cat 6

[Intrinsic0ne]

Where do these people get these answers?

Where do these people get these answers?

My friend,

First thing. All 8 conductor utp cable can be terminated 568a or 568b no matter cat3 cat5 cat5e cat6 or systemmax 400% head room solutions.
You must however purchase a cat6 quality jack on both ends. If you terminate a cat 6 cable onto a cat 5 jack you just downgraded the cable to cat 5.
The installation is the most important. If you sell a customer a cat 6 solution and you don't know what open ys or open eyes are then you are probably on your way to degrading the cable via poor installation habits. Cat 5e is very forgiving since its perfomance can be achieved even with poor installation practises. Be prepared to test the cat6 cable with a 15000 dollar fluke tester when the customer wants to see the results also. Watch your bend radius, dont use wire lube, don't let cable kink (messaging back doesnt help) Make sure you understand cable. I recommend the systemax website for cable information.

I hope this helps.

 

[hbiss]

IntrinsicOne, you hit the nail on the head. Most electrical contractors have a much too caviler "can do" attitude towards this kind of work. They assume their installation is a success when the customer plugs in a computer and can access the internet. Unfortunately that's hardly going to indicate whether they got what they were supposed to get in terms of performance.

The proof is in certification, and that is the only way to determine whether or not the job was done properly so that it will actually be a category 5e, 6 or whatever system.

I disagree with anyone who says installation and terminations are easy. Either they don't know what they are doing, have never had their work tested for certification or have only done small jobs. In my opinion current ethernet UTP termination methods are not at all "installer friendly". You might not consider it a problem if you do small jobs but consider a project with hundreds or thousands of jacks. Who on earth would you get to do the repetitive mindless job of terminating them and hope that they don't file a comp claim because they got carpal tunnel because of it?

And, installation methods are going to get even more stringent and difficult with CAT7 and beyond if copper ever goes any further. The jacks and plugs for versions of CAT7 are nothing like the 8 pin plugs and jacks used now. They will require even more time and effort and look like they will make terminating fiber easier by comparison.

The point I am trying to make here is that data wiring has become a specialty. To be able to give your customers what they pay for requires training, experience and the proper testing. You are only wasting your time and your customer's money if you say "yeah, I'll give you CAT6, no problem" unless you understand what you are doing.

-Hal

 

[dbuckley]

Data wiring has been a speciality for a long time now.

Back in the early 1990s when I had a building wired for Cat5 I got a handover of a lot of paper that was printouts and graphs of the test of every cable, along with a ten year performance guarantee.

I've been idly watching a company do some Cat5-ish wiring in the building I now sit in, and I've not seen the appropriate test equipment. No-one cares anymore.

 

[hbiss]

No-one cares anymore.

More appropriately there are so many people out there today that claim to do data wiring and really have little idea what they are doing. Sit through a two hour seminar given by a manufacturer at the supply house and you are an expert.

Cost is also a major factor. Who do you think is going to charge more, a company who employs an RCDD and certified installers or some outfit who has no one with any real training and uses helpers to pull and terminate?

-Hal

 

[petersonra]

Data wiring has been a speciality for a long time now.

Back in the early 1990s when I had a building wired for Cat5 I got a handover of a lot of paper that was printouts and graphs of the test of every cable, along with a ten year performance guarantee.

I've been idly watching a company do some Cat5-ish wiring in the building I now sit in, and I've not seen the appropriate test equipment. No-one cares anymore.

The reason they don't care so much is that experience has taught that it really does not matter so much any more. The hardware is far more tolerant of mediocre premises wiring these days. Add to that the plain fact that the cable is better, as are the connectors and tools, and that you can buy a cheap cable tester that actually works pretty well, and you find that there is little reason to waste the resources on unnecessary testing anymore.

 

[hbiss]

Sorry, I'm not buying that, it just proves my point. Without testing, if only on random samples, how would you ever know that what you installed meets specs?

The experience you talk about is if you plug in a computer and it works your job is a success. Customer doesn't know any better either and you get paid. Unfortunately that is the level of competence we have these days.

-Hal

 

[tallgirl]

Hal,

I pay a lot of attention to network error rates and throughput. In the end, that's what matters -- not some test results. If I can plug a computer in on one end and a switch on the other, and they talk to each other at some obscene speed with no errors, the wiring is installed correctly.

 

[hbiss]

That is testing, isn't it? Do that and find that you have problems all over you new network and you just busted the installer.

But what about the installations where the owner doesn't have the knowledge you have. I would say this is probably the majority of small to medium sized networks. There is no accountability on the part of the installers so they never learn what they are doing wrong or even care to know. They just keep doing what they do as long as the owner thinks it works and they get paid.

-Hal

 

[petersonra]

That is testing, isn't it? Do that and find that you have problems all over you new network and you just busted the installer.

But what about the installations where the owner doesn't have the knowledge you have. I would say this is probably the majority of small to medium sized networks. There is no accountability on the part of the installers so they never learn what they are doing wrong or even care to know. They just keep doing what they do as long as the owner thinks it works and they get paid.

-Hal

Our guys have made hundreds, maybe thousands of cables. we test each cable on a simple and cheap fluke tester. So far, we have found only a few bad ones, and none failed in the field. We have even used our cheap tester to debug stuff in the field that "real" electricians have installed that did not work. The cheap and quick tester is very good at finding bad cables.

I do not know what level of testing you think is necessary for simple cat5 installations, but any thing beyond a simple cable tester is just plain overkill and is unnecessary IMO.

As another poster said, if the cables are bad, the network statistics will show it very quickly.

To be honest, cable TV wiring is far more susceptable to cabling problems and virtually no one bothers to do any testing on them beyond plugging in the cable box and seeing if it works.

 

[tallgirl]

As another poster said, if the cables are bad, the network statistics will show it very quickly.

I thought I'd posted these the last time network reliability was brought up, but here they are again in case I hadn't. None of the cables in my house were tested with anything more involved than a basic tester that tests for miswires, shorts and opens.

junk-box-> uptime
 08:48:32 up 74 days,  5:00,  9 users,  load average: 0.04, 0.03, 0.01
junk-box-> ifconfig eth0
eth0      Link encap:Ethernet  HWaddr 00:13:D4:7A:CF:9B
          inet addr:192.168.0.1  Bcast:192.168.0.255  Mask:255.255.255.0
          inet6 addr: fe80::213:d4ff:fe7a:cf9b/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:137385535 errors:0 dropped:0 overruns:0 frame:0
          TX packets:125892917 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:114135468334 (108848.0 Mb)  TX bytes:52886471965 (50436.4 Mb)
          Interrupt:169

The protocol level statistics show the same overall situation, though they do include network failures as computers are turned on and off and connections are lost as a result. But if several hundred millions of packets can be sent and received over the course of 2 1/2 months and no errors show up, something is definitely right.

 

[hbiss]

Bob and Julie, obviously you know what you are doing. That's not always the case with everybody wanting to be network installers today. My point is that some kind of performance testing needs to be done on some basis at least to validate that the installer is using the right techniques. The IT industry will tell you that every cable on every job needs to be certified. I don't agree with that either, but the other extreme of installers who don't even understand the basics and rely on the customer connecting the computers to see if they work doesn't cut it. When I look at a job and see the jackets stripped back 4" on all the CAT5e cables behind a patch panel I wonder what is going on. Computers work so I guess its OK.

As for cable TV wiring, I've written plenty on that on this same theme also.

-Hal

 

[tallgirl]

Hal,

If someone wants to pay for all that expensive testing and certification, that's fine. Take your money and run. That's the capitalist way

I'd probably require that a sample -- maybe 1 or 2 percent for a large installation and 10 percent for a small (under 50 terminations) install -- be tested, but when experience says that Cat5 is basically bullet-proof, except in the face of extreme stupidity or gross incompetence, I just don't see a reason to test or certify every single last piece of wire. I make, repair, annoy, install, route, rearrange cables on a regular basis, totally ignoring all the conventional wisdom, and the stupid things just keep on working. Go figure.

Oh -- and stripping the jacket back isn't a problem. It makes for an ugly termination, but not an electrically bad one. Even untwisting a few inches doesn't seem to matter. Or separating pairs. Or all kinds of things.

 

[don_resqcapt19]

Julie,

It makes for an ugly termination, but not an electrically bad one. Even untwisting a few inches doesn't seem to matter.

But it will fail a test. The first time I installed any Cat-5 we rented the fancy test instrument. I didn't know any better at the time and used a standard 8 pin coupling. Our tested showed cross talk fail because the coupling had an 1" of so of straight connection in place of the twist. After we found that out everything tested fine. Now we double check once in a while with a good tester, but for the most part if passes a cheap pair test it has been fine.
Don

 

[dbuckley]

If I can plug a computer in on one end and a switch on the other, and they talk to each other at some obscene speed with no errors, the wiring is installed correctly.

Yeah, I think that pretty much about mains wiring too. If it works then who cares how it's wired. The NEC stuff is all just bullsh*t so some overpriced jerks with big expensive trucks can rip me off. If it works, then the wiring is installed correctly.

 

[tallgirl]

Don,

Here's a picture of the wye-cable I used to use for my den. It's about a foot of Cat5 with one RJ45 on one end and two pairs each going into two RJ45s on the other.

Now, the den computer is where I store a lot of video that I capture from on-air broadcasting. I made a digital video recorder years ago, before they were common enough you could just go to the store and buy one. Typical video files are on the order of 1GB per half hour, so I would copy several gigabytes at a go -- sometimes huge moves, like I have "Ben Hur" on disk somewhere. That wye, and the other one on the other end, worked just fine until I put speakers in the ceiling and fished another Cat5 from the den to the utility room. Typical transfer rates were pretty close to drive speed -- right around 50MB/s, or well over 400Mb/s, plus FTP, TCP, IP and Ethernet protocols overhead. All going through a totally non-conformaning pair of wye-cables.

 

[tallgirl]

Yeah, I think that pretty much about mains wiring too. If it works then who cares how it's wired. The NEC stuff is all just bullsh*t so some overpriced jerks with big expensive trucks can rip me off. If it works, then the wiring is installed correctly.

Well, that's a bit extreme. I think we all agree that the NEC is a very conservative standard. That's why if you make some minor mistake and the inspector doesn't catch it, 999 times out of 1,000 the house doesn't ever burn down, the breakers don't trip, and no one ever knows. Does it mean the NEC is invalid? Nope, it means that it's a conservative standard.

What I'm saying here is that the "you must test and certify every cable or else it isn't going to work!" requirement is excessive. You can take a typical quality assurance approach to your install -- test a subset of the cables to insure they meet the requirements. If you find they don't, examine your processes and procedures, identify the causes of failure, and go again. If you do this carefully and methodically, eventually you can do just as good a job as those guys with the expensive trucks and expensive test tools.

 

[dbuckley]

Well, that's a bit extreme.

Provocatively so, no doubt about it.

What I'm saying here is that the "you must test and certify every cable or else it isn't going to work!" requirement is excessive.

Based on hard experience, I can categorically tell you that it isn't. Back in the day, an organization I worked for expanded from three buildings to eighteen buildings, and I specified the networking for most of those buildings. This is City of London stuff, so the flood wiring requirements were (for their day) generous, usually four outlets per desk, rising to sixteen for a trader desk. There must have been hundreds of thousands of terminations, if not millions. Lots and lots and lots. I had a bookcase full of test reports for every one of those cables, along with the ten year warranty. The system was 100% tested with 100% pass.

But what I know is this: The same guys installed bunch of cable after bunch of cable, and they were good at their job, they understood you dont go below bend radius specs, and you dont pull the stuff like a tug-o-war. Nevertheless, despite all this care and effort, sometimes, a cable would fail testing, requiring its removal and replacement. Sometimes a termination would be required to be redone. Sample testing would have been unlikely to find the faults.

But you are (sadly) correct in that it probably wouldn't have made any difference, as many of those cables were for phones or video, rather than 10 mbit data, which was what it was all for back then.

But the bottom line is this - I like to do a job properly, and I like others to do the job properly. If some sparkie installs me a socket, I'd expect him to have tested it properly. If I pay someone to install me a data point, I want it tested properly. Anything less is just a lousy job. And like I said, no-one cares anymore. If it has continuity it'll do.

 

[don_resqcapt19]

Julie,

Typical transfer rates were pretty close to drive speed -- right around 50MB/s, or well over 400Mb/s, plus FTP, TCP, IP and Ethernet protocols overhead. All going through a totally non-conformaning pair of wye-cables.

I didn't say that it wouldn't work...only that it would fail the test. I think the failure mode on my straight coupler was near end crosstalk.
Don

 

[petersonra]

Julie,

I didn't say that it wouldn't work...only that it would fail the test. I think the failure mode on my straight coupler was near end crosstalk.
Don

IMO that is not necessarily a failure. It is an example of over done test specs.

 

[tom baker]

Actually the testing is often to meet the jack and punchdown block manufacturers 5-10 or 25 year warranty. And what does it matter as in ten years we will be installing some product that has not been invented yet.