Noisefloor Project: One firmware for 3 chips.

[maxv]

Hello,

This isn't actually a problem report, but might be interesting for readers of

this forum anyways.

I'm writing a series of posts about how to get a unified firmware executable to

run on the Nordic nRF51422, nRF52832, and nRF52840 chips, with no modifications.

In other words, the same binary file loads and runs on all three chips, deciding

what capabilities it can use at runtime. This is different than the usual way of

compiling one firmware file for each microcontroller, and it may not be very

useful outside of my deliberately simple use case, but I thought it was a cool

challenge anyways.

In the writeup, I discuss the hardware differences and software techniques

required to make this work as well as how to document the requirements.

The first two posts are complete, and I'm partway through the third now. I'll

update this post as I make progress.

Here are the first links:

• Introduction
  https://fromplantoprototype.com/r/noisefloor/index.html

• Boundary Conditions and Peripheral Interfaces
  https://fromplantoprototype.com/r/noisefloor/boundary-conditions.html

I would be happy for any feedback and questions about the posts.

Cheers,

Max Vilimpoc

ps. Weirdly enough, I didn't follow the instructions about the REG0 and

REGOUT0 registers here, but somehow my nRF52840 Dongle still works fine when

debugging it with an nRF51-DK as the J-Link probe. Lucky me!

[Dmitry]

Hi Max,

Very interesting, thanks! Curious about your "top secret" project

Concerning your RSSI sampling time question - it's the same (0.25 usec) for 52832 and 52840. There is a new parameter in 52840 spec - "RSSI settling time after signal level change" (when an internal filter settles down) that is 15 usec.

[maxv]

I could tell you, but then it wouldn't be secret.

Aha, I see what happened with the RSSI sampling. I took the value from the Product Spec 1.0 document for some reason. It's corrected now and the post is redeployed.

Thanks for spotting that!