How I made a 60fps Eink monitor, the Modos Flow
109 points by clean-rope5995
109 points by clean-rope5995
I have met Alex and seen a demo of this at FOSDEM this year, and it looks pretty good.
Same, it felt basically "at last".
I wasn't fully sure because it was still a prototype and the case wasn't done yet but it held a lot of promise. I'm confident that my next e-ink purchase will most probably be something from modos.
I have a dev kit from their earlier Crowd Suppy run, anything y'all want to know?
Good to code on?
It's okay; I do take it outside and use it for that.
There are a few different modes of how it renders. The "fast" (default) mode does B&W, but everything is dithered, which makes text look bad- there are some dots that shouldn't be switched but are. Buttons on the side can switch to "text" mode, which is fast and sharp but just black and white; or my preferred "reading" mode, which does a fast just-black-and-white paint of the changed regions, then does a better re-render with grayscale.
So- not as much syntax highlighting ss I usually have, have to retheme accordingly.
any quantitative power draw measurements? Only specs I could find on it were "high" power compared to "low" of typical epaper.
How's the contrast in bright sunlight?
Have you run into any problems on the software side?
No measurments, sorry, I don't have a USB interposer with a power meter or anything. (Unless you know of a way to read it from a USB port on Linux? I could give it a whirl!)
I take it outside and use it, in preference to my builtin screen. I'm trying to avoid direct light (I burn) but it's absolutely usable.
The one "gotcha" for me was setting the right mdoe; there are a few buttons that can switch between different speed/dithering/grayscale tradeoffs, which I didn't know until poking at it. I haven't tried re-flashing it.
There is a host-side SDK with the idea that you can plug it in to your window manager or what bot and tell it to use different modes for different regions. I tried that and didn't get it immediately working, but haven't spent a lot of time debugging; could be a problem that I personally have.
If the power draw is relatively stable, you can use a usb cable like this to measure the draw: https://www.amazon.com/dp/B09HC4K3KP
What's the power draw like? One of the advantages of them is their zero power draw when not being updated, but I assume at high refresh rates they might actually draw more power than an equivalent low-power/low-brightness/Transflective LCD or MIP (memory in pixel) displays?
this is a really cool display! would be great to see it in a laptop or a tablet, I’m not sure I have a usecase for a small external display on its own.
On the go, you can power Flow at up to 40 Hz with a single USB Type-C cable. At a desk, you can connect additional power and take advantage of its full 60 Hz refresh rate.
this sounds a bit weird to me. I can’t find the numbers for power consumption, but one USB C cable can power an entire high-end laptop, surely a single display can’t consume more?
If I understand correctly (I don't), it's a matter of whether the input has Power Delivery. A desktop monitor with an AC/DC input will output PD at 60W or 100W or whatever. By my laptop won't output PD.
I have a camera that will only charge via PD; my laptop will happily connect to it by USB-C, but won't charge it, because it needs 25W or 30W or something.
That said, yeah, I have other portable monitors in my house that can run off USB-C from a laptop. E-ink needs higher voltages than 5V to clear cells, maybe that boosts the power draw too high at 60Hz?
I'm kind of incredulous about this. From what I understand, the slow frame rate is basically a software limitation, not a hardware one? Why hasn't this become more widely available?
Cost and reliability. The main usecase of eink displays is low power. Having a power hungry high bandwidth controller goes against the main usecase, so less effort is powered into it, having a more expensive controller that your customers don't need isn't worth it then. Driving eink displays harder drives their longevity down, which shifts their value proposition. Eink displays tend to be expensive to begin with, through a combination of patent cost and tricky mechanical design. You could manage to get a Toyota Corolla to go at 150 mph, but you'd have to modify things in the interior meant for comfort as well as remove some safety mechanisms and everything that moves in the car might be done after a single lap of that.
I see, thanks for the explanation.
Do you have a sense for the longevity of the display driving it like this?
No, this isn't just basically software. He had to design a special-case hardware controller to deal with the physical constraints of the pixels, at carefully tuned update rates.
Ultimately, "all hardware is just an algorithm, thus starts with software" is true - but in the case of e-Ink, the controller market has not caught up to the same extent that, say, a DVI or VGA controller market, which has decades to develop.
Since Alex' work is open source, I would expect that this is going to get reproduced at scale, at some point soon ..
That's fair. The hardware controller needs to change. But my point is that the e-ink display itself is capable of doing this type of refresh. It was purely an interface issue.
This is very interesting to me as I’m working on a physical version of Ensō (enso.sonnet.io) and reducing typing latency on e-ink panels is… possible with some hacks/being clever about what to render etc., but it still feels very clunky.
I’d love to get this in a custom size (basically something roughly the size of an Apple keyboard).
If you happen to know how to get one, shout! I’ll take any info/tips
Clearly there's some desire for a high refresh rate, high contrast reflective screen. I wonder why most of the approaches we're seeing are "lets make do with normal epaper or hack it" (like this) rather than manufacturers (perhaps Sharp's Memory Display line?) making 6in or larger reflective LCD panels?
edit: also please electronics industry make a purchasable electrowetting display....
My impression is that e-ink looks more like paper than reflective LCDs do.
yup, though not enough to disqualify reflective LCDs for direct sunlight use (at least for me, based on my Playdate handheld). e-ink is definitely better on that front than LCDs, so if this took off and was robust over long term, that would be extremely cool.
I've been considering making a hand held console in the same vein as the playdate (which uses the sharp memory display) but with one of these st730* displays . I'm still investigating and testing with them, but I think there are people pushing in this direction given how cheap these are.
Ooh. I've never seen those! I need to get better at my aliex-fu, digikey and mouser will only get you so far.