Slashdot Mirror
CERN Engineers Have To Identify and Disconnect 9,000 Obsolete Cables (vice.com)
An anonymous reader writes: CERN, home to the Large Hadron Collider, has grand plans to update the world's largest particle accelerator complex in the next few years. But engineers have identified a barrier to the upgrade: there's no space for new cables in the injectors that accelerate particles before they enter the LHC. In the past, when parts of the accelerators have been upgraded or added to, engineers would often additionally replace the cables that connected them. In the process, they would leave in place the old cables that were no longer in use. Now, a heap of obsolete cables are blocking the way to install new ones needed for the accelerator’s next big upgrade. To make space, CERN engineers have set out to identify and remove the old, unused cables. All 9,000 of them.
Sell them to audiophiles. You have a limited supply of cables used in a unique, world-class esoteric application. That's a perfect match for people with deep pockets and shallow skulls.
Praise Buddha that removing abandoned cables is now a code requirement in the US. I remember an old server room where the manager wanted to raise the floor 6" so they could fit in more cables. 12" apparently wasn't enough...
Sounds like my rack.
Watch this Heartland Institute video
Sadly, I was thinking the same thing... might be time to get started doing that before I also have 9000 cables to identify and remove... :P
cern manager: we've disconnected 4 cables...can anyone confirm on the console that these are disconnected?
cern engineer: nothing new here chief.
cern researcher: my panini press stopped working.
cern manager: ok wrong cable
cern engineer: janice had a panini press in her office?! I want one
cern manager: guys lets not get off track here...
cern mathematician: Where do I file a report about the espresso maker? its seemed to quit working entirely.
cern laureate: my jack lalane power juicer just cut out and im mid-smoothie, this is urgent...
cern manager: just use the vitamix in my lab.
cern engineer: vitamix?! am i the only one here whos been drinking freeze dry sanka for 5 years?!
cern mathematician: of course not Ive been drinking your sanka too...
Good people go to bed earlier.
LHC.....No wireless. Cost more than Fermilab. Lame.
If you post as Anonymous Coward, don't expect a reply.
Are you looking to get into the fast-paced, exciting world of experimental particle physics? Then CERN might be looking for you! We are currently in the process of hiring an unpaid intern to help us perform maintenance on the Large Hadron Collider (yes, you read that right--THE Large Hadron Collider!)
Pay: This is an unpaid position, but the experience you gain will look great on your resume!
Hours: Don't Ask.
Required Qualifications: At least 10 years experience working with high energy supercolliders, 15 years experience with networking enterprise-level systems, and a minimum of a Master's Degree in Theoretical Physics (student experience will be considered)
Position: Entry-level.
9000 cable... and no labeling
Sadly this is common...
As a lab manager I had to institute a rule that ANY cable that didn't have a label was going to be removed when found with no warning. Any cable which was incorrectly labeled, was subject to be connected to what the label said, or if that wasn't possible, the label would be removed and then the cable was pulled for not being labeled. Label content was defined and all where trained on how to make proper labels, and retrained when they came to ask why their system suddenly stopped working.
Maintaining a lab is a daily discipline, like cleaning house. You have to pick up after yourself as you go along or at the end of the day the mess is huge. Hey, where you born in a barn? Your Mom doesn't work here, clean up your mess!
"File to fit, pound to insert, paint to match" - Aircraft Maintenance 101
One of the few silver linings of the Hurricane Sandy damage: they finally pulled tons of old copper out of the tunnels and cable-runs and replaced it with fiber, because there was finally no way to be sure which was obsolete and which was current-but-damaged.
If they know they are 9000, that would suggest they have already identified them.
It's not like the New York City Subway, where a combination of age and bad record keeping in the early years, combined with the fact that it's 3 or 4 different systems that merged into one system has led to most of the engineers not knowing what's down there at all.
You ask how many unused cables are in the NY Subway, and you'll get a shrug. Nobody knows. Hell, my favorite is when they break through a wall and find track and a train that nobody knew about for 60 years.
9000 unused cables? Pfft. That's not that impressive.
If telephones are outlawed, then only outlaws will have telephones.
Imagine trying to deal with this.
Nope, no sig
I don't know for sure but I bet this was part of a penny pinching cost analysis up front.
I recall when moving to a new site setting aside some time/budget to ensure that every cable was labelled (so, for example, we could trace ethernet from port on switch to patch panel to underfloor cable to floor jack to desk cabling to desk port) and set up a simple database to keep the details.
Work was killed off by accountants as an expensive luxury, after all cables didn't move often did they?
Fast forward to a minor flood under the false floor taking out some (but not all) systems. Fortunately some of them were in the finance and commercial group.
Suddenly it was "why can't you reconnect me NOW??". Money was paid for an 'after the event' recording of wiring by external people (which cost about 5 times the 'saving' up front).
Still at least it was better than a LONG time ago [Vax and VT220 era] when I saw one person labelling connections by yanking out an RS232 cable from a patch panel, waiting for a call "My terminal's died", asking which room they were in and making up a label and then plugging it back with "I think that may fix it" and getting pathetically grateful responses in return.
One more cable and they could have gone Super Saiyan on the task.
They are all colour coded... All grey ones are cables.
It's amazing how much cabling gets forgotten about when you have a chaotic lab environment and new stuff coming in all the time (we do hardware evaluations and other systems integration work.) There's never any money left over for structured cabling once it's been spent on all the fancy new hardware. Even if we invested in structured cabling it would turn into an unstructured mess quickly. I have racks that look like those Magic Eye pictures; the only thing that will solve it is unplugging everything. I'm sure world class scientists can't be bothered to label anything if we can't!
i use patchsee cables. a bit more expensive but you'll NEVER disconnect the wrong cable again. you get a specially shaped torch which you shine at 1 end of the cable and the other end lights up. the torch has 2 modes - stable light and flashing.
there are other brands (e.g. evo6, belden) doing similar things but they tend to be overpriced and overly complicated (cable with its own buttons and batteries in jack boot).
I am reminded of the days of wire-wrap circuit boards. hunders of wires in a few colors at most forming a rats nest of interconnects on the back. All done by hand from post to post where you had to count pins by eye to find the right post each time. Chance of 100% correct wiring was geometrically vanishing.
The problem was not discovering the connections you had failed to make (which is easily done with a continuity tester) but finding the connections that were mistakenly wired the wrong pins.
So what you did was go find a filament transformer (these were high current low voltage transformers used to power the filaments in tubes). then you put one probe on one pin, and another probe on every other pin it was not supposed to be connected to. This is not as complex as it sounds since normally one pin is not connected to more than 3 or 5 other pins. So once you eliminate those, you can just slide the probe along the sides of all the other socket pins.
The current was so large that even a momentary connection would vaporize the wire if it was incorrectly wired. A continuity tester would not have worked well because the response time for the human to test all N^2 connections and look at the continuity tester was too long.
Some drink at the fountain of knowledge. Others just gargle.
So THATS where the PFY has been all morning. I thought he was in the tape safe again.
in anima Apparatus
a device that could potentially eventually produce a particle that could turn our planet into a blob of stranglets
Potentially you could stop writing catastrophic nonsense, too.
a tone generator would make the job far easier
This is the least of your concerns. If you can merely safely access both ends of the cable, that's already a big win. Remember that different cables go between different locations, and some of these locations are unsafe to access during operation, some of them are unsafe to access at any time due to presence of radioactive dust, and many teams keep different things inaccessible at various times for all sorts of reasons. You're basically assuming problems that aren't there. They have mounds of documentation detailing the routing of these cables. The cables do have identifiers, and I can state this quite categorically. This is CERN bureaucracy, they sometimes serialize their pencils for all I know.
The "have to identify" phrase is simply sensationalistic wording. They can point to each of these cables on the plans, they know what labels they have, that's not an issue. The issue is to physically get to the ends of each cable run and find them among everything else that's there, without breaking any rules and without disrupting anything where downtime can cost a thousand Euros per second.
A successful API design takes a mixture of software design and pedagogy.
and it'll be twice as much work later on.
Wanna buy a shirt?
https://www.redbubble.com/people/stealthfinger/shop?asc=u
Since the cables have been bombarded by high energy particles,....
The facility has been active for over 60 years, and the cables are in some of the oldest parts (the injectors). So yeah, some of this stuff is obsolete; If I'm not mistaken, they ripped out a lot of 60s control electronics during the consolidation last year...
LHC.....No wireless.
The magnets in the LHC require a ~9,000 amp current and the ability to dump it somewhere fast in the event of a quench. Care to explain how you plan to do that wirelessly? It's also worth pointing out that the part of the accelerator complex they are recabling was built in 1954, 13 years before Fermilab existed and 17 years before the first wireless packet network.
Book: Puzzles for Pleasure
Chapter: "Wire Wizards", page 73
* https://books.google.com/books...
Yes, but there's a point that doesn't work any more. The original injectors at CERN are more than thirty years old; some sites are probably 40 or 50. You find me a cheap and effective way of labelling cables that doesnt' fall apart over that time span...
On the boring side: I'm pretty sure that scratch tape/disk derives from scratchpads, pads of paper used solely for doodling or intermediate calculations and stuff that you don't want to enter into your (paper) engineering notebook.
https://app.box.com/WitthoftResume Code: https://github.com/cellocgw
patchsee cables aren't transparent. they have 2 strands of optic fibre running from one end to the other. you shine into one end, light comes out of the other (no need to even disconnect the cable). the longest patchsee cables i've seen were 30 metres. i don't think they make longer ones.
Telephone companies used to have this exact problem. (Maybe they still do). Central offices contain a "mainframe", essentially a huge patch panel that connects cable pairs coming in the building to the switches. Technicians activated a given local loop by running a cross-connect pair. When service was discontinued, they'd often just disconnect the pair but leave it in the mainframe to clog things up for the future. I suspect this problem is decreasing with the growth of remote switching. E.g., AT&T U-verse terminates the customer loop in a VRAD cabinet in the local neighborhood instead of carrying it all the way to the central office.
9000 cable... and no labeling
They have angered the Flying Spaghetti Monster, and will be punished.
Don't waste your vote! Vote for whoever you want, unless you live in a swing state it won't matter anyways
Sadly this is common...
Dealing with this right now. Wiring an 80 foot carbon fiber sailboat, I think I'm about the 6th electrician that has worked on this.
Bundles of wires about half a foot in diameter, all unlabeled, all white until you cut the outer insulation off. I see why the last several quit after a few weeks.
None of this makes any sense. One of them somehow thought 4 gauge cable would be enough to run 500 amps over 60 feet. The wire chases are under the deck which has already been glued on years ago, there is no access to put new stuff in, no chase or anything, unreachable by hand.
Send help.