On Thu, 11 May 2006, Mike Andrews wrote:

Roy Lewallen wrote:

This might be a good time to mention that light bulbs, which I've seen
recommended here as a test load, have a resistance when cold that's only
about a tenth their resistance when hot. So you'll get some pretty
awesome current when you first connect one. A supply with foldback
current limiting probably won't light one at all unless the voltage is
raised slowly.

That's a cool thing to know.

FYI, vacuum tube filaments show the same thing. Lower ohms when cold.

More below...

Hmmmm ...

Bobbie Barmore was discussing IOT power supplies and their crowbars on
the GowBugs list a few months back. Seems that the test to see if the
crowbar's good is pretty simple:

1) Turn everything off
2) Safety tag and padlock all the breakers and the START switch
3) Ground everything that even might think about biting
4) Put a piece of (?) 32-gauge wire between HV and ground.
5) Pull the shorting sticks, button everything up, unlock and
untag the breakers, untag the START switch
6) Make sure everyone's clear and safe
7) Push the START switch

The power supply's good for something like 700 mA at 35 KVDC, and it is
a _really_ bad idea to get across it if you might pass less than abount
700 mA. But the wire's good for more than 700 mA, and so will look like a
short to the PS.

If the wire melts, vaporizes, or explodes, then the crowbar circuit is bad.
It should be unaffected by the test. Put 35 KV across a piece of 32ga wire
without blowing it up.

Hydrogen thyratrons to dump the power, and fast-acting circuit breakers to
Shut Things Down NOW, appear to be the way to protect the wire from the PS.
Fascinating technology.

Homebrew _that_.

35 KV at 700 ma? Exploding wires might not dump all the charge (what size
capacitor?). Maybe need 2-3 dozen carbon resistors in series with a
milliamp meter to monitor remaining voltage?

--
Mike Andrews, W5EGO

Tired old sysadmin