Sonoff Pow crash with high-wattage appliances

[xoseperez]

Originally reported by: Dave Arter (Bitbucket: davearter, GitHub: Unknown)

---

Thanks for the espurna firmware project, it's awesome. After reading your blog post about the Sonoff Pow I thought I'd give it a try on some Pows I've got here.

I understand the Pow support is very experimental but thought I'd let you know of a crash I'm seeing in case it's helpful. When running a device such as a toaster or kettle via the Pow the firmware crashes soon after switching the appliance on. The Pow restarts and subsequent power output readings are all 0, despite the appliance being powered on.

I managed to capture the serial output when the crash occurs, which I've attached. I've removed the device id and tidied up the whitespace/linebreaks but other than that it's as it appears.

I'll keep experimenting to see if I can find the threshold at which this bug is triggered.

[xoseperez]

Original comment by Dave Arter (Bitbucket: davearter, GitHub: Unknown):

---

For more context, I ran the stacktrace through xtensa-lx106-elf-addr2line and this was the result:

epc1:
0x4021ab8c: hlw8012_cf_interrupt() at ??:?

stack:
0x40107218: interrupt_handler at ??:?
0x401009c2: ppEnqueueRxq at ??:?
0x4010078f: ppProcessTxQ at ??:?
0x401071e0: interrupt_handler at ??:?
0x4000050c: ?? ??:0
0x400043a8: ?? ??:0
0x4021ab98: hlw8012_cf_interrupt() at ??:?
0x4010727c: interrupt_handler at ??:?
0x40107266: interrupt_handler at ??:?
0x4000050c: ?? ??:0
0x401048f9: ets_timer_disarm at ??:?
0x4000050c: ?? ??:0
0x401071e0: interrupt_handler at ??:?
0x4000050c: ?? ??:0
0x40000f68: ?? ??:0
0x40000f58: ?? ??:0
0x400042db: ?? ??:0
0x40004b31: ?? ??:0
0x40104e5e: spi_flash_read at ??:?
0x401076b8: pvPortZalloc at ??:?
0x40222105: pm_set_sleep_time at ??:?
0x402225a6: pm_get_sleep_type at ??:?
0x40222653: pm_get_sleep_type at ??:?
0x40203648: delay_end at ??:?
0x40225325: ets_timer_handler_isr at ??:?
0x4022536a: ets_timer_handler_isr at ??:?
0x40202e83: loop_task(ETSEventTag*) at core_esp8266_main.cpp:?
0x40000f49: ?? ??:0
0x40000f49: ?? ??:0
0x40000f49: ?? ??:0
0x4010f000: ?? ??:0

[xoseperez]

Hi

Thanks for reporting this!
Can you copy the code you are using to test the library? Also: what's the nominal power of the toaster?

[xoseperez]

Original comment by Dave Arter (Bitbucket: davearter, GitHub: Unknown):

---

The toaster is marked as 1000W on the bottom, and the kettle is 2750W. I'm using the standard espurna master branch, deploying to a Sonoff Pow with platformio run --target upload -e sonoff-pow-debug

[xoseperez]

I have managed to reproduce it with a ~1kW load but I'm still debugging.
In the meantime I have changed the code to use non-interrupt driven mode in the HLW8012 library and it works pretty well.

[xoseperez]

Version 1.4.4 uses latest HLW8018 library that fixes some volatile variable assignations. Back to interrupt driven approach runs now fine with 1kW loads, tested over the last 24h.

[xoseperez]

Original comment by Alexander Christian (Bitbucket: alexander_christian, GitHub: Unknown):

---

I'm using just the HLW8012 lib with my own sketch, which includes a webserver presenting the measurements...

When using interrupts and having "nothing" attached to the POW (=no load at all), I face the same stacktrace when opening/loading the webpage.
So I guess it's not really load-related, but rather a kind of interrupt issue. Maybe in combination with the webserver.

If I switch to non-interrupt-mode, there is no crash. But Having precise measurement without interrupt will block the webserver :-(

So I'm a bit lost... :-(
As this is not related to espurna, I will open an issue on the HLW lib itself.

[xoseperez]

Original comment by Alexander Christian (Bitbucket: alexander_christian, GitHub: Unknown):

---

Link to the issue on HLW lib: https://bitbucket.org/xoseperez/hlw8012/issues/1/esp-crash-when-using-interrupts-webserver

[xoseperez]

Fixed with HLW8012 library update