Surge protection device failure mode

[newfutile]

I have come across a failed SPD, I thought I would take it apart and post pics for people to discuss.

It appears to work by melting an alloy or type of solder to break the connection.

I guess this is after the MOV decreases its resistance and increases the current.

 

[DPG]

Very simple mechanism by the looks of it. I sort of expected more complication.

 

[newfutile]

Yes I guess I did as well, I'm not sure what's in the potted area on the back, presumably a spark gap, or on a type 2 an MOV, maybe I will take the dremel to it when I get home

 

[pc1966]

Any idea what killed it? Any lightning activity nearby or maybe open-PEN sort of event?

I suspect the potted area has a big-ish MOV disk but would be interesting to see.

 

[newfutile]

No idea what killed it as such but the mov reads similar to the working one at 250,500 and 1000 volts with resistance decreasing with voltage

 

[Lucien Nunes]

Presumably heat from the MOV is conducted to the tab from which the sprung link desolders itself. So rather than detecting an overcurrent, it is actually sensing when the MOV itself has been thermally over-stressed. For a very brief surge event the heat dissipation is adiabatic which gives this type of setup a ballistic response; it can desolder even after the supply has been interrupted if enough energy has been dissipated.

A similar mechanism was used in various electronic applications. For example, some 1970s TVs had power resistors in certain critical positions equipped with a spring-loaded arm that would desolder and disconnect if the resistor overheated, rather than trying to protect the circuit against overcurrent with a very tiny and frail fuse. These could be reset by the service technician once the fault had been found. There's another one right behind me - some Leak amplfiers have a smoothing resistor soldered to the tags in such a way that if it overheats, it desolders itself and drops off the chassis.

 

[DPG]

Presumably heat from the MOV is conducted to the tab from which the sprung link desolders itself. So rather than detecting an overcurrent, it is actually sensing when the MOV itself has been thermally over-stressed. For a very brief surge event the heat dissipation is adiabatic which gives this type of setup a ballistic response; it can desolder even after the supply has been interrupted if enough energy has been dissipated.

A similar mechanism was used in various electronic applications. For example, some 1970s TVs had power resistors in certain critical positions equipped with a spring-loaded arm that would desolder and disconnect if the resistor overheated, rather than trying to protect the circuit against overcurrent with a very tiny and frail fuse. These could be reset by the service technician once the fault had been found. There's another one right behind me - some Leak amplfiers have a smoothing resistor soldered to the tags in such a way that if it overheats, it desolders itself and drops off the chassis.

Which Leak amps have you got? I do like Leak stuff - probably stems from my dad's Delta 70 that he had on the late 1970s.

I've had several different ones over the years (mainly transistor, but i did have a pair of TL/10s) but the prices have shot up lately unfortunately.

 

[Lucien Nunes]

I was referring to a pair of TL12+ that I have had since my teens and have followed me around everywhere I have lived. These have the 120R 5W wirewound smoother. There are also two Stereo 20s one of which is also in the room with me but not in use, a couple of original germanium stereo 30s and a Delta 70 that is currently lacking a few transistors. The PA collection includes an original TL12. Also various preamps, Point One, Varislope, Troughline tuner, but none of those are in use.

I have been trying to think of other components that desolder themselves like the SPD, of course there is the traditional spring-loaded anti-surge fuse, but perhaps others too.

 

[DPG]

I was referring to a pair of TL12+ that I have had since my teens and have followed me around everywhere I have lived. These have the 120R 5W wirewound smoother. There are also two Stereo 20s one of which is also in the room with me but not in use, a couple of original germanium stereo 30s and a Delta 70 that is currently lacking a few transistors. The PA collection includes an original TL12. Also various preamps, Point One, Varislope, Troughline tuner, but none of those are in use.

I have been trying to think of other components that desolder themselves like the SPD, of course there is the traditional spring-loaded anti-surge fuse, but perhaps others too.

Nice collection of amplifiers.

They started going down hill after the Delta series. The 2200 and 2300 were made more cheaply, and also had DIN type input sockets which I don't like. Although saying that a certain very high end amplifier company also seem to love them.

 

[timhoward]

It appears to work by melting an alloy or type of solder to break the connection.

Were you tempted to whip out the soldering iron and join it up again? ;-)
But seriously, thanks for posting those. It's good to know
a) they do actually operate sometimes. I've probably fitted 100 now and never seen one "go red".
b) how they work

 

[newfutile]

This is the first I have seen to fail as well, I wonder if it was not caused by a surge but maybe incorrect soldering, this is only a guess of course, does anyone know what happens to a mov that has been subject to a surge?
I do have a digital microscope in my electronics workshop at hope maybe I have a much closer look.

 

[pc1966]

This is the first I have seen to fail as well, I wonder if it was not caused by a surge but maybe incorrect soldering,

That does not look like a well-soldered joint that has simply melted.

I would also expect some other evidence of overheating like a bit of melting or discolouration of the plastic.

 

[newfutile]

It may not be traditional solder , it’s probably a low melting point alloy of some type ,at some point I will take it to my electronics workshop and use hot air to determine its melting point

 

[newfutile]

I tried hot air which started to melt the case at 250C the alloy did not soften at all, I switched to the soldering iron which was was easily melting my lead/tin solder ar 250C, this stuff eventually started melting at 350C though it was sinking a great deal of heat via the metalwork.
These test were conducted on the good one of the pair I have but I expect there the same.