Hacker News Reader

Comment by ant6n | Hacker News Reader

I wonder whether it would be possible to create a tablet/laptop with an e-ink display, where the processing switches between a low-powered ARM and something like an Arduino. When the Arduino is used, the ARM would be turned completely off.

View on HN

> I wonder whether it would be possible to create a tablet/laptop with an e-ink display, where the processing switches between a low-powered ARM and something like an Arduino.

You'd have to deal with different architectures and a lot of headaches, but it's the idea behind ARM's big.LITTLE, where you pair high-perf high-power and low-perf low-power cores. So your tablet's SoC might have a ~200mW A53 for mostly-idle use with a ~1W A72 for more active "high-performance" work loads. And the only additional hardware is the additional cores and the interconnect, the SoC is shared between "low-perf" and "high-perf" cores, which probably wouldn't be the case for an arduino/arm mix.

True. But ~200mW is still a fair amount, compared to, say ~20mW of something like the esp8266.

That's the thinking behind the big.LITTLE concept, and it's in use today. It's nice because in addition to power savings, it doesn't take up as much of the PCB as an ARM+Arduino combo would.

Apple does something similar, uses a ARM M-series microcontroller to do sensor readings (activity detection etc.) when the device is sleeping [1].

[1] https://en.wikipedia.org/wiki/Apple_motion_coprocessors

This does actually somewhat exist now. Most modern phones/tablets/ereaders have a low power core for monitoring and sleep modes. The problem is as soon as you want to 'do' anything you need the performance of the main core. Especially anything with graphics (like eink).

The Arduino is underpowered for anything more than a supervisory role, and even those applications are tending towards Cortex-M series.

I understand the power savings of e-ink go away when you update often eg typing, scrolling.

Maybe that can be worked around? eg if updating only a small area. Or as a second display for docs.

It depends quite a bit on how the screen is updated. If it's like a traditional raster image, then you have to clock out each individual pixel in the image along with some blanking periods. The controller on the display is then responsible for taking that pixel data, syncing on the blanking periods, and drawing lines of pixels whenever data enable goes high.

If the eInk display doesn't use this raster model but rather an addressing model then it might be possible to support partial updates. I doubt very much that it does though.

Eink uses a raster mode, but with (somewhat) tri-state pixels. You can send a 'no change' to any particular pixel. The disadvantage is you still need to scan the entire screen, so it isn't really useful for changing individual pixels, but partial updates are common.

The other difference from a traditional raster display is there is no minimum refresh rate.

Eink controller chips use an addressing model and handle the partial updates for you, so the main processor sees a bare framebuffer.

Note the link I submitted above, where it says that yes you have to send all rows with every refresh, but for every row you can tell it not to do anything. Presumably that means if you skip all but few rows, you'll get a faster update.

Apparently you can do local updates, i.e. update a particular row. At least on the 6.7" e-ink that the Kindle uses:

http://www.essentialscrap.com/eink/waveforms.html

Reply on news.ycombinator.com