help me understand my oven fire?

My oven was preheating this morning when the bake element started to arc, somewhere along the middle of its length. Not arcing to the oven wall, but somehow internally. Despite hitting the off button, the arcing continued, and slowly moved along the length of the element. I thought it was just a fire, so I poured some water on it (I had an extinguisher nearby, but I knew that would make a giant mess so I wanted to save it as a second-to-last resort (911 being the very last resort)). Water did not make the fire go out, but while I had the oven door open I noticed the classic 120-Hz buzzing, and realized this was not a combustion fire, but an ongoing electrical arc, presumably from the element to the grounded metal sheath. I hit the breaker in the basement, and the arc went out. Whew.

So now I’m puzzled. The oven of course runs on split-phase 240. But shouldn’t the off switch (i.e. the relay controlled by the membrane-keyboard off button) interrupt the connection for BOTH legs of the 240? Should I be replacing a relay somewhere in there? Or maybe the whole damn oven?

Was it storming outside?

I had this happen with the heating element on a dishwasher in the early 1970s. A lightning bolt hit very near (boom and flash simultaneously), then I heard some buzzing in the dishwasher. I opened the dishwasher and saw the arcing. I believe it was “ball lightning”. The dishwasher was not damaged.

Does the oven work, now?

No rain/lightning at the time. The element is of course trashed. I don’t know about the rest of the oven (I haven’t turned the breaker back on yet), but I suspect it’s all fine.

How far could this arc have migrated before extinguishing? It was walking up the length of the element like a fourth-of-july sparkler. What’s the contruction like at the terminal of the element? Would it have finally drawn enough current to pop the breaker, or would it have progressed outside the oven enclosure and possibly started a house fire?

Your description makes no sense, electrically, since an element is a single-conductor item. There’s no second conductor or neutral for an arc to travel to (from the electrically hot conductor). I mean, I believe your description, but can’t make sense of how it might happen.

I am trying to remember if some elements can burn (up) in this fashion, from a crack or internal flaw. I am dimly recalling elements with badly burned or bubbled sections.

Shutting the oven off should have stopped the event, though. I wonder if excess voltage was making it to the relay/element/controller and jumping contacts?

Good luck in all, though. Sounds like the kind of thing that could get out of control. Water’s not the recommended solvent for electrical fires, though. :slight_smile:

A couple of things, from my own experiences. I’ve had heating elements go bad before, and they self-destruct like that.

I’ve also had circuit breakers go bad, and while not tripping, fail to pass current through themselves. I imagine a circuit breaker could fail the other direction as well.

Lots of modern circuitry doesn’t wait to supply voltages to the components until you turn them on, they supply voltages to the internal components all the time, and only supply the GROUND RETURN when you hit the power button. So a short-circuiting element might well fail in a way that seems to bypass the on off switch.

I would replace the element, and maybe the breaker (breakers are relatively cheap, and easy to swap out), and see if that resolves everything.

Sounds like maybe there was more than one thing going on here? An oven element has a thin wire inside it that carries the current, which is then surrounded by tightly packed magnesium oxide that prevents electricity from passing to the outer casing of the element (see this clip from The Secret Life of Machines at 16:45 ). Theoretically if the magnesium broke down somehow, or was compromised because something was dropped on the element, crimping it for example, it should just trip the breaker. Maybe there was some sort of ground/neutral fault and malfunction in the controls, possibly combined with a compromised element?

Isn’t that totally backwards from the way it should be? Light switches for example are spliced between the hot from the circuit breaker and the hot to the light fixture. The neutral just passes through the box going the other way. The point is so that turning off the switch stops any current there before it gets to the fixture. If you put the switch on the neutral, then when you turn the switch off and go to change a light bulb or work on the fixture, you can electrocute yourself if you touch the ground wire or something else that’s grounded or neutral. Why wouldn’t a toaster or washing machine or oven also be wired with the cutoff as early on the hot as possible, to keep electronic controls powered?

Notice at 17:15 they mention that the magnesium oxide is an insulator “to stop the outer casing becoming live.” That outer casing is metal, presumably grounded to the oven chassis. My working theory (right now) is that there was a sustained arc from the element to the casing, with the amount of current below the trip threshold of the breaker. What bothers me is that voltage was still present at either end of the element after pushing the off button.

But this was split-phase power, with current flowing from one live leg to the other. During normal function, ground wouldn’t have been involved at all.

If the element had just failed open-circuit, it’s likely I would have thought to check for voltage at the connectors before trying to replace the element, but that’s not true of everybody; if the oven is deliberately designed to leave one leg live, that seems like a recipe for electrocuting your customers. If it really is designed to interrupt just one leg of the split-phase power, that’s…shocking.

If it was a phase-to-ground fault, perhaps just one phase like you suggest, meaning it was only 120 volts, maybe the fault that started the arcing also fused something in the control circuity, perhaps arcing behind the panel as well. Maybe in that case there wasn’t enough current flowing to trip the two-pole breaker since it was only half of its potential. I think in a lot of 240-volt appliances the controls and ancillary devices (fans, lights, rotisseries, etc.) run off 120-volt, and for computerized electronics they’re stepped down to fairly low voltage DC, though I don’t know if the DC step-down transformer would be fed from 240 or 120.

From 120V generally. 120V ones require less wire turns, and are produced in higher volume, and so are cheaper for a stove maker to buy.

Without being on hand to diagnose:
As the arc was traveling back towards the source wire, I 'd guess it was slowly drawing more current The heating element has internal resistance. Shorter the element, the lower the resistance. Sooner or later the breaker would have tripped. As to why it didn’t turn off, I’d guess the on/off relay that is controlled by the electronics had its contacts welded by the original instantaneous overload when the arc was created. I work on industrial equipment where that is a common failure. Electronics themselves tend to burn open. Usually immediately after they shorted out. Relays burn shut until sometime much later when, if the load is heavy enough, they burn the contacts off and open. Burning open probably wouldn’t have been the case in your oven.

Water was a very bad decision to try to put out an electrical fire. Also it was a 60Hz buzz, not 120Hz. Household supply line voltage and frequency are not related. US frequency is fixed at 60Hz by law with very little deviation.

Yeah, the only deviation that comes to mind is 400Hz power used in aircraft repair facilities.

Yep, I concur. Pressing the off button didn’t turn it off, the relay was fused shut or you wouldn’t have gotten these symptoms. So at a minimum you need a new heating element and new relays. Relays are probably soldered to the control board. So the whole board may need replacement.

You are correct (I have looked at the circuit diagrams for 240 V appliances and specced relays for them) : there were 2 relays. Both had to fuse shut, or possibly what you were seeing was between one leg and ground. It is not legal to make a 240 volt appliance in the USA that uses only 1 relay, because both legs are always energized.

The ironic thing is, you can get an oven for $400 on sale sometimes. It may in fact be cheaper to get a new oven, replacement parts like a control board tend to be massively overpriced to encourage this.

The buzz in a situation like this is likely to be 120 Hz. Although the overall waveform is 60 Hz, the current rises to a maximum and dips to zero at 120 Hz. Put another way, the acoustic energy is related to the instantaneous power of the waveform. Power is V[sup]2[/sup]/R, and the square of a 60 Hz waveform is 120 Hz.

Presumably older ovens with a manual thermostat and on/off switch in a single dial wouldn’t have this problem, correct?

On the flip side, is this a failure mode that could occur in any oven that incorporates relays in the heating circuit?

I suppose both relays might have welded shut, but only one arc became established/sustained. It kept migrating toward one end of the heating element as the broken end of the element got consumed by the arc; if both legs were involved, there would have been two arcs, one migrating slowly toward each end of the element.

Ours will cost more to replace as it’s a double wall oven, but that’s fine by us. We’re a little rattled by this, and likely to regard this particular oven with suspicion going forward, even if it’s repaired and deemed safe by the [del]exorcist[/del] repairman.

This is reassuring. I assume the last inch or so on each end of the element assembly is low-resistance wire, so hopefully that would have popped the breaker before the arc consumed that wire and migrated outside of the chassis.

Nope, it was 120. 60Hz power, you get two current peaks/magnetostriction events per cycle, so 120 Hz. 60 Hz is a nice bass note, something most people can’t hit with their voice; 120 Hz is a nice low hum, classic transformer noise, something most post-pubertal men can easily hit with your voice.

Over here it is ‘code’ to connect a high current device like an oven to the mains via a suitable switched connection which is easily accessible to someone standing in front of the appliance. The idea of having to go down to the basement, or in my case, out to the garage to isolate it, horrifies me.

It is also ‘code’ to always put the switch on the ‘hot’ side. Of course, there isn’t really a hot side, but it’s important that all switches cut the same side of the supply.

I’ve had dial switches fail shut as well. The switch becomes welded inside and the knob resists turning. Yes, any oven with mechanical relays can fail this way. SSRs, or solid state relays, can fail as well, and I think are considered less reliable than the mechanical ones. (though I have the impression that all the SSR failure modes fail open)

The actual thing that failed is probably a component that costs about $5. (a quick search on Mouser)

Low-side switching is only used for DC, because of the connivence of using high-power n-channel FETs to do the switching.

A few factors fed that decision:

  1. I assumed (at first) that the element was entirely de-energized by the off button - a crappy assumption in retrospect.

  2. I was ungrounded (standing on the kitchen floor, not touching the oven chassis).

  3. I was holding the (dry) handle of a porcelain coffee cup.

  4. I was pouring from high above the element, so the water stream would have broken up before reaching from my cup to the element.

All in all, it’s a move I would not repeat. I just knew what a mess a dry-chem ABC fire extinguisher would make, and didn’t want to go there unless nothing else worked (and in this case it wouldn’t have worked anyway since it wasn’t actually a combustion-related fire).

Not about ovens I would have to look at wiring diagram, but on many 240 appliances and equipment the contacts are single pole. A two pole relay cost more than a single pole relay.

On a electrical device never assume the power if off. Always test for voltage 1st before putting your hand in to work.