• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Minimum cable length for Cat5

TechnoPro

Golden Member
This link would have me believe that the minimum length for Cat5 is 2.5 meters. Any merit to this? In my mind, less distance = less attenuation, so short is good. And if this 2.5 meter length was an industry minimum, why are so many patch cables sold shorter than that?

Anyone have answers?
 
Well I'm not thinking well enough at 8am, but I can point out that if it is 2.5 meters, that distance is likely to be the entire non-repeated run. Meaning a 3 foot patch cable running from your PC to the wall doesn't matter, as long as the combined distance to the next device is long enough.
 
The absolute minimum is ~18 inches. The recommended minimum is one meter.

The issue is the related to near-end crosstalk, far-end crosstalk, and the ACR of the cable.

There are also issues with the necessary load / impedence to propagate the signal and interface to the transceiver.

"It's Technical" 🙂


FWIW

Scott
 
odd I have used cables about 6-8" in length going from router to router in my rack at home. I have had no issues I am running at full duplex not sure if that makes a difference.
 
actually I went and looked and some of my routers are 10 meg half duplex and it worked there. Maybe it's one of those YMMV things so intel plays it safe.
 
Originally posted by: w0ss
odd I have used cables about 6-8" in length going from router to router in my rack at home. I have had no issues I am running at full duplex not sure if that makes a difference.

as with all things cable related running out of spec usually appears to "work" without problems. Its when you look at error rates and what not that you see the real problem.
 
Originally posted by: spidey07
Originally posted by: w0ss
odd I have used cables about 6-8" in length going from router to router in my rack at home. I have had no issues I am running at full duplex not sure if that makes a difference.

as with all things cable related running out of spec usually appears to "work" without problems. Its when you look at error rates and what not that you see the real problem.

Take it from someone who has spent an entire weekend rewiring a a server room.. use at least 18"

The network administrators and LAN guys before me thought it was cool to arrange the patch panels and hubs on top of each other, and use 6" and smaller wire jumps from the patch panels to the hubs/switches.

And not surprisingly they were never sure why sometimes the network was flaky, and why they got increasing errors on the ports. The clean up alone was like a network upgrade.

Those strange network drops, sluggishness, and errors all disappeared..
 
Could be subjective, but on my own setup, I had been using a short (~6") patch cable from DSL Modem to Router. Periodcally, I would lose internet access. I would also lose connectivity to the router, suggesting that the cable from router to PC was bad... In any case, I switched to a longer patch cable from Modem to Router and have experienced far fewer, if any, drops. Coincidence?
 
It's more than just plastic-coated-metal. There's some science involved.

If you follow the rules (or rules of thumb), generally speaking, you have one less system to worry about.

(refrain from previous posts🙂 It doesn't matter how good the equipment is, if the cabling is flakey and / or out-of-spec, your network is crap. It's generally the least expensive component/system, it's usually the most overlooked component/system, and it's the component/system that can cause the worst, glitchyist, biggest pain-in-the-butt problems. (Flaky power is a close second).

Build from a solid foundation; use quality components and implement according to the rules.

My .02 (same ol' .02)

Scott
 
Originally posted by: ScottMac
It's more than just plastic-coated-metal. There's some science involved.

If you follow the rules (or rules of thumb), generally speaking, you have one less system to worry about.

(refrain from previous posts🙂 It doesn't matter how good the equipment is, if the cabling is flakey and / or out-of-spec, your network is crap. It's generally the least expensive component/system, it's usually the most overlooked component/system, and it's the component/system that can cause the worst, glitchyist, biggest pain-in-the-butt problems. (Flaky power is a close second).

Build from a solid foundation; use quality components and implement according to the rules.

My .02 (same ol' .02)

Scott

Scott:

Your .02 are worth quite a bit, thank you. It's so true what you say about flakey cabling can render an otherwise good network inoperable or low-functioning.

I am working with this small business, getting them up to speed technologically. I mentioned in another thread that they were running both a wired and wireless network. Wired for file sharing, wireless for internet... Brilliant. So I took care of that and transferred everything to wired. Speeds improved. Then we replaced the Windows 98/ME boxes (which were infested with AOL and adware galore) with new machines running XP Pro. More improvement, but things still felt slower than they should.

I noticed that almost all cabling was out of spec. Whoever assembled the patch cables was sloppy as hell and left considerable untwisted cable in front of the connector. Once replaced with normal cables, the network, albeit small, was much smoother.

As for why I didn't address the cabling from day 1, my priority was to eliminate this idiotic hybrid wired/wireless situation as quickly as possible. I was doing my best not to laugh at the designer of such a mess.

 
hell hath no fury than a poor cable infrastructure.

"don't fvck with the physical layer" <---spideyism.

There is a load of sciene behind the cable specs, like how long a bit is on the wire - work it out. 1/100 million bits a second = how long the bit is on the wire. factor in speed of light and a general 70% propogation factor (electricity on the wire moves at "about" .70c) and you can figure it out.

Then factor in the minimum frame size of 64 bytes for ethernet and you get to all the fun stuff.
 
I've never seen the OSI model useful outside of a classroom. Considering most teachers can't keep it completely straight I guess it's not that great in the classroom either.
 
Originally posted by: cbehnken
I've never seen the OSI model useful outside of a classroom. Considering most teachers can't keep it completely straight I guess it's not that great in the classroom either.

The OSI model is the most used thing in networking bar none. We live and breathe the OSI model in network design. I can't stress enough how important it is.
 
I have personal experience with failed tests on "handy" little short cables. The pinout was correct, the continuity good, but it failed in crosstalk as expected. I stick with the 1 meter rule, so far so good.
 
So those freebee one foot cat5 patch cables we've all been getting, more than likely for the modem/router patch, are worthless? Figures.
 
Originally posted by: TwoBills
So those freebee one foot cat5 patch cables we've all been getting, more than likely for the modem/router patch, are worthless? Figures.

heh, they're patch cables. Just not CAT5 patch cables.
 
Originally posted by: spidey07
Originally posted by: TwoBills
So those freebee one foot cat5 patch cables we've all been getting, more than likely for the modem/router patch, are worthless? Figures.

heh, they're patch cables. Just not CAT5 patch cables.

Ah, havn't got mine yet, so I missed that. Hopin' for a pink cat 3. 🙂
 
So does each section of cable need to be 1m or just the total run? i.e. If I have a wireless AP going to a wall jack and then the wall jack runs another 100 ft to my closet, can the AP to wall cable be less than 1m or does that need to meet the minimum requirement as well? If it does, how do you guys keep things neat when you want to plug something in right next to the wall (like a wall mounted AP)?
 
It would apply to the entire span.

I.e., AP --> Switch (must be > / = one meter)

AP --> wall jack --> <some normal length of cable> --> Panel -->switch (you're OK)

Assumes proper termination at each junction. Improper termination will cause reflections (SWR) that can also degrade the performance.

Excess cable on the rack / cabinet is usually handled by using cable management products (browse the Anixter catalog @ www.anixter.com and search for "cable management")

Good Luck

Scott
 
Why could a short cable be any more 'unreliable' than a longer one? I dont understand that.
I mean, I can see why there are MAXIMUM cable lengths, but not minimum lengths.
 
One of the main problems (at a very basic level) is that the transmitted signals from each end bleed over to, and interferes with, the received signal from the other end. (Picture two people standing nose-to-nose screaming at each other).

There are other issues relating to impedence / loads, and the feedback loop for each transmitter.

It's not a Good thing.

FWIW

Scott
 
Back
Top