I can remember it being a scary thing in my youf!
Kind of late to the party anyway, with EVs (hopefully..?) being the wave of the future. But in that case, the article still serves as a historical review or summary.
And there'll still (always..?) be the rough, remote-environment, mining, military, etc. applications where engines and alternators get used.
However modern semiconductors have largely put it to bed insofar as automotive electronics are concerned, you simply have to use the appropriate "load dump proof" regulator in front of your delicate micro or whatever.
Yeah, when you don't need much load current, just handle the voltage -- or drop it with a crude pre-reg, like a depletion MOS. Up to loads of a few watts, that's an easy way to get ride-through -- have done it *and tested* before.
Gets harder when you need much more than that. Beefy dropper? Switch off the load (ridethru not required)? Wide range converter?
I wonder what happens with somewhat larger alternators/generators with a wound field, do they overvolt in the same way if the load is removed ?
I guess there are not many we are all familiar with, I don't suppose a welder or arc lamp cares to much!
It may be more common than we like to think -- just this year I had happen, turns out it was the battery itself -- weak or broken internal connection. At first, I thought I had left my lights on while parked or something -- had ECU lights, but wouldn't start (solenoid chatter), managed to push-start it (manual transmission), but was really wonky, didn't seem to be taking a charge, would sometimes cut out completely? Particularly concerning / unexpected, the car also has electric power steering, not hydraulic, so when the bus tanked during a turn or maneuver.. gotta crank on that wheel all of a sudden!
The real scary thing (with respect to the electronics) is how often it can happen: if the battery is just fuckin' toast like that, the nature of nonlinear loads on the bus (e.g. relays turning off when a sudden load e.g. headlights tanks the voltage) means that it very easily oscillates. The same effective inductance that causes the load dump at turn-off, causes a complementary voltage collapse at turn-on. Then relays cut out, loads turn off, process repeats at ~1Hz or a bit less. The poor electronics in that thing, that day!
But with a battery, no one's the wiser. Basically a big enough bypass cap to deal with the inductance, and who cares. Momentary loads are sourced from the battery, and the alternator picks up the slack eventually.
(Car wasn't "poorly maintained", well, maybe parts of it I won't argue about, but... the battery itself wasn't very old, under 5 years IIRC. Freak failure I would say.)
But do the converter electronics in say a wind turbine have to be wary ?
Don't know offhand, would guess the electronics have enough capacitance to handle it (perhaps even supercaps, depending on how much and how long they can ride through for -- if a whole lot, potentially dealing with gusty conditions too). Depends on generator type of course, with that only applicable to wound-rotor types.
The trick, or the annoying part really, about automotive alternators, is they're hella reactive. Large gap between rotor and stator, not particularly well optimized, relatively high frequency (100s Hz), piss efficiency (~50%?). If field current were constant and the output were low-leakage, the output EMF would be stiff -- a reasonable voltage source as you'd hope to see, and typical of most direct-grid alternators. Instead, with so much leakage between the rotor EMF and windings, it's got a high impedance or CC characteristic, hence the large OCV when step-loaded.
AFAIK, the mechanism is only energy storage in the field winding, and the slew rate thereof; a sudden load tanks the voltage, field current is adjusted up (at maximum slew rate dI/dt = V/L) and it stabilizes, sudden unload peaks the voltage, field current is adjusted down (at some maximum slew rate) and it stabilizes. Probably the falling slew rate is given by a clamp diode and the field winding resistance; the regulator could potentially unload the field very quickly, say into a flyback clamp TVS, or at a symmetrical rate by H-bridge chopping the field winding (which returns excess energy to the bus, no TVS dissipation), but I suspect they just use the diode for simplicity, and live with the slow response.
It's conceivable that a given application could control field current very quickly, greatly limiting such an excursion; given the standards, it seems automotive ones at least aren't doing this, but it's possible that industrial examples (besides having better magnetic design) do.
Tim
Home
Help
Search
Login
Register
