How to do fail safe power switch?

[John Fields]

Well, my "bottom of the barrel" suggestion is - monitor the "green"
lamp current, if on for more than a pre-determined time, then fail to
all lights flashing amber - this seems to confuse most drivers in a
satisfactory manner.. Or dont worry about it and let Darwinian natural
selection weed out the gene pool a bit......

 

[Rich Grise]

I saw a guy fishtail off the street and knock over a signal post.
The remaining signals all went flashing red.

 

[David Lesher]

Lamps do fail both "open" and "shorted". Usually shorted blows a fuse
or trips a breaker somewhere...

I wish. Many moons ago, at LeRC, I designed and had built some
16-channel cards for a Q-Bus system; to drive T1_3/4 lamps in those
rectangular push switches. You had 4 lamps per switch and with gels
could "change color" by powering different lamps.

These were driven by a Falcon on the backplane; I think via a DRV-11
card. I was issued a hard requirement that the lamps had a minimum
on-time sufficient to be seen; say 1-5 seconds, as some feared the
sw would blink them on and off so fast that...

I used #330 lamps. I designed a card with decoding and ....NE555's...
as both the timers and drivers.

Since there were 32 cards per system x 16 lamps/card; it turned out
I needed a fair sized supply, but so what...I found a unit with
SCR-regulation of the primary side -- far cheaper than direct DC
regulation and more than adequate.

Buying the 555's was an issue. You Shall Not Procure that which is
in the stockroom; Get It There. Stockroom err.. blew a fuse...
themselves and accused me of "hoarding" when I req'ed 5000 NE555's.
That was many times the predicted annual demand for the Lab. They
refused. Procurement said "You Shall Not..." again. I finally got
them from DSA in Columbus...learning that if you buy^H^H^H trade
virtual money for real things but WITHIN the USGovt, that's NOT
"procurement" no matter how big...[Later, moi, a lowly GS4, signed
a PR for a $0.5E6 order of custom graphics terminals, proving it's
easier to spend more than less...]

Finally, we got everything built and the chief tech & I went to the
test bed, in the big computer room. Mounted the backplane, ps's,
switch array, connected the Falcon to the PDP-34 that drove it, and
started it up. All went well until I turned on the 50A lamp power
supply.

WACK went the chassis... pieces spit out the fan toward us, and a
significant puff of smoke exited the cabinet and went toward the
smoke detector above. I recall thinking "Let's see... Set off two
non-adjacent smoke detectors and the sprinkler system will rupture
the air/H2O isolation disk. There are about 150 people who WILL be
PO'ed when the EVAC alarm goes off. Plant Protection [ie fire dept]
will be here in about 90 seconds. Shit, the paperwork this will
take.."

The good news is the 2 Zillion CFM HVAC system that kept the resident
370-3033, Cray 1, DG Nova, 40-odd PDP-ll's and several VAXi cool
had just-below-the-ceiling air currents that broke up the cloud and
spread it around....no horns...

We powered down and looked inside. About a dozen to 18 NE555's were
in pieces or just plain gone. It turned out that some high percentage
of lamps were hair-shorted. We rigged up a direct-to-lamps power
cable and tested each batch that way; the many hundreds of instantaneous
amps from that 50A supply made sure none were REALLY shorted, at
least not for long.

But I went back and justified my decision. The 555's were something like
27c each from DSA, and even if I'd had room for picofuses, they were
over a dollar fifty each.

 

[Tim Shoppa]

Lamps do fail both "open" and "shorted". Usually shorted blows a fuse
or trips a breaker somewhere...

I wish. Many moons ago, at LeRC, [...]
We powered down and looked inside. About a dozen to 18 NE555's were
in pieces or just plain gone. It turned out that some high percentage
of lamps were hair-shorted. We rigged up a direct-to-lamps power
cable and tested each batch that way; the many hundreds of instantaneous
amps from that 50A supply made sure none were REALLY shorted, at
least not for long.

But I went back and justified my decision. The 555's were something like
27c each from DSA, and even if I'd had room for picofuses, they were
over a dollar fifty each.

In our TCR's, there are fuses on all the local control panel
bulbs (28V nominal supply... some of the big panels, especially at towers,
must have many hundred bulbs.) We have the slight advantage that in normal
use only a fraction (10-20%) would ever be lit at once. Newer TCR LCP's use
white LED's.

I've often wondered how manufacturers of table lamps etc. get away with
using 18 gauge zip cord, when everything else on the circuit is
sized for 15 Amps and 14 gauge wire? Do 120VAC bulbs have some very well
known characteristics when they fail shorted such that UL, CSA, etc will
let them get away with this?

Christmas lamps also use skinny wire, but there are so many bulbs in
series that it seems like a single failed-short bulb wouldn't have much
effect, as all the other bulbs make a pretty good current limiter.

Tim.

 

[John Woodgate]

I read in sci.electronics.design that Tim Shoppa <shoppa@trailing-

I've often wondered how manufacturers of table lamps etc. get away with
using 18 gauge zip cord, when everything else on the circuit is sized
for 15 Amps and 14 gauge wire? Do 120VAC bulbs have some very well
known characteristics when they fail shorted such that UL, CSA, etc will
let them get away with this?

Some lamps (bulbs) contain a fuse in one lead-wire. But just apply Ohm's
Law. The supply impedance at a typical 120 V outlet is about 0.15 ohms.
The resistance of a table lamp lead may be another 0.15 ohms, depending
on length. So the prospective short-circuit current if a lamp shorts out
is 400 A or so. This causes the supply fuse or circuit-breaker (if it
hasn't been replaced by a bit of pot spindle) to clear in such a short
time that the lead doesn't heat up to any hazardous extent.