Background

The Backyard Hero receivers may be the heaviest use I've given the NRF24L01. Most projects needing RF were either low bandwidth enough to stick with 433mhz or LoRA, long-distance enough to just leverage a GSM/3g/4g connection, or high-bandwidth enough just to use WiFi. But I had this drawer overflowing with ever shape and color of these transcievers so I figure they'd finally get their shot.

And largely it's been a pleasant experience - thanks to the RF24 library, interfacing is fairly simple and for the most part, things just work. But what I'd been experiencing for the last few days is that they dont always work. Something changes in the weather and they start dropping packets... As always, I assumed it was me, in fact I completely overhauled the communication protocol because I figured I was over the packet size and reassembly was goofing up somewhere.. Then I started blaming the decoupling capacitor configuration I put close to the pins... and so on and so on. Seems like whatever I did fixed it, but inexplicably things just broke again, randomly. I was about to implement a frequency hopping strategy (not a small task) when I really did a deep dive into the internet, and it turns out I'm really not alone in my loathing of these gremlins.

The Magic Touch

One consistent behavior that I observed is that some of my chips would have transmit issues, but I would go to touch the chip and they would be resolve themselves. That is, until I took my hand away. So even before researching, I knew there was something odd with the RF bits, namely in the module somehow jamming itself (as I assume my finger was either mildly shoring voltage to drop the intensity of the signal or attenuating it somehow).

It wasnt until I really took a look - all of my modules looked the same, like any other NRF24L01+ breakout... but then I took a known working chip from a project years ago and compared it to the ones I'd taken out and used in my test receivers --

Here's my tried and true chip:

And here is the other:

The two together --

Impostor!! Amazon, how could you do this to me! Turns out that in my grabbing from the bin, I'd only grabbed these imitations. The general consensus is that these knockoffs chips are notoriously flaky, especially at higher-than minimum TX power. Even then, it obviously needed help to actually work. The short story here is that they overwhelm themselves with their power! This guy came up with the idea of an attenuator for the signal in the form of wrapping aluminum foil around the PCB antenna... Wouldnt ya know it, that works!

I need to actually figure out a solution to this - though for immediate testing in-house this is okay. It may push me to directly integrate the NRF24L01+ directly on-board and figure out the whole antenna mess on my own. I think that will mean I get to talk to the FCC though.