Anybody remember Flux? That keyboard with the maglev buttons over an HD IPS screen.
Did you want one? Yeah me too.
Since it still hasn’t come to market I’d like to work on a similar capability for Framework.
This is exploratory for the moment as I don’t even know if this can be made thin enough to work. But any ideas, feedback, help would be appreciated.
First I need to figure out the dimensions of the keyboard module.
Then try to decypher how thin the components can be and find sourcing options.
Power demands are also something I need to figure out.
Welcome to the forum.
Why are you looking to make the keyboard for Framework specifically? Would it align with Framework by being an open hardware design and / or open source firmware? Framework Laptop 13 or 16?
Is it possible to have powerful enough magnets for maglev without triggering the lid-closed sensors? Personally, I have no interest in the screen because I don’t look at my keyboard, but I’d potentially be interested in a maglev keyboard if the experience is better or if it eliminates difficult/fragile parts like scissor switches.
The lid close sensor is near the bottom on the headphone jack board in the Framework 13, so I don’t think it will be too much of an issue. The one possible issue I see is the fan, as the magnets in the keyboard could possible interfere with the magnets in the cooling fan’s motor. The cooling fan is directly under the keyboard, so the magnets may have a larger effect on it.
I’m open to suggestions on key alternatives. Response time is important but what most appealed to me was the screen and projection onto the keys combo.
I think the magnetics should be fine. The keyboard controller can filter out the (predictable) noise from the fan, and the lid sensor shouldn’t be that sensitive.
This will likely be the biggest dealbreaker.
But I think you underestimate how sensitive the lid sensor is. Several people have made threads about their Framework going to sleep randomly & it was discovered to be a nearby magnet. Or have commented about it. Placing the Framework on top of another (closed) laptop, a cell phone next to the Framework (magnet in the speaker maybe), etc.
The hall sensor that is listed on all the Framework 13 variants might be configurable, but I can’t find the datasheet for that component.
If it’s a switch, not a sensor, then replacing it with a switch which requires a larger magnetic field might be doable.
I’ve been looking into creating something like the Flux keyboard for the Framework 13. Here is a summary of my initial findings.
The stock keyboard has a 3.8mm height, of which 1mm is the base pcb and 1.4mm is the keycap. This results in a compressed height of 2.4mm and a key travel of 1.4mm.
The smallest magnets I’ve been able to find are all disc/cilinder magnets. I haven’t been able to find any cube magnets smaller than 5 mm. All sizes are available in grades from N30 (“weak”) to N52 (“strong”). Useful magnet sizes (diameter x thickness):
2x0.5mm
1x1mm
1x0.5mm
For the screen we start running into problems. The framework panel is 1.9mm thick which without any room for sensors or a pcb would already result in a maximum keytravel of 0.5mm. With the smallest sensor I’ve been able to find of 0.4 mm, there would be no keytravel at all.
So I’ve been looking at Eink panels, which seem to be around 0.8mm thick and therefore might be usable.
The sensor I’ve been looking at is the TI TMAG5253, which is a tiny sensor and using TI’s magnetic simulator seems to have usable output for the movement the magnets will have. I was thinking of giving each row or column their own ADC and pulsing each sensor on one at a time. At maximum performance (and maximum battery draw) that would give a polling rate of 1.9kHz for rows-wise and 4.8kHz for columns-wise.
So combining all that, it would be something like this:
Keycap with magnet (1.4mm)
Travel space (1.09mm)
Eink panel (0.8mm)
Sensor (0.4mm)
2 layer flexpcb (0.11mm)
Which seems doable, I’m going to order some samples and experiment with some 3d printed mockups.
