Hacker News Reader

Comment by zokier | Hacker News Reader

Why use 6 chips when you can use one? It's not like those discrete chips would be cheaper in total

They probably would be, actually. In quantities of 1, a discrete logic chip is around 50 cents, while that microcontroller is $2.50. In higher quantities the logic gets cheaper faster. For this project, you would need:

1 quad NAND gate

1 quad NOR gate

1 quad XOR gate

1 hex inverter

The cost isn't the main thing here, though. A better reason to go with discrete logic is that the MCU also requires an entire software and hardware toolchain to create and load firmware.

Never underestimate the cost of adding nonvolatile memory to your system.

> In quantities of 1, a discrete logic chip is around 50 cents, while that microcontroller is $2.50

The article also mentions that ATMEGA809 should work here, and costs about $1.53 in singles https://www.digikey.com/en/products/detail/microchip-technol...

> I built two prototypes, one with an AVR128DA48, and one with the older ATmega4809 which is pin compatible, and their performance was indistinguishable. The program only uses about 1500 bytes, so an ATmega809, ATmega1609, or ATmega3209 should be fine

For the record, you could also use a much cheaper MCU like e.g. an STM32F030C6 which costs around $0.60 for a single chip and around $0.45 for 30 chips at LCSC the moment. Of course this is completely irrelevant if you build a small number of boards like here, because then you'll probably waste most time with programming the chip and you need the necessary programming connector + tools too.

It would be fun to redesign this board for the lowest manufacturing cost with machine assembly and compare it to how much more expensive it would be to use discrete logic chips.

For me, the ideas become a little more concrete when I see a discrete package and can exactly map these inputs to that output.

Otherwise, why use one chip when you can use none? This could all be little virtual switches in a web app.

> Otherwise, why use one chip when you can use none?

Because you're actually designing a hardware product to sell.

In this case the behavior is not equivalent, I know the author said it's close but with polling the inputs I doubt it would replicate various feedback configurations faithfully.

It would indeed disappoint the user hoping his CMOS fedback inverter gets faithfully destroyed.

Whereas less destruction of electronics -> more faithful "experimenting with and learning about logic" as per the device's purpose.

You could also use 0 chips and create an app that does the same thing.

I think you'll find an app needs more than zero chips to run...

In theory, if you have the source code (it is the easiest), you can run the app in your head, so no chips needed.

Those 6 discrete chips will probably use less power.

Reply on news.ycombinator.com