SuperNoobyQuestion: Patch cable signal deterioration

Forums › LAN (ethernet/Wifi/routers/Bluetooth) › SuperNoobyQuestion: Patch cable signal deterioration

McStag

8 posts

Wannabe Geek

#17932 18-Dec-2007 18:00

Hey im new to this website and just an aspiring geek.
Im trying to set up a conection to our router from my sleepout via patch cable.
It looks as though im going to have to use a 30m cable.
Will there be any signal deteriation i.e will my internet speed be slower than a direct conection??

grant_k

3539 posts

Uber Geek
+1 received by user: 124

Trusted

#100717 18-Dec-2007 18:32

If possible with your router, it would be good to force the port to 10Mbps rather than 100Mbps. However, many of the cheaper routers don't support this so it may not be possible.

I tried to connect a Linksys router directly to 90 metres of Ethernet cable and it wouldn't work. However, by using an old Wireless Router with a 4-port Hub and Uplink port -- basically just using it as a hub -- I was able to get the Linksys talking over the 90 metre cable.

Hopefully you won't have that sort of problem at 30 metres, but it will depend on how much electrical noise is present in the vicinity. If you still can't make it work, I would suggest using a WiFi link, but a wired link will definitely be cheaper so best to try it first.

To answer your final question: Will it be any slower?

No, not at all. Because the slowest speed on Ethernet is 10Mbps whereas your typical ADSL internet connection is around 5Mbps or slower. So your internet connection will always be the bottleneck, not the other way around.

P.S. Welcome to Geekzone Sam

cyril7

9075 posts

Uber Geek
+1 received by user: 2499

ID Verified

Trusted

Subscriber

#100726 18-Dec-2007 19:18

Hi Sam, no there should be no reduction in performance on a 30m length, infact the standard for ethernet is that you can have 100meter runs and still maintain full 100Mb/s performance. I have run lengths of over 130meters and still had 100Mb/s connections.

Things to keep in mind, treat your cables with care, dont put tight kinks in them, dont place them under pressure (ie dont put heavy weights on them) all these things will effect the return loss (ie put reflections on the line).

Just for you info, my office is down at our garage, there is a 80meter cable between all the routers/switches in the house and the garage. Using wsttcp.exe as a test tool this cable maintains a 87Mb/s TCP through put which is as good as a 1meter cable.

I have had cat5e cable at 70meter pass 1Gb/s circuits with no limitation, 100Mb/s will be a breeze, just treat the cable with care.

Cyril

grant_k

3539 posts

Uber Geek
+1 received by user: 124

Trusted

#100732 18-Dec-2007 19:45

cyril7: Hi Sam, no there should be no reduction in performance on a 30m length, infact the standard for ethernet is that you can have 100meter runs and still maintain full 100Mb/s performance. I have run lengths of over 130meters and still had 100Mb/s connections.

Hmmm, I'm very interested to hear you say that Cyril. I wonder why I had problems with my 90 metre run of Cat5E cable then?

The situation was this:

- AirSpan Wireless Transceiver at the far end (no ground connection)
- Running 10Mbps over the cable back to our cottage
- 3 different ZyXEL routers were quite happy communicating over this cable for more than 3 years until 2 of them got spiked out by lightning on their WAN ports during 2 separate strikes this year
- Replaced with Linksys WRTP54G so I could reduce the amount of wall-warts and cables behind my desk (PAP2T + ZyXEL = WRTP54G)
- Hmmm, Linksys will not communicate reliably with AirSpan
- Plugged in my laptop and that wouldn't communicate over the cable either
- Plugged in my wife's laptop and ditto
- Plugged in my desktop PC, and that worked fine
- After experimenting a bit, tried an old Wireless Router with 4-port Hub and separate Uplink port
- Plugged cable from AirSpan into Uplink port, then plugged Port 2 of the Hub into Linksys and all happy

Sometime later, I finished installing Lightning Protectors at each end of the cable as well as a 4 metre piece of water pipe clamped to the mast extending around 0.5m above the 2.4GHz grid antenna. Connected water pipe via a THICK piece of cable to a 2-metre ground spike.

At this stage I thought it might be a good idea to connect the antenna pole to ground as well, just to keep any lightning from frying the AirSpan transceiver. However -- this stopped the ethernet connection dead in its tracks. Disconnected the link to the antenna pole again and the ethernet started up again

What on earth is going on here, I have been wondering. Relevant points are:

- The AirSpan transceiver's case is connected to the shield of the antenna coax as you would expect.
- The shield of the antenna coax is connected to the antenna frame.
- So, by grounding the antenna pole, I was effectively grounding the case of the AirSpan transceiver and this was causing the ethernet link to fail.
- This is very strange because ethernet is transformer-coupled, so it shouldn't have been affected at all.

Clearly my ethernet connection at 10Mbps over 90 metres of Cat5E cable is right on the boundary of go/no-go.

But why?

If you have any ideas Cyril, I would be most interested to hear them

dan

1134 posts

Uber Geek
+1 received by user: 105

Lifetime subscriber

#100736 18-Dec-2007 20:00

30 metres is no problem, i have a customer who is running a cable from 1 warehouse to the next, to connect 2 LANS,
atleast 60 metres, at 1gbit speed.

cyril7

9075 posts

Uber Geek
+1 received by user: 2499

ID Verified

Trusted

Subscriber

#100749 18-Dec-2007 20:47

Hi Grant, interesting issue. Just to confirm I have installed quite a number of 100meterplus lines no issue. You definitely from your description have an earthing issue. What I have found is that a lot of routers and single port adsl modems infact dont have isolated ethernet ports. I suspect this is for two reasons, firstly there is only one port, and secondly they are hanging on a isolated power supply. It might be worth putting a switch between the two ports to see if it isolates the problem, switches always are transformer isolated. Worth a try.

But I definitely stand by my cable length expectations. The standard says you should expect 100m, in practice I have found you can exceed this by 50% and still get full performance. I did do an exercise with eye closure monitoring some years ago on ethernet ports, again I found 100-120meters was well within spec.

Cyril (aka Stephen)

grant_k

3539 posts

Uber Geek
+1 received by user: 124

Trusted

#100752 18-Dec-2007 20:58

Thanks very much for your comments Stephen

(yes, I did remember your real name, but wasn't sure if you wanted it publicised on Geekzone)

It's interesting that you have found that single port modems frequently don't have isolated ethernet ports. From what I have seen in my installation, I suspect that this is the case with the AirSpan. I am powering it via a DC-DC converter, so there is no earth connection at all, unless I deliberately provide one via the antenna pole. Going by what happened when I tried grounding the pole, it sounds like I am best to leave it ungrounded.

By using the old WAP as a switch I have in a back-to-front kind of way, done what you suggested. The transformer isolation on the switch ports has definitely done the trick and the system is 100% reliable as it stands.

Once again, thanks for your comments, especially on the lack of isolation issue. It's good to have an explanation for what I have observed.

Cheers,
Grant.

P.S. If you're ever up our way, be sure to stop in and pay us a visit