Slashdot Mirror
How a Wiring Rack Should Look
Julie Jacobson writes, "It's so much fun to deride some of the worst home wiring jobs in existence. But once in awhile, we should salute some of the cleanest, most perfectly labeled cabling jobs in U.S. homes. At the recent CEDIA Expo, the association for home-technology integrators handed out awards for the Best Dressed Systems, each featuring miles of cable, hundreds of connectors, tons of steel, and a clean aesthetic that could make the most finicky designer swoon. Show them to your own installer for inspiration."
When they wired their mainframe, they spent about $2000 for a bunch of bix panels.
When it was my turn to do the same job, I took $5.00 and went to the hardware store, I picked up a 1ft by 4fr plywood scrap and bought a box of finishing nails and brought that in the office (the canadian head-office of a fortune 500 company, btw) and started hammering away neat rows of nails to which I soldered wires from a 100 pair cable we ran between two floors.
On hearing the hammering, the boss of the other department (who happenned to pass by by chance) came to have a peek, and he sees me hammering and soldering and asks me "what are you doing???"
- I'm doing a patchboard for the serial lines.
- Why don't you use a BIX board like we did in the plant?
- Because yours cost $2000 and mine only $5.00.
He left without saying a word.
Here is a picture of a site in Dallas, TX. This picture belongs to a HUGE telecom company. A baby bell if you will. ;) How they maintain this I will never know.
http://www.waystupid.com/item-378.htm
What is more amazing is that after several attempts by staffers, the management refuses to let people clean this up. And they show this to prospective customers on a daily basis!!!
Mike @ The Geek Pub. Let's Make Stuff!
Old story -- Long time ago a Vax 785 / RS232x9600 installation in Tasmania had a problem with perfect crosstalk -- one VT220 terminal was displaying & accepting keypresses identical to the one on a desk near it, with the latter terminal being unplugged from the computer. Turned out the cables were bound together neatly along their entire length, and the bits just jumped across inductively.
Do not mock my vision of impractical footwear
Zip ties are single use.
Until I started working with Dell rackmount servers, I would agree with you. Dell throws in reusable zip ties to organize the cables on the back. There is a little button you can push to release the zip tie.
I thought it was pretty neat for 1 cent of plastic.
wait a minute...these are HOME installation? Who needs 48U of rackspace and that much cable in their HOME?
Sig removed because it was obnoxious
... sometimes it has a tendency to get to your head. You didn't get your EE degree for hammering some nails to string RS-232C, any more than a doctor gets their degree to treat a common cold in a healthy 8-year old. In the same fashion, I didn't get my CS degree to write Swing UIs. Did our educations tangentially cover these things? Yep, they did, but they're a) not our core competencies and b) can be done by someone who is literate and capable of following a simple single sheet of instructions.
Why are our degrees important? Well, one thing they let us do is properly identify edge cases. A self taught programmer implementing a Swing UI with a sorted combobox might decide to use a bubblesort on it, which would work fine through testing right until it got to a customer who put a couple hundred items in it, when the application would just start to unexpectedly hang. The doctor hopefully catches that 1 kid out of 10,000 who doesn't actually have the cold and needs treatment within the next 48 hours to save his life. And you, as an electrical engineer, identify when impedence would be an issue.
Ah, but here's the rub: edge cases are edge cases for a reason, and purported experts who cry wolf regarding the edge cases get ignored by a public which sees solutions which work perfectly for 2.5% of the price. And, as several folks have pointed out, you're crying wolf here. The reason the solution appears to work isn't because the grandparent was ignorant of impedence, its because its just physically impossible for that to be a problem for that device.
Or, as I learned in Engineering school (in tech writing, of all places): "You're going to graduate with a degree from one of the best schools in the country, and you'll be working your first job with tech-school grads who have 15 years of experience, and in your first two weeks one of them is going to say something you learned in school is wrong. You might disagree, perhaps vehemently. But before you voice your disagreement, figure out exactly why he thinks his way will work, because odds are it will. Remember: he's worked there for 15 years and hasn't blown it yet, or he wouldn't still be there."
Help poke pirates in the eyepatch, arr.
>Who needs 48U of rackspace and that much cable in their HOME?
I have half that much rack gear and probably a quarter mile of audio cable, just in my modest
home recording studio. Except for my piano, I'm sure that patchbay cabling has been my largest single expense.
-fb Everything not expressly forbidden is now mandatory.
You're totally right. That's something I've learned at my new job. I was always trying to do things perfectly, and I realized that it took too long to do them as well as I wanted, and even then they wouldn't be *perfect*. I thought back to my metals class when my teacher talked about the level of accuracy which is necessary for particular jobs, and how you wouldn't build a house while measuring lengths in micrometers just as you wouldn't build an engine while measuring in inches. (I forget the name of the principle, I was in Jr. High.)
What really drove that point home was when somebody plugged an ethernet switch into itself a few weeks after I'd done a moderately good wiring job in our closet and I had to tear it all out because I couldn't even get to our management console on the switch to see which port was causing the traffic storm. (Netgear FSM750S if you want to know.) So, just as many people had pointed out, my zip-tied bundle of cables did me no good and they are now hanging off the side of the switch in a mess, no longer matched up to the numbers on the patch panel.
I guess it really does matter what kind of job you're doing... If you're not going to be changing anything, zip tying it all up might make sense. They staple AC wires inside the walls of houses, so I'm sure we can find an instance where zipping up cables would be appropriate. For me though, I'll take velcro and a slight mess.
It is hard sometimes to make service loops look neat, but they're absolutely worth any clutter they cause.
Of course, with a little creativity it's often possible to bundle everything up so that one or two snips releases plenty of extra length.
My "favorite" though is people who pull fibre cables "nice and tight" then zip them within an inch of their lives while the equipment is warm. As soon as it's powered off for a few hours, fibres start breaking. It sure looks pretty until you have to cut a zillion ties to do anything.
I know this is not going to be a popular opinion but..
Tightly bound cable bundles at the rack are NOT helping anyone but the anal person looking at the pretty pictures. Try troubleshooting or replacing a wire in that bundle at the rack. A rack is a dynamic environment, that is why there are jacks there! If it was not meant to be dynamic, why even use jacks at the rack? Just hard wire everything. If your rack is attached to the floor, movement and cable chafing is not a problem either. If you have to spend more then 5 minutes to get the bundle the way it looked before you replaced a wire at the rack level, you are wasting every ones time including your own to get it to look absolutely perfect. There is NOTHING wrong from a technical standpoint by using the plain old cable management hooks on the front of a typical rack that the cables route through, each wire does not need to be zip tied or Velcro every few inches and each wire spreading off of the bundle with a cable tie for each one, in fact, you can cross over the width of a typical 19 inch rack without the need for any additional bundling other then standard wire management trays and hooks. Side rack mounting depends on type of management you have for that, I've seen some better then others.
Like I said, my opinion is not going to be a popular one but can someone give me a technical reason why every rj45 plug in something like a 24 port blade needs to tied individually or even in groups of two? Is that "unsafe"? Is it a hazard in your environment? If so, what the hell is going on in your equipment room and why do you not have a door on your rack? Is it harder to track down then a huge bundle 20 deep and zip tied to 10ft lbs every 3 inches?
Oddly enough, I've seen many installations where the rack looked pristine but the out of sight areas or covered parts for the actual runs like under floor or overhead in partially covered trays or inside the rack vertical sections looked like spaghetti. If your goal is neatness and you justify the clean rack area for some technical reason, what is your excuse for the other areas that are out of normal sight looking like crap, do those technical reasons for a pristine rack not apply to areas others can not easily see? In order to get the rack to look nice, the extra cabling is balled up and hidden elsewhere. It does not make sense.
For reference, I do neat work now and I've had to replace and re bundle cables and wires inside nuclear reactor instrumentation control panels and rack mounted electronic instrumentation shelves using nylon string and shellac so I am very familiar with the concept and the goals of proper wire management.
Yeah, I have done time in TV studios. You seem to have seen more of it then I. They are even worse then sound studios, because there is fibre, cat5, and related for IP stuff. I remember seeing bundles of BNC carrying composite, SDI, those triplicate YCbCr packs, and some other weird stuff. Then those snakes that combined 50 cables carrying everything from audio to teleprompter feed.
The sound studios aren't as bad, it is more a question of volume. Once everything is at +4 dBV, all is well. The small studio I used to work in had a 144 point bantam bay in 'studio a', and it got fairly complicated sometimes.
The thing that shocks me though are the large multi-operator post audio consoles, with a half dozen 56 bundle MADI inputs, video sync, and feeds to ADR rooms. Intimidating.
Sig (appended to the end of comments you post, 120 chars)
No doubt, jeeves. I've been there, done that.
In a collocation facility that used to allow customer to rack their own gear... yeah, how ever bad the rack looks in your imagination at this point - double that. Now we strongly suggest they allow us to do it the first time, or we shut them down and do it over - and it typically takes about 150% more time.
I can appreciate the neat and clean wiring/racking job - for a full 48u rack with 1U servers and network gear- expect about 15-25 hours of labor to get all the systems installed and wired. This is assuming redundant power, redundant network, KVM, and management connections.
Or you could just let the customer throw it in, and they'll have it plugged in before lunch time. But it'll look like crap, have hotspots, crimped wires and fiber, overloaded PDUs, and nothing labeled.
Then they expect it to stay up and running for 3-5 years like that.
(And if you do need to change a subrack for something different, you pretty much have to replace all the cabling anyway.)
Having said that, those pics look like nothing more than (what used to be) standard telco cabling practices. We used to do that, day in, day out, with 100/200 pr cotton-braided cable - and we didn't have zip-ties, we did block-lacing. Hell, if I'd done work as sloppy as some of that 20+ years ago when I was an apprentice, I would have been failed!
What part of "a well regulated militia" do you not understand?
The critical issue here is: with the latter terminal being unplugged from the computer.
You should never, NEVER unplug RS232 equipment from the computer and leave the port enabled. The interface spec wasn't designed for it. Even with proper cable, what will happen is that the computer will receive the output it is sending, and thus receive its own login prompt.
The discussions of maintainability reminded me of a funny story.
At the first company I started, we had an excellent ops fellow who did all our wiring. The racks were immaculate, on par with the the winners in the competition. We never found maintainability a major factor, as things were wired right, and patch panels routed things as changes dictated.
However, on one occasion, I do remember his obsessive compulsive approach annoying. We were doing some moving around, so he was coming in and out of my office every few minutes for various changes, as was I. I typically don't screw in my monitor (or other cables), because, well, I don't need to, and I often change things around. Anyhow, the work I was doing that day involved plugging the monitor into a few different units to check things out. At one point, I couldn't remove it from the PC. It had been screwed in. I undid it, and moved it to the next PC I was checking, went to the bathroom. When I came back, I couldn't remove it, it had been screwed in again. Every time my employee walked by, he was screwing the monitor cable in tight, the way it "should be." This went on for about four or five times. The fact he even spotted it was amazing, much less the inability to walk by it without "fixing" it.
Love many, trust a few, do harm to none.
again Yes! Like others have said, if you work in an enviroment where you have to deal with machine faults etc (I'm actually a broadcaster but its increasingly server based now and the same principles apply) it makes it so much quicker and easier if the rack is wired neatly and in such a manner that you can trace cables (numbering helps too!). Unfortunatly what I've found when Broadcast and IT kit start to combine is people can't be bothered to make CAT5 cables and so the premade patch cables get used, for a temporary bodge this is fine however on permenant installations having cables the correct length and with appropriate numbering makes life so much easier!
Yeah, because the cables in that winning system were all store bought - NOT! How hard is it to test a cable after you've terminated it to the exact length you need for the neatest, tidiest install possible? I used to do a lot of that, and I would have to re-terminate maybe one out of every fifty cables, it's not that hard.
If you want a really nice looking install, you need to terminate yourself.
- None can love freedom heartily, but good men; the rest love not freedom, but license. -- John Milton
For truly permanent installs, like wiring from the comms room to offices, then skilled (and tested) field terminations are pretty much the only option. There is way too much slack if you factory terminate end-user wiring.
For data centers, all the really good installs I have seen use factory-made trunks going to patch panels interconnected with factory-made patches. This gives you MUCH more flexibility than point-to-point cabling, and makes box installs go much faster when you can just break out a crate of more-or-less appropriate-length patch cables.
Ideal in my mind for a data center: Server -> factory patch of a good length (use several different sizes and select whichever is appropriate) -> mini patch-panel -> factory trunk -> central patch panel -> factory patch -> whatever... (usually your switch). This involves more cables than a field-terminated point-to-point setup, but it makes changes SO much easier.
I can say that for fiber, you should only use field-terminated cables if absolutely necessary. It is simply too easy to screw up optic fiber terminations. 90% of the cabling issues I deal with in my line of business (enterprise storage support) are because of poor field-terminated 50u (or *shudder* 62.5u) cabling.
SirWired