Shelly 1PM and power surge

[uberlinuxguy]

The Shelly 1PM (bought through the US store) has a voltage rating of up to 240V, IIRC.  I had a Shelly 1PM installed on a US 220V circuit for a heater (well within the rated wattage.)  In the US, 220V circuits are wired Line-to-Line 110-to-110.  Everything worked great until one day, the power company had a surge roll through that took the 110V lines to 180V.  Line-to-Line that works out to 360V which is well outside the Shelly.  The Shelly 1pm had what appeared to be a catastrophic failure, of the explosive kind.  I am wondering if this is the expected failure mode in such a condition.

[thgoebel]

In most cases, the fusible resistor in the Shelly will trigger and stop the current due to short circuit in the voltage suppressing device (VDR). But in some cases the resistor does not react as quick as possible which leads to a massive short circuit in the Shelly. This short is controllable via the mains breaker only. Like in this example…

[TheRegister Unique]

Was the heater on during the surge? On top of what thgoebel mentioned above (the varistor-based protection against voltage spikes) there could have been a sharp increase in the load (the heater) current ...

[uberlinuxguy]

I used the 1PM to monitor the usage on the heater, it was definitely off at the time.  Another 1PM further down the line on the same circuit didn't fail.  I forgot to mention that.  Judging by the graph of the voltage measurement, it did seem to be a rather instantaneous event. I am glad the main breaker tripped, I think it saved a much bigger mess.

 

This incident definitely gave me a little pause about the IoT things in the house.  Especially the ones on Mains power.  I have also invested in a whole house surge protector which I am kicking myself for not doing to begin with.

[TheRegister Unique]

Normally I would not expect the varistor to cause such explosion when reacting to overvoltage. If it did not reacted fast enough, such a high voltage (360 RMS, over 500 peak to peak) could have caused a catastrophic failure of input electrolytic capacitor (rated for 400V). really hard to tell for sure though.

 

[thgoebel]

Most likely it’s the fusible resistor. If this device doesn’t trigger (or not fast enough) it plays the role of a match: It exhibits ionizing gases which causes arcing. This can be stopped with the mains breaker only. Using a heat shrinkable tubing for the fusible resistor may diminish such catastrophic events. Alternatively, a Picofuse would give better triggering results…

[terae]

Plus devices have 500mA fuse and VDR to catch real power surges. Used here with constant 220-240Vac 50Hz.

[uberlinuxguy]

If I can find some time, I might do a postmortem on the board to see what actually exploded.  I will post pictures if I get around to doing a postmortem.