Dangerous Prototypes

A friend had an accident, his extension cord (the one with surge protection) caught fire...

new: http://imageshack.us/photo/my-images/528/sweex.jpg/ (http://imageshack.us/photo/my-images/528/sweex.jpg/)
burned: http://imageshack.us/photo/my-images/90/dscn3430e.jpg/ (http://imageshack.us/photo/my-images/90/dscn3430e.jpg/)

anyhow, going further we discovered that this is a "normal" behavior for surge protectors as mov's catch fire lightly and they are the key element in the surge protector ... if you have a very short peak trough it it will work ok but a longer peak (like for e.g. you get overvoltage from 230 to 270V or 200V instead of your 110) it will catch on fire ?!?!?!? not really protecting you from anything :(


things to see:
http://youtu.be/j53qtYc5ZeE?t=3m19s (http://youtu.be/j53qtYc5ZeE?t=3m19s)
http://www.davesieg.com/?p=226 (http://www.davesieg.com/?p=226)
http://www.sfowler.com/investigations/S ... ectors.htm (http://www.sfowler.com/investigations/Surge%20Protectors.htm)

so what's your take on this?

wow....hopefully not too much damage was done.

Sounds like the surge protector wasn't UL approved - they require MOV's to not fail catastrophically, or to include a series thermal fuse.

Evidently this is a common problem:
http://en.wikipedia.org/wiki/Varistor#Hazards (http://en.wikipedia.org/wiki/Varistor#Hazards)

I had always assumed that the MOV conducting for a period of time would trip the breaker or blow a fuse.  It seems like the current protection and MOV should be spec'ed together for this purpose?  Honestly, I haven't run across this since I use universal input supplies that require a fuse and maybe some filtering at the connector.

that's the problem, looks like mov catches fire before fuse has time to react

as a general rule i replace mine if there has been a surge or not every few years, I replace with ones that have a connected device warranty, and always UL and CE ratings.

like with any device manufactured in an uncontrolled setting, these are prone to penny pinching manufacturers using cheap parts.

I prefer the metal encased ones, they tend to be more robust(and more $$) but lessen the chance of fire spreading outside the surge protector.

these are one time protection devices, if you ever have to replace one for any reason cutting the cord(unplugged) as close to the body as possible will help prevent them from being used and possibly causing damage.

NEVER use one with the neutral pin removed, this is part of the protection circuit.

I have seen and repaired some pretty serious damage from faulty safety devices, a 100A mains breaker failed in a manner where it was putting 220V into a 110V branch whenever the hot water heater kicked on destroying several household apliances. fortunately only fire to happen was inside an outside breaker box.

Yeah, I've seen Mov's go up, but certainly the proper UL, CSA, etc. approved ones (i.e. name brand stuff like Siemens, etc.) will withstand more and should be self extinguishing material. Kapton is good here and I usually wrap a layer over the MOV's, if for no other reason than I've had fair size one go 'bang' on 415V supplies and copping the bits in the face certainly isn't something you want if you're working on it at the time!

Looking at that picture of the MOV (I'm assuming it is a MOV and not a Ceramic suppresion cap?), it looks like they are only protecting the downstream side. Looks to be a MOV on each line (A & N) to Earth, an X2 Cap, then a small fuse leading to a 3rd MOV on the downstream side.  MOV's are designed for very short duration pulses (Spikes) and can handle huge fault currents for that short duration. If it's a longer duration Surge or a very intense Spike, the MOV will eventually fail and likely fail in a short-circuit condition.  Under those circumstances a local fuse is needed to blow (or breaker trip) and stop that fault current.

From those pictures it appears like there is no local fuse (unless the switch is also a breaker), so the only thing that can blow (or trip) is the Fuse or Circuit breaker back at your switchboard and that might be a 20A or more breaker, perhaps with a C characteristic (relatively slow). Chances are the relatively light duty lead on the breaker would be light enough to make it hard for the breaker back at the board to trip because of the leads resistance. The lead will heat up, the MOV will still be trying to clamp (or may have gone short by now) and so it will keep heating and ...            1 x BBQ powerboard.     

Sometimes the switch is also a breaker, but if it is it will be a cheap Thermal breaker and will have a very slow trip characteristic, typically 2 x Rated Current for 50 seconds or so, 4 x for about 8 seconds, 6 x for about 4 seconds and 10 x for nearly 2 seconds. If it was a breaker and was rated at say 10A, that could be passing 40A for 8 seconds or even 100A for 2 seconds which is plenty enough to turn the powerboard into an electric BBQ!  Chances are the circuit board traces went up before the MOV and it may well have been the resulting fire that did most of the damage, not damage to the MOV may have been secondary to the PCB traces vaporising.

Actually I've collected a few fairly spectacular failures over the years, maybe we could start a new thread of spectacular failures.  :)

[quote author="Sleepwalker3"]Actually I've collected a few fairly spectacular failures over the years, maybe we could start a new thread of spectacular failures. :)[/quote]

sounds like a good idea, I'd have to see if the failed breaker is still around so i can take pics.

there is a small fuse inside with the mov, donno if it broke or not but I don't see how a mov would catch on fire before fuse pops out :(

this is scary

http://img836.imageshack.us/img836/3492 ... resize.jpg (http://img836.imageshack.us/img836/3492/dscn3561resize.jpg)

http://img826.imageshack.us/img826/3769 ... resize.jpg (http://img826.imageshack.us/img826/3769/dscn3563resize.jpg)

http://img341.imageshack.us/img341/7963 ... resize.jpg (http://img341.imageshack.us/img341/7963/dscn3565resize.jpg)

http://img32.imageshack.us/img32/2548/d ... resize.jpg (http://img32.imageshack.us/img32/2548/dscn3586resize.jpg)

http://img51.imageshack.us/img51/6903/d ... resize.jpg (http://img51.imageshack.us/img51/6903/dscn3562resize.jpg)

This reminds me a little bit of a power strip that I heard a rather loud "pop" from a while back. 

Each outlet in the strip have a switch and a red LED connected to it.  It seems like the resistors are severely underpowered to drop the led current down to a few mA in a 230 volt circuit.  It seems like they don't even have reverse protection diode so the LED will have it's polarity reversed with a voltage far beyond the safe limits every cycle .

[attachment=1]

[attachment=2]

[attachment=0]

---

[quote author="matseng"]Each outlet in the strip have a switch and a red LED connected to it. It seems like the resistors are severely underpowered to drop the led current down to a few mA in a 230 volt circuit. It seems like they don't even have reverse protection diode so the LED will have it's polarity reversed with a voltage far beyond the safe limits every cycle .[/quote]

those look like they were 1/4 watt resistors, don't think i'd trust anything under 1 watt for a a led at that voltage.

Correct me if I'm wrong, but for a 230 volt circuit at say, 5 ma, would reguire 230*.005=1.15 watts? What were the resistor values?

5mW is really high for an indicator, I would expect to see 200-500uA

They are 150K resistors.  1.5mA = 350mW @ 230 volts. 

They look smaller than regular 1/4w resistors, maybe 1/6w?  But they are underdimensioned even if they are 1/4 w rated...

there is a small fuse inside with the mov, dunno if it broke or not but I don't see how a mov would catch on fire before fuse pops out :(

That's what I was saying Arhi, I think there are two MOVs on the *Upstream* side of the fuse and so the first two MOVs (which may potentially me Y1 Filter caps) are connected via the Switch (which may also be a thermal breaker) to the incoming Active wire.

The fuse looks like it comes after all that stuff and will only blow if the MOV on the left side (*Downstream* of the fuse) also clamps hard.  Put another way, looks like the first two MOVs (or caps) which go to Earth are connected directly to Line in, only the MOV on the left is downstream of the fuse.

Pretty crappy design I'd suggest.

perhaps some nasty letters/email to manufacturer are in order.

@Matseng - Looks like 1/4W or 1/2W to me, they've been gradually increasing the Power Vs. Size ratio for a while, so much so that at work we had to change our system so that we wouldn't get one batch of older small wattage ones mixed up with higher wattage ones that looked near identical.

It's unfortunately all too common to see things like that and as most plain LED's have a PIV closer to something like 5V, it's just crazy. I had a powerboard that started smoking too years ago. It had one LED at the end and the resistor had started to cook and melt the plastic. Around here the voltage fairly commonly can creep over 250V, so it's not surprising. I think the resistor was something like a 1W, but I calculated the power it was dissipating and it was over it's rating. How these things get 'Safety Approvals' I'll never know. Likely the approvals are given out by pen pushing engineers who drive a desk and never get out and get their hands dirty in the real world.

@Skybeaver - I'll see what I can find. Unfortunately I end up with so much of that sort of junk I tend to just chuck a heap of it when there's a clean-up, but there's probably a few things around. Kind of wish now I'd taken more pics of all that sort of thing, but a lot of it would have been well before we were all walking around with hi-res phone cameras in our pockets.
Anybody else got in good spectacular failures?

[quote author="Sleepwalker3"]@Matseng - Looks like 1/4W or 1/2W to me, they've been gradually increasing the Power Vs. Size ratio for a while[/quote]
So how do they do this? Just upping the maximum allowed temperature?  I'm sure they haven't invented some kind of nano-tech surface cooling enamel yet ;-)

Mainly improvements in materials and processing I believe, they do state something to that effect. It's mostly for things like metal film resistors, not so much Carbon types. Whereas the 'mini' resistors used to be 1/16th or 1/8w, the are commonly now 1/4W and what used to be 1/4w size is mostly 0.5W now, some we get are 0.6 and I've even seen 0.7W is the same size.
The tend to be able to tolerate higher temperatures, presumably more flexible with better terminations and other improvements.
Take this one for example, they now rate it at 0.6W, but when I started in electronics, this would have been a 0.25W size. The 'Range Overview' says "This series of resistors is now rated at 0.6W (commercial spec) and 0.25W (CECC military spec).
Axial - 0.5 - 0.75W & Zero Ohm Links.  High stability metal film resistors. Extremely low current-noise level with low temperature coefficient and close tolerance." http://http://australia.rs-online.com/web/p/through-hole-fixed-resistors/0148506/?origin=PSF_433011|fp  It's been very common over the past decade or two to see ratings for sizes like that doubled or even more, but it's more about the quality of the part.
Likewise Electrolytic Caps have shrunk dramatically over the years for a given size and capacity.

Here's another example - This body size used to always be either 1/16W or 1/8W, but now is commonly 1/4W (.25W) or 1/2W (0.5W), this one being classed as 0.5W.  The data sheet goes into the tests done and references some of the IEC test requirements for determining all this crap.
http://http://australia.rs-online.com/web/p/through-hole-fixed-resistors/1650151/

Seems to me like using no surge protection is a safer option, expecialy in countries where the voltage swings are often...Just remember to backup often...

Odd, my post from before doesn't seem to be here.

That "product" would NEVER be allowed under BSI approval

1. Plastics are not flame retardant
2. The linking of the sockets with wire is WRONG.
If you used the end socket ONLY and then pulled the full current through the cable, those internal wires would BURN.
3. Not allowed to "solder" wires like that, incase a socket overheats and the wires disconnect.
4. Clearance, where the "N" wire passes close to the bare "L" connections
5. Fake/poor quality resistors.
6. single strand wire used internally

This looks like some cheap mass produced Chinese shit made for the China market and re-branded.

well it was allowed into EU for a while :( .. now is banned from EU ( http://ec.europa.eu/consumers/dyna/rape ... m?rx_id=97 (http://ec.europa.eu/consumers/dyna/rapex/create_rapex.cfm?rx_id=97)  )

yeah.. nice thick ground wire in that... :-/
Now i have to open all our surge strips to check.

[quote author="hardcore"]That "product" would NEVER be allowed under BSI approval
2. The linking of the sockets with wire is WRONG.
If you used the end socket ONLY and then pulled the full current through the cable, those internal wires would BURN.
6. single strand wire used internally
[/quote]
Hmmm...  Why would linkage of the sockets with a wire (instead of a busbar) be a problem?  If the wire is of the same area as the cable leading into the powerstrip the cable would probably burn first due to less cooling and two wires being in close proximity.

And why would a single strand wire be bad internally? Within the unit there are no flexing, just as there are no flexing in the wall sockets outlets.  Single stranded wire is very common as the fixed installation wiring in building walls.

[quote author="matseng"][quote author="hardcore"]That "product" would NEVER be allowed under BSI approval
2. The linking of the sockets with wire is WRONG.
If you used the end socket ONLY and then pulled the full current through the cable, those internal wires would BURN.
6. single strand wire used internally
[/quote]
Hmmm...  Why would linkage of the sockets with a wire (instead of a busbar) be a problem?  If the wire is of the same area as the cable leading into the powerstrip the cable would probably burn first due to less cooling and two wires being in close proximity.

And why would a single strand wire be bad internally? Within the unit there are no flexing, just as there are no flexing in the wall sockets outlets.  Single stranded wire is very common as the fixed installation wiring in building walls.[/quote]

As far as I remember single stranded wire is forbidden for home instalations in Spain. I don't know the exact reason but it's been forbidden for quite a while. I'll ask my father (an electrician) to check why. Maybe there's something wrong with it...

what is "single stranded wire" ?

I know about "solid wire" (one copper wire inside insulation) and this one is usually used for house installation (wires in wall's are solid, usually 3x3.5mm2 or 3x4mm2) and I know about "stranded wire" where there's a bundle of copper wires inside insulation and these are used for everything else (all your "soft" cables are with stranded wire, like your computer power cable).

what is "single" stranded wire?

That's just another term for "solid core wire" but solid wire/solid core/single strand/single core/1 core, all means the same thing.

[quote author="erdabyz"]As far as I remember single stranded wire is forbidden for home instalations in Spain. I don't know the exact reason but it's been forbidden for quite a while. I'll ask my father (an electrician) to check why. Maybe there's something wrong with it...[/quote]

Yes, please do - it would be interesting to know.  As you say it seems that modern installations where the cables are pulled in tubes are done with single wire multi strand. But a bit older installations where the cables are nailed to (or inside the drywalls) are with multiple wires single strand.

interesting, solid core copper wire is still the standard for electrical installations in the USA AFAIK.  I'd be curious as to why you'd want stranded wire in walls ( more flexibility for a cable that never moves and less current carrying capacity).

For the surge protector in question, most times you'll see stranded wire used in this application, since it's easier to get a good crimp and the wire is flexible, but I believe with all the right contacts, wire gauge, etc you can also crimp connectors onto solid core wire.

thanks for the explanation about "single strand" :)

I do see reason why would manufacturer use stranded and not solid wire inside the appliance but I don't see a reason for anyone to "forbid" using solid wire. As for contact resistance I'm 100% sure that when you properly screw solid wire the contact is not worse in any way then with crimped multistrand wire... not to mention that I seen many cases where multi strand wire is not crimped but directly screwed in the terminal and that is million times worse imo then any other case .. or the issues with poor crimping ..

As for the soldering of the wire I got explanation about that one while back from a friend. It has to do with how solder reacts under pressure. Do a simple test, use a screw terminal, take a multi stranded wire, twist it and solder the end, then push that soldered and into screw terminal and properly fasten it. Check again after a minute and tighten the screw if needed. Then leave it for a week and come back to it, you will find that screw is loose and that you can now twist a whole turn or even two before you have a proper pressure on the wire (this will now be properly tight and after another week the wire will not get loose again).

Yes, the solid/ soldered stuff tends to deform slowly under pressure. You find the same thing on industrial terminals. I frequently find 415V 3 phase connections coming loose in 'rising cage' style terminals where the wire has deformed slightly over the years and is no longer tight, in some cases this has resulted in 'BANG!' and chunks of metal blown out, other times it may just be a malfunction or it may not even have caused a problem yet. It happens with stranded wire also, but usually it's when the strands are fairly thick, say 6mm Sq. wire with 7 strands. So yes, soldered wires under a terminal are not a good idea and unsoldered stranded wires under a terminal is just asking for trouble.  I don't see any problem with soldering to the busbar though, providing it's done properly. If the temperatures are high enough to melt solder, the solder joint coming apart would be the least of your worries! By then the insulation would be well and truly heading south, even for the 105C stuff I generally use for general fit-outs (you average power cord crap is typically 85-90C rated wire).