- Feature Articles
- CodeSOD
- Error'd
-
Forums
-
Other Articles
- Random Article
- Alex's Soapbox
- Announcements
- Best of…
- Best of Email
- Best of the Sidebar
- Bring Your Own Code
- Coded Smorgasbord
- Mandatory Fun Day
- Off Topic
- Representative Line
- News Roundup
- Editor's Soapbox
- Software on the Rocks
- Souvenir Potpourri
- Sponsor Post
- Tales from the Interview
- The Daily WTF: Live
- Virtudyne
Admin
Yup, here's your problem. Someone set this thing to 'Evil'.
Admin
I once supported a HP optical jukebox that had four drives and two robot arms. The jukebox itself had a glass panel on the front so that you could watch it whizzing around swapping disks in and out. (It was especially impressive if you told it to audit itself.)
One day I noticed that retriving from optical was not working so I headed over to the server room to have a look. What I found was that the jukebox was in a self-test mode and was slowly moving the robot arms up and down the shelves. This is what it does when it thinks a platter is half in its slot and doesn't want to slam into while moving around. The problem was that it wasn't able to find a platter sticking out so would get to the end of its test run, reboot and then try again. I talked to the support people at HP and they sent out one of their expensive engineers to have a look at it.
What he found, on opening the jukebox, was that one of the sticky labels on the spine of a platter had peeled off slightly and was cutting the laser beam that detects a disk sticking out. Obviously the arm was not able to physically detect it in the same way the beam could...
Admin
Admin
The robots were running around on a carpeted floor? Does nobody else see a problem with this?
Admin
Not if the robots are wearing sneakers.
Admin
Admin
Admin
I once worked in a lab doing research in assembly robot arms. It was before my time but one guy thought it would be fun to program the arm to switch itself off. There was a big push button on the side of the base.
The program worked perfectly steering the hand to the off switch and depressing the button... and then carried on until it had crushed the button into the housing. He had not reckoned on the standard behaviour of most push buttons which is they only disconnect when the button is released!
Admin
I think the static discharge caused the CPU to execute the SSM instruction (opcode 0xde, 0xad) which stands for 'Switch to Skynet Mode'
Skizz
Admin
I've heard many stories like this, usually involves misogynistic mainframe terminals that don't like women- or at least their static inducing combination of silk dresses and nylons, and produce gibberish instead of reasonable output.
Admin
At one site, they had new, STK NearLine robot tape silos (ObAnecdote: "NearLine" = "Nearly a silo"). We put the cartridge into a slot and requested it. "No such tape" error. Tried again. After much futzing around, finally figured out that the red lasers on the robot really, really didn't like the red cartridge. We had to take the entire silo down to manually remove the cartridge, during prime shift. Folks were NOT happy...
The best hardware debugging episode I ever witnessed was in 1980, at Bell-Northern Research. A test RLM (Remote Line Module) for a DMS-100 switch (in development at the time) was misbehaving. This was a box the size of a largish beer fridge, with many circuit cards in it. The hardware guru was called in. He looked at the symptoms, thought about it, then pulled out a card and rerouted a wire on the back -- didn't change any connections, just moved which pins it ran near -- and put it back in. That fixed it.
I was in awe, still am...
Admin
Admin
Admin
That machine was called an IBM Data Cell. http://www.columbia.edu/acis/history/datacell.html
Admin
If it happened today I probably wouldn't be able to email the problem resolution report back to the home office, as the proximity of "lady", "bosom", "stretch", and "wang" would choke the porn filter.
-Harrow.
Admin
I remember a similar machine in my first adventure against the Sariens (or was it the second....)
(Someone has implemented several of the early Sierra games {in javascript, I think} at sarien.net)
Admin
Or the static surge could have flipped a bit in the instruction pipeline on the 8080, causing it to emit this infamous instruction:
Which was first uncovered by the governor of California when he defeated Skynet.
Admin
I ran into the same thing on a job in Mexico. The electricians down there laid all of the 480V AC right on top of the 24V DC signal wire. Evertime a motor was turned on, it activated every photosensor along that wire run which would signal the load was ready for transport and startup about 12 load conveyors.
I was cool to see a days worth of production hit the floor for no reason. It presisted for about 4 days before we found out what changed.
Hey for eight bucks a day for an electrician what do you expect???
Admin
I wonder if there's an Asimov story where he considers the possibility of a buggy robot.
Admin
That bit juice mentioned above was trademarked MagnaSee.
Admin
Admin
I was involved in replacing word processors called Micoms with something a little more modern for the time. One job was to convert the 9-inch floppy disks they used to the new format. I called the company that did it and they said it would be $200 per disk. I said I guess that's not too bad (everything to do with computers cost a lot back then). He said, "no, it's pretty bad. Micom disks only hold 9K".
9K on a 9-inch disk. I could write that small.
Admin
A few years ago i was at a customer site. They had a several WORM drives (each transaction was burned as soon it was made) with a robotic arm to pick used WORM cartridges from their shelters so they can be read. The whole thing was installed in a unit that looked like a big closet with doors with contact sensors, so the robotic arm will be disabled when the door was open. In that day a tech from the manufacturer was working with the unit - and of course he used a bit of duct tape to disable the door sensor. Just when he was looking inside the unit, someone elsewhere in the building made a request to access some old data: the robotic arm comes to life and knocked the tech guy right in his forehead...
Admin
I'm the Craig Landrum mentioned in this story and I can assure one and all that this is not BS - actually happened. The Fichetrieve units (manufactured in the 1980's by Lektriever - Google them for a list of their dealers) were heavy duty items, employing a large robotic picker that ran on rails along the center of the unit. The robot was actually quite heavy since it had to pick up and move large trays of microfiche weighing a few hundred pounds each. The robot really was driven by an S-100 card 8080 system and it really did react badly to static shock on the outside of the unit. I believe the issue was fixed with the use of additional grounding or isolation caps, etc. I believe Lektriever built the three units in response to a government RFP to update the processing line.
That whole NMPC processing line was a fascinating project in it's day, almost completely automating the production of strip-up microfiche. There was a mounting room that had units that would take the 24x microfilm and cut it into individual frames, apply glue to the edges, and - under computer control - glue the frame to the fiche, which was then refiled in the Fichetrieve units. Fiche mounting robots!
My original story contained more detail, but I understand why Alex rewrote it to give it more immediacy and punch.
In any case, this story actually happened.
Craig Landrum, CTO Mindwrap, Inc. craigl at mindwrap.com
Admin
<snip> Sorry, buy 8080 microprocessors did not have a firmware... <snip>
Sigh. Kids. Ever since chips could run instructions, there has been firmware to store the programs that run them. I know this because I've restored numerous 8080 and Z-80 Altair and IMSAI systems and used to write firmware for a living.
The 8080 didn't have firmware on board like the modern PIC chips and other microcontrollers, but they could easily use external PROMS or EPROMS.
Admin
cool you stole my robot photo for this article...
Admin
I can't believe that in 2½ years, no-one has pointed out that the problem would have been solved by the admiral re-booting.
Admin
It's not really a robot or mainframe story, but I do recall a time when there was a brand of one armed bandit in some pubs that was susceptible to static shock. Usually it would just crash and need a power cycle, occasionally though it would dump out cash. Not all of its cash, only what was in its float chamber. Pretty handy when I was skint student. You could trigger one with the piezo sparker from an electric lighter. They got rid of them quick though, once the trick became well known.
Admin
oh, this reminds me of a "darwin award" story...a warehouse had a robot that zipped around on rails moving stuff around. since it moved very fast, there was a safety system: the robot would stop whenever the door was open, and the door was designed so it could not be latched from inside... but some IDIOT , for unknown reasons, thought that was too slow, and went to a LOT of trouble to shut the door from inside-and was KILLED by the robot.
Admin
I've had plenty of experience with ESD (electro static discharge) knocking out equipment. It's almost always a reset line that's most vulnerable, because it's usually pulled high (inactive) with a weak resistor: 10k or so. A noise pulse looks like a reset command. One vulnerability test involves sparking not to your equipment, but to a nearby metal plate. The current surge through the plate throws a strong "plane wave" that can be picked up by wires inside your system. Take it from an old hardware fart: this story is believable.