Neopixels”) could be leveraged for color feedback on the panel.
Old cellphones could be also used for displays and touchscreen controls. (Possibly a rooted android, coded to run a browser on startup that’d load a webpage with the controls and websocket-based realtime displays?)
Today I learned a beautiful new word.
Glad y’all seem to like my panel. Lemme know if you have any questions about it.
Not only that, but you can’t use a touch panel without looking. Where as this you can remember the location of switches and preform actions by memory.
Etching wise, you’re saying the difference is that you did the cut from the flip side? You still have the paint on the outside of the final panels?
THAT. A very important part of the user interface designs is the consideration for tactile feedback and muscle memory. This is something that is leaving the field of consumer electronics fast, and I think there was an attempt to make it go even from industrial control systems but then the operators were making too many mistakes which luckily hindered the adoption of such abominations.
In my designs (prototype instrument panels) the paint is on the back, on the inside, behind the acrylic. I work with a cheapo Chinese K40-III laser cutter, a staple of cash-strapped hackerspaces. My experiences are that the paint engraves perfectly, but generates smoke and plasma during cutting, which absorbs the beam. So it is better to cut from the acrylic side, and engrave from the paint side.
Having the paint on the inside has the advantage of being less susceptible to damage, easier to clean, and the grooves aren’t subject to accumulation of dirt.
On a side thought, there are some applications out there for e.g. displaying computer load and other variables on an external display. You could leverage that for one of the panels, to have a realtime display there. And/or add a salvaged LCD with attached Raspberry Pi on each side of the console with the switches, and show the stuff there. I had something similar for supervising my company’s computers and their logs.
Awesome! How thick is your acrylic? Mine’s like half an inch…
I’m looking forward to adding rotary dials on the next version, but LCD and Rawsbetty Pi stuff is still way out of my league.
I usually work with 2 to 2.5 mm. Cheaper, cuts well on my machine (I don’t think it could handle the 1cm or more at all, at least in reasonable number of cut passes).
Arduino-class microcontrollers are good for interfacing of potentiometers, which opens door for both knobs and sliders. If you can source pots from e.g. old oscilloscopes, you can even have two in one - an outer and inner knob. 3d-printing mechanical adapters with simple gears would also do a good job here. Another thing to interface are the incremental encoders; two switches that switch on and off in a phase-shifted way. Excellent to interface via interrupt pins, doable over pin polling.
Knobs can be also easily improvised from old mice. They have two smoothly running rotary encoders, and output data in PS/2 format that can be converted to USB with off-the-shelf adapters. Newer mice are all optical, but that would just give you only one sensor and require a 3d-printed adapter to translate the controller knob movement to something that moves under the mouse sensor.
Raspberry Pi is a cool thing. A small linux-running computer. Ideal for attaching controlling and controlled stuff to network via ethernet. All sorts of devices then can be attached to it via GPIOs or USB adapters. You could then run an assortment of daemons on it, and e.g. execute a HTTP request (or send a UDP packet, or a MQTT event) on a controller change.
This topic was automatically closed after 5 days. New replies are no longer allowed.
