Welcome to EnviroDIY, a community for do-it-yourself environmental science and monitoring. EnviroDIY is part of WikiWatershed, a web toolkit designed to help citizens, conservation practitioners, municipal decision-makers, researchers, educators, and students advance knowledge and stewardship of fresh water. New to EnviroDIY? Start here


  • neilh started the topic Stability Testing ~ how to do it? in the forum Mayfly Data Logger 1 month ago

    I’m just wondering who might be doing some system stability/reliability testing ? and what their setup is.

    By system stability: I define it as test setup representing real world conditions that can exercise a majority of core code.

    Traditionally, in open source this is such a sqiggy area, that it requires a community of supporters to exercise it.  In commercial organizations, this is typically performed by a quality assurance group that reports to a different level of management than the software designers.

    My core stability test setup consists of a Mayfly 0.5b, with a wifi Digi XB2B, a temperature/humidity sensor AM2320 and a battery 4A, with a uSD, and a modified FTDI monitoring cable.  My 2nd stability system consist of the same but with a Digi LTE modem.

    The modified FTDI monitoring cable is to be able to provide non-intrusive monitoring for system status.

    Mayfly 0.5b is the latest, and for want of any other definition, only the latest revision of hardware should be used.

    Digi WiFi XB2B a local stable wifi  may provide the most reliable access to the internet, and the WiFi provides the ability to manage a gateway error conditions. Big advantage with WiFi is unmetered connection.

    4AHr LiIon Adafruit-354, Charging temperature 0~45C, discharge -20~60C .  I’ve been having problems with 2A batteries in cold temperatures that I can’t explain.

    Pimoroni AM2320 , Adafruit-3721 is so that the I2C subsystem is exercised Adafruit-3721

    For real-world conditions I define it as varying equipment temperature,  0C(-5C) to 40C OK,  varying wireless connectivity, varied solar including no-solar till the battery exhausts.

    This generally looks like this – https://github.com/neilh10/ModularSensors/wiki/Testing-overview

    The other part of a  stability/reliable  system is having a clear set of test objectives with a repeatable build process to produce software, which I believe I have. I always start from no libs and let it pull in everything.

    I am building off 0.27.0  While ModularSensors is amazing, it has a diverse ecosystem of supporting modules and a reliable reference system is the basis of any reliability discussions.


    I had been hoping to get a stability test for 0.27.0  on my standard system over the hols, but after running for a week  of  testing 24-48hours, finding some issues, tweaking , trying again, all I can say is I am running into issues on the basic stability system.

    I did have some pretty good stability tests in the  past (0.25?) , so now I’m questioning those, and thinking how I might check them out again. One big difference from my last tests is its mild winter, temperature drop closer to 0C than before.

    The issues  are the XB2B hybrid is not always responding, and I’m getting some unexpected RESETs reboots.

    The XB2B hybrid seems to get busy, and a series of “+++” do not result in a response. I’ve solved some of them with some strategic delays(), but I think I need to explore TinyGSM 0.10.9

    I’m getting periodic resets ~ at this point I can only think it is the WatchDog kicking in. However since it doesn’t advertise when its going to bite, I need to do some work on this and also how to figure out what caused the RESET (read MCUSR early) . However after reset, then the XB2B hybrid is not able to find the WiFi SSID  sometimes for a couple of hours. If I turn off the power, and then power it back up it immediately connects.

    The link to the data.monitormywatershed.org often doesn’t get a response over many Posts. I’m sort of assuming this may mostly on the MMW side, but could also be the XB2B links not always going up, so trying to figure out a reliable configuration. There maybe something happening with MMW response, that is then making the XB2B busy such that it doesn’t respond to “+++”

    So really just putting this out there to see if anybody else on a defined release has got a setup that they are getting some good longterm results with?  🙂