Hacker News Reader

Comment by numpad0 | Hacker News Reader

Depends. I2C only takes two I/O pins.

That’s only part of the equation. Grid displays require a lot more memory and often power to drive. By contrast, a seven segment display requires 7 (or 8) LEDs and 8 bits per digit on the output stage, 4 bit in BCD, which can be fed via a shift register.

Almost any other display that you will reasonably use in a application that makes sense for segmented multiplexed LEDs will integrate internal controller, so instead of having to refresh the thing somehow you have a thing that is simply an peripheral hanging of some or other bus that just does its thing. And well, such things tend to be cheaper than seven segment LEDs (there is a reason why random almost disposable crap has graphical monochrome OLED panels)

Granted, but as a sibling states, the other factor is firmware complexity. I can theoretically drive a high-resolution display on an 8-bit micro (although it may require with some clever bit banging to get I2C to work at all) but its trivial with a 7-segment display.

And as a bonus, you don't have to involve anymore UI/UX design than to answer the question "what number should go here?"

It'll be microcontroller equivalent of running full GNU/Linux to control an Internet of e-waste tea kettle, but CH32V003 + SSD1306 OLED + passives takes less than 10 parts total, no crystals needed, has plenty sample codes online, and costs ~$2 combined. Depending on the price of the final product, it could make financial or product design sense.

In a medical or aerospace application, that added firmware complexity could be a non-starter. Not from a price, or even e-waste standpoint. Its that proving the thing is safe is a lot harder when its complex. That's the real point of keeping it dumb.

Reply on news.ycombinator.com