On Monday afternoon, Dec. 19th, there was a strong smell of hot smoking plastic in the NCSA hallway outside room 1005.   It was traced to a lighting control panel - apparently a smoldering wire.  Urbana firefighters were called, and they and some of us from NCSA were peering at the smoking panel when...

<blockquote>
<a href="relay-pop.3gp">Video of lighting relay giving up the ghost</a>
</blockquote>

<p>

Below are successive 1/30th-second images from the above cell-phone camera video.  The
image isn't read out simultaneously across all parts of the camera sensor, and you can
see that on the first frame -- where the left edge abruptly becomes bright
white -- and from the fourth frame, where Luke's head looks stretched out as
I start to leap back.
<p>
No one was hurt, fortunately. Luke Scharf, standing closest, was dazzled by the glare for a few minutes, but soon recovered.  (I wonder now: were there tiny specks of copper left on the floor after those sparks cooled?)
<p>
We learned later that the real fireworks had come from behind the panel to the right (see <a href="#below">below</a>).
The electronics visible here controlled an array of relays, behind that right-hand panel,
one of which was (I suppose) beginning to short out.  If we'd had that right-hand panel open,
we might have all been showered with molten metal.
<p>

<table border=1>
<tr><td>
  <img src="Video.0019.jpeg"> <br>
  Seemingly normal, but note the bright strip along the left edge.
  That's the fireball ramping up fast, toward the end of the time this frame was grabbed.
<td>
  <img src="Video.0020.jpeg"> <br>
  Cheesy camera's detector is totally saturated.  This is a solid white image.
<tr><td>
  <img src="Video.0021.jpeg"> <br>
  Still saturated for a few percent of the frame time, 1/30th second later.  As it cools down, we can see some sparks (hot droplets of wire, maybe?) flying.
<td>
  <img src="Video.0022.jpeg"> <br>
  More sparks.  A fireman's flashlight (which had been shining on the electronics 1/10th of a second earlier) is starting to pull away, shining on the panel.  I'm pulling away too...
<tr><td>
  <img src="Video.0023.jpeg"><br>
  ... so the next few frames are pretty dizzying ...
<td>
  <img src="Video.0024.jpeg">
<tr><td>
  <img src="Video.0025.jpeg">
<td>
  <img src="Video.0026.jpeg">
<tr><td>
  <img src="Video.0027.jpeg">
<td>
  <img src="Video.0028.jpeg">
<tr><td>
  <img src="Video.0029.jpeg"> <br>
  The last of the sparks stuck around for at least 1/3rd of a second.
<td>
  <img src="Video.0030.jpeg">
</table>

<hr>

<a name="below">
<table border=1>
<tr><td> <img src="IMG_20111219_174701.jpg" width=90%> <br>
	Lighting control panel, pre-fireworks.  The firemen are pointing some sort of IR imager at it, looking for warm spots.
	<p>
<tr><td> <img src="IMG_20111219_174708.jpg" width=90%> <br>
	This panel controlled lighting all over the building, apparently.
	Some lights went out when it failed, including in some adjacent offices (but not others), and including the Liebert chiller in nearby room 1003.
	<p>
<tr><td> <img src="IMG_20111219_191144.jpg" width=90%> <br>
	Post-fireworks.  Right panel is now open, exposing some cooked relays.
	<p>
<tr><td> <img src="IMG_20111219_191154.jpg" width=90%> <br>
	Closeup of cooked relays.  I'm not sure what the white deposit is -- maybe ash from the smoking insulation/relay body?
	<br>
	The terminals on the next-to-topmost relay are clearly discolored; are they also on the one below it?
</table>
</a>