Last December the Minnesota Streetcar Museum had an electric arc in the ceiling of one of their wooden trolley cars that started a fire. Fortunately they smelled it and put it out with very minor damage.

Decades old insulation on wires at 600V! Sure could have been far worse.

Four years ago my family built my sister and her husband a house in rural Oregon. I wired the house and learned about Arc-fault Circuit Interrupters (AFCIs). The National Electric Code requires them for all circuits with bedroom outlets. They are similar to Ground Fault Interrupters (GFIs) in breaker boxes for circuits with outlets near bathtubs/sinks, i.e., they are a form of circuit breaker. The duration of an arc is very short thus an arc is not going to throw a 15 or 20 Amp circiut breaker or a GFI. But an arc is quite hot where it occures.

For more background on AFCIs I suggest you start at: http://en.wikipedia.org/wiki/Arc-fault_ ... nterrupter

Obviuosly household 120V AC is very different from 600V DC. Theoretically, would a 600V DC AFCI work? Are any made? What is/would be the cost of one?

I don't know the answers to these or the many other related questions. Hopefully some of you do.

So, is this something worth digging into? Or is it a weird idea?

Arcing faults in DC powered transit can be a real problem, especially at the higher voltages. Since the arc is not consistently being interrupted like it is with AC, it can burn smooth, for a great distance, and at low currents. At the last system where I worked, the newest cars had current in, current out, comparison equipment to watch for arcing faults.

I doubt there is enough demand in the DC field for the development of Arc-fault Circuit Interrupters.

There are things that can be done to reduce the chance of an arcing fault on antique trolley cars: inspecting the wiring and fixing marginal areas, testing the wiring with a megohm meter regularly, and using proper street railway lamps. Poles should not be left up on unattended cars. Smoke and heat detection devices and sprinklers should be present in your car house.

Maybe they will be high cost but if they save a one car they will have paid for the development cost.

Thanks for your thoughtful comments.

Is it even possible on DC? Can anyone explain what an AFCI even looks at?

Comparing current out and current in will not tell you about an arc fault if the arc current is returning by the normal path not to ground. That's pretty much what a Ground Fault Interrupter does, which in 120VAC wiring looks at the current going out the hot and compares it with current coming back on the neutral.

If the former is more than the latter some must be leaking to ground (the neutral being bonded to ground further up the system like in the panel box and back at the utility's transformer). If a person who is also in contact with the ground makes contact with the hot side, they will have current pass through them which rejoins the system back at those grounding points bypassing the neutral. The GCFI sees the greater current going out on hot than comes back on the neutral and trips. Current-out / current-in would work the same way. (It also requires a hard wired return path, not using the chassis ground for return but I suspect that wouldn't be done with HV circuits anyway.) But you won't detect an arc that way unless it also happens to be a ground fault.

How the AFCI works...what it looks at to determine something is arcing...when the arc current is neither excessive (which would cause an overcurrent trip or fuse to blow) or a ground fault (which would trip a GFI, if equipped)... is something I do not fully understand. But I suspect it may be something that only applies to AC circuits. Anyone know?

Steve

I looked into arc fault breakers while building my new home. I found I could not use a drill or saw in any bedroom outlet without a trip. Within the breaker there is a fairly sophisticated electronic circuit that looks for several types of arcs and decides if they are normal, like a hairdryer, or bad, such as intermittent break in your electric blanket circuit.
I have only seen residential AC versions of ACFIs but a call or email to Siemens or Square D’s industrial tech support line could help.
Trouble I see is how anything will deal with normal trolley line and controller arcing.

Close call on our part for sure!

Below are some photos of our Museum's "mother car," the subject in this story of arc faults (I've never heard of them--not an electrical engineer). One shows the burned wood where the lighting fixture was mounted. We surmise that the arc formed from the steel reinforcing piece to which the wood carline is bolted. This steel band, which is tied to the steel girders making up the side frame of the car, goes up the window post, forms the curved side-roof and clerestory and then down to the opposite side frame girder. Aside--TCRT's standard cars main frame was built like a bridge member and were produced by the American Bridge Company in the beginning. After 1898 all of TCRT's passenger cars and most work cars were built in TCRT's own shops.

The bad wiring in question is probably original from the early 1930s, when the interior lighting configuration was changed when the car was converted to one-man-two man operation. The wiring was covered with painted ceiling panels so there was no easy way for us to inspect same. After this near-miss, our shop guys took out the ceiling panels and re-wired the ceiling lights using modern material. In the process of removing the panels, the old white paint flaked off in hunks revealing the original pin-striping and faux-wood finish of what appears to be an early masonite-like material. Since we maintain No. 1300 in its June 1954 appearance, this old paint has been removed and the panel re-painted.

Thanks!

Here is an article that explains a bit about what the device is looking at in determining that an arc fault is happening.

http://www.pyrogen.com/arcfaultsae99.pdf

They say that this could be applied to DC but I really don't see how it's directly applicable as there is no expected waveform that an arc current would deviate from. I saw another SAE site that says there is a need for a high voltage DC arc fault interrupter--presumably for use on hybrid and battery automobiles-- but it doesn't sound like it exists yet.

Steve

How many of us isolate circuits and meggar the old wiring inside the framing of our cars? I think many of us would be unpleasantly surprised.

dave

Would the arcing of the trolley pole on the overhead wire trip an arc fault interrupter?

Not if it only senses downstream. Contacts opening and closing and motor brushes are another matter. But again, what does it look at to know?

Steve

As the device is current sensing, yes, a trolley pole arc upstream would be indistinguishable from a grounding arc downstream, both of which would trip a dc version of this device.

JD

Only if it is a stupid device. The trolley pole arc is, after all, upstream. All the device needs is a ground to be able to distinguish upstream from downstream. For a lighting circuit, it would be asking "Is the downstream load acting like a resistive load for the upstream voltage available?" The device could also do some learning once it is in-circuit, so it could store in NVRAM basic facts it has observed about the load, resistive, inductive, types of noise detected normally on the circuit etc. Then it listens for abnormal and, on finding it, trips.

Hey, I'm even pretty sure we could build CFLs that work 5 in series on 600VDC. Geez with LED that wouldn't even be hard.

Are we looking for a technological fix for the low tech problem of aging failing insulation? Seems like just replacing all the wiring wholesale every half-century or so would do the trick.....

dave

You're kidding, right? ;o)

Actually, our current restoration of a single-truck car will have A/C lighting circuits using an inverter. I also understand that the McKinney Avenue Transit Authority is going to convert one or more of their heritage streetcars to use A/C motors.

Thanks!

Not at all. It's all in the ballast. Common fluorescent light ballasts are 120-277 VAC because the sales distributors don't want to have to stock 2 separate ballasts. That reflects just how easy it is for modern switching power supplies to work on a variety of voltages. It's a question of components rated for the highest voltage; wires thick enough for the current needed at the lowest voltage. And the rest is software. In streetcar applications with 5 in series, the software would need to contend with a variety of unusual conditions (including arcing). But that is it.

I for one would prefer LED as it would be easier to conceal the anachronism inside a frosted glass globe that is only somewhat wrong in a streetcar. A twirly CFL in a streetcar would simply be an abomination. Technically feasible, but ... ... no.