esp32 with a pi zero-W "master" ?

[miran2782]

I am running reef-pi v6.0 on a pi-zero-w on my main tank in the living room. Am I able to create a esp32 unit to go on another tank in another room and have that be a "slave" to the reef pi software I have running on the pi-zero or do I need a linux pc to host the reef-pi software for the esp32?

If I am able to do that, is there any limit on the number of esp32 units I could run from the same reef-pi on that pi-zero-w?

Wishing I loaded up on more pi-zero-w units for $5/each a couple years ago at Microcenter.

Thanks!

 

[Sral]

I am running reef-pi v6.0 on a pi-zero-w on my main tank in the living room. Am I able to create a esp32 unit to go on another tank in another room and have that be a "slave" to the reef pi software I have running on the pi-zero or do I need a linux pc to host the reef-pi software for the esp32?

If I am able to do that, is there any limit on the number of esp32 units I could run from the same reef-pi on that pi-zero-w?

Wishing I loaded up on more pi-zero-w units for $5/each a couple years ago at Microcenter.

Thanks!

Absolutely ! It's just the same driver that you can use on any ReefPI.

There might be a limit regarding CPU load. If you want to check a lot of parameters on a few different ESP32 every 10s, you might slowly run into problems. However, I don't quite know the limits, you would need to find that out by running one unit and seeing how much that increases your PI's CPU load.
Another thing might be WiFi stability, although I have no experience with that. @Ranjib for example mentioned once that he runs quite a few systems with quite a few Kasa power strips in the same WiFi, so that might not be critical.

What's more important is that the ESP32 is currently not perfectly stable. With me it crashes and reboots every 5 minutes on average. Reboot is very fast, so that's not a huge problem, some sensor values return as 0 on these instances, however, if I remember correctly. That is something to consider when running a full setup on an ESP32, as it might cause problems on dosing, etc. . E.g. anything that is actuating and time critical that needs to change somewhat reliably.

I have identified where that problems comes from though, so we might be able to fix that in the future.

 

[theatrus]

I am running reef-pi v6.0 on a pi-zero-w on my main tank in the living room. Am I able to create a esp32 unit to go on another tank in another room and have that be a "slave" to the reef pi software I have running on the pi-zero or do I need a linux pc to host the reef-pi software for the esp32?

If I am able to do that, is there any limit on the number of esp32 units I could run from the same reef-pi on that pi-zero-w?

Wishing I loaded up on more pi-zero-w units for $5/each a couple years ago at Microcenter.

Thanks!

Honestly, not any real practical limit you'd run into. You're going to first have WiFi coexistence issues well before any limit on the Pi (e.g., 2.4GHz only, b/g only, limited data rates), but you'd be at 20+ devices before going crazy.

As for unstable ESP32s, a lot of power supplies/regulators are underspecced, especially if the ESP32 falls back in 802.11b mode at max transmit power where its a massive energy hog. Outside of software bugs (which of course absolutely can happen), there isn't an intrinsic ESP32 stability curse especially with new revision processors.

 

[Sral]

Honestly, not any real practical limit you'd run into. You're going to first have WiFi coexistence issues well before any limit on the Pi (e.g., 2.4GHz only, b/g only, limited data rates), but you'd be at 20+ devices before going crazy.

As for unstable ESP32s, a lot of power supplies/regulators are underspecced, especially if the ESP32 falls back in 802.11b mode at max transmit power where its a massive energy hog. Outside of software bugs (which of course absolutely can happen), there isn't an intrinsic ESP32 stability curse especially with new revision processors.

Sry, I should have specified. The ReefPi ESP32 code is currently not perfectly stable, it throws critical errors within the AsyncTCP library. That however, can absolutely be fixed.

Although now that you mention it, the power issue on WiFi changes could be a reason for it. I have posted the error reports here.

 

[miran2782]

Thanks! wasn't sure it running multiple would bog down a pi zero. Probably will have to get a few to give it a try. Just setting up my 2nd tank in another room. I would rather just have a single reef-pi running both. I like the idea of distributed nodes. Might even try to put one in my DIY LED light for PWM to cut down on the number of cables running to it.

 

[theatrus]

Sry, I should have specified. The ReefPi ESP32 code is currently not perfectly stable, it throws critical errors within the AsyncTCP library. That however, can absolutely be fixed.

Although now that you mention it, the power issue on WiFi changes could be a reason for it. I have posted the error reports here.

Yeah, those are actual quirks in the software stack. The ESP32 is complex, and the Arduino libraries are also in a weird state of being on old ESP-IDF versions. I haven't dealt with that framework (ESP-IDF only) so don't have any recommendations.

To the OP, yeah this can cause issues then. The interface here isn't very complex so if its a matter of using a different framework to make it stable its pretty straight forward. Honestly, PlatformIO is probably simpler than managing things in the Arduino space anyway.