Depletion Mode Power MOSFET?

But I bet it was strong when it was all bolted together!

It's a retrofit. We're changing the motor assembly along with it's board. So yes, there are many logical reasons.

If the FET's routinely blow then I'll have a bad name in the service department, but I don't think that'll be seen as quite so egregious an error as a battery running down (consider your Triumph with half a twisted-off bolt holding the head on -- it's just not the same as that damn stud, is it?).

Aren't there diodes fast enough for that yet?

:-)

Beat me to it. You can also power a hunk of CMOS logic, to manage the braking logic, from the same lithium battery. 30 years life should be easy.

John

I see. The inaccessibility of the board had escaped me. (And I didn't mean that you were being illogical, only that the ideas of 'battery' and 'control board' don't necessarily include 'inaccessible'.)

Quite so. One stud did stick, and I had to bend the timing chain guide into a pretzel getting the head off. (Fortunately it was made of soft aluminum, so it lived.)

Yes there are, but I'm not selling them yet.

I recently showed that I can get wide modulation bandwidths (several gigahertz, limited by my test equipment) in a 200 THz (i.e. 1.5 um telecom band) waveguide-integrated detector. Other folks have seen bandwidths up to 176 GHz, but that's on a 10.6 um (28 THz) carrier, in free space rather than waveguide, and with much lower efficiency. It's a slog, but I'm getting there. (It's sort of interesting that the limited charge density in semiconductors limits their response to carrier frequencies below 10 THz, at least when considering normal collective behaviour e.g. in a rectifier vs single-electron behaviour, e.g. in a photodiode. Metal-insulator-metal devices are potentially many times faster.)

Cheers,

Phil Hobbs

how about manually shorting them - is the techs time so precuiuos that he can't fit a jumper cable?

can you get away with less viscous drag by using an impact wrench?

I buy the board space point, but not the reliability nor the cost (because further on you said that the solution need not be "terribly cheap").

Technically the relay could be very small because all current it has to swallow is the amount created by someone manually twisting the mechanism, not the motor current. Taking care of what happens when someone pulls the plug on your device when the (possibly big) motor is at full rpm: It gets shorted out by your wimpy relay, screetches to a halt, ruins the mechanics and welds the relay short forever. But in a more controlled situation the reöay would always switch under zero-load conditions, with a lifetime close to infinity.

This beats all the battery-plus-MOSFET ideas by a long shot.

robert

Huh? I thought you were a relay fan, like me.

I don't know how long the service typically occurs after power down, or if the service guy routinely powers up the thing before going to work on it, but a capacitor could be a solution as well.

As would be a manual switch, this being a serive situation anyway.

robert

Yup, I was about to suggest this as well. One small relay plus a bit of power management. That could be pretty small.

-- Thanks, Fred.

I couldn't see anywhere that you mention voltage, a while ago I bought some TO251 MOSFETs from Farnell that had RDSon in the tens of milliOhms range, the 15V one's had ID rating over 50A.

Wow; can you say more about the method? I just know the basics here, so maybe I am misunderstanding but 1500nm is IR, so you must be talking about IR LEDs/lasers, and modulating the intensity, right?

No, it's a plasmonic antenna coupled to a silicon-on-insulator waveguide, with a metal-insulator-metal tunnel junction at the antenna. You can read our paper from last year here:

Those ones turned out to be thermally dominated, but the recent ones are quick.

Cheers,

Phil Hobbs