Very slow connection with CAT6 connection between buildings

[Innovate1]

I have a run of about 230 feet in the ground in conduit between buildings. It is cat6 cable. I think it is gel tape rather than gel filled - checking on the details of that. BTW, I know I ordered my building wire from wireandcableyourway and maybe this wire too but when I log in it says I have no history of orders. It was about 2 years ago and surprised it is purged so quickly and have sent them a request for order info.

I'm not sure what the connection speed was at first but it was working but has slowly gotten worse and is pretty unusable now. I used a scope and pulse generator as simple TDR and didn't see any issues. Checked continuity with a tester and all lines show good. Checked resistance of all wires end to end and that looked ok. When I do end to end connection between router and pc the status shows 10 Mb for speed. If I connect with a short cable between equipment I get Gb speed.

Wondering if water has gotten into the cable. Anyone have experience with burying the gel tape type cable. Apparently it isn't quite as good as the true gel filled? Thought about pushing a small poly tube down the conduit on the low end and sucking out any accumulated water but it might take forever for water to work itself out of the cable if that's the issue.

I have another run of the cable used for much lower speed low voltage connection and it is working fine.

 

[loganb]

What router are you using? It may be the limiting factor. CAT 5 is rated for 100m runs generally so length wise you should be fine.

If you're afraid there is water in the line, can you blow It out with compressed air from one end?

 

[Innovate1]

The router has 1G ports and is the same equipment I am connecting to with a short patch cable for comparison. I am not testing by putting data through but just using what a windows PC reports that status of the port at. So any bottlenecks in the router don't come into play in that status (but would of course in actual operation, possibly making the connection slower than the reported status speed if the status speed was where it should be).

 

[Stuff]

Did you check resistance wire-to-wire? Regardless you have the cable in conduit so pull it out, examine it, test it, and replace it. It is a remote possibility that the conduit failed and is pinching cable. I assume the cat6 isn't run in the same conduit as power.

As a workaround try putting a switch/router in front of the PC. Sometimes dedicated network hardware negotiates speed differently/better than a PC.

 

[Wrench97]

Is the Cat6 wired to the correct standard on both ends?

 

[Innovate1]

Did you check resistance wire-to-wire? Regardless you have the cable in conduit so pull it out, examine it, test it, and replace it. It is a remote possibility that the conduit failed and is pinching cable. I assume the cat6 isn't run in the same conduit as power.

As a workaround try putting a switch/router in front of the PC. Sometimes dedicated network hardware negotiates speed differently/better than a PC.

I had routers/switches at both ends and wasn't getting problems with connection and speed. The PC was just used as a test and not the normal configuation.

 

[Innovate1]

Is the Cat6 wired to the correct standard on both ends?

Yes.

 

[CoogarXR]

Is this a pre-made cable, or did you install the ends?

 

[Innovate1]

Is this a pre-made cable, or did you install the ends?

I installed the ends. Reterminating the cable is something that is usually suggested with user installed ends. I have a space run and will probably terminate that instead and see how that cable works.

 

[jeepxj]

if its only connecting at 10mb then you got something funky. i'd re terminate one end, see if it connects better. if not then do the other. i'd also try a different switch or do a direct connection from a laptop to the modem.

 

[Miss the Pontiacs]

Have you a picture of your installed ends. Anyway possible that cable is/was kinked or stepped. on prior to installation. Cable quality?? Gel tape so cable is flat, no twisting of the four pairs within the sheath?

 

[Noltz]

Did you bury anything else in the trench when you ran the wire?

 

[theoldwizard1]

I know on most 80s/90s mini-computers and workstations, you could actually get the number of errors/retransmits on an Ethernet port. Not sure about today's ports. Also, something WinDoz probably does not do !

 

[haveissues]

I suggest triple checking your terminations. 1+2, 3+6 need to be twisted pairs. If it isn't wired that way continuity will be fine but the cable will do exactly what it is doing-link at 10mb and sort of function.

 

[CoogarXR]

I installed the ends. Reterminating the cable is something that is usually suggested with user installed ends. I have a space run and will probably terminate that instead and see how that cable works.

I personally prefer installing cables to a punched-down connection. Either a punch-down patch panel or a female keystone jack. And then use pre-made patch cables to connect to the devices. The crimp-on RJ-45 ends can be a little flaky. I still use them here and there, but for "critical" connections, I prefer a punched-down connection over a crimp.

And I also prefer solid copper cable. Some of the cheap stuff is copper-clad aluminum (CCA). I'm not sure if it's junk, but it sounds like it would be, lol. I only ever buy solid copper.

 

[Innovate1]

Did you bury anything else in the trench when you ran the wire?

AHJ said gas line had to be min 2' from electric power. Excavator had a 2' bucket. Electric was on one side, gas on the other with low voltage in the middle. All electric was in conduit so no chance for it to move around much. We made sure things stayed in place as we covered it and put temporary stakes in a few places to keep things in position until they were covered. So spacing was pretty consistent - I don't see how we would have significant interference from the power.

 

[Innovate1]

Have you a picture of your installed ends. Anyway possible that cable is/was kinked or stepped. on prior to installation. Cable quality?? Gel tape so cable is flat, no twisting of the four pairs within the sheath?

I inspected the ends carefully - it looked like all the wires were in the proper positions. No, the cable isn't flat. It's a typical round cable. The gel tape is just inside the outer sheath. I think it has a bit less gel than gel filled.

Reterminating with keystone jacks. So much easier and sure way to get all the wires into the correct positions.

 

[walta]

Have you checked the cable for open circuits, shorts, miswires, reversals, and split pairs?

This is possible with a meter but for a human but given 8 wire you have 64 tests to run and analyze the chance of mis interrupting the data is very high.

They make cheap testers ($35.00) to run all the combos and throw a red light in the event of any errors.

https://www.amazon.com/dp/B08DZW212Y/?tag=atomicindus08-20

This tester will not tell you about the quality of the connection this is about length of the wire and the highest frequency the wire can carry. For this you need Verifier a 600-2000$ tool.

If I had this problem I would lay a temporary cable on the surface and see if that would eliminate the problem if it passed the cheap tester.

Walta

 

[wyliesdiesels]

What router are you using? It may be the limiting factor. CAT 5 is rated for 100m runs generally so length wise you should be fine.

If you're afraid there is water in the line, can you blow It out with compressed air from one end?

he stated he connected the PC with a short cable to the router and it was working fine.

 

[wyliesdiesels]

have you tested the cable with an ethernet pair tester?

 

[Noltz]

AHJ said gas line had to be min 2' from electric power. Excavator had a 2' bucket. Electric was on one side, gas on the other with low voltage in the middle. All electric was in conduit so no chance for it to move around much. We made sure things stayed in place as we covered it and put temporary stakes in a few places to keep things in position until they were covered. So spacing was pretty consistent - I don't see how we would have significant interference from the power.

Try the Keystone terminators. You're 12" from a strong power line & running parallel for 230'. That worries me. Grab a multimeter, unplug both ends and connect your ground lead to earth. Check your pins with a strand of copper wire (don't booger jab the test lead into it ) and look for induced voltages. I am worried you're too close.

You mentioned "..another run of the cable used for much lower speed low voltage connection and it is working fine.". Is this in the same conduit? Is it AC low voltage?

 

[justsam]

Try the Keystone terminators. You're 12" from a strong power line & running parallel for 230'. That worries me. Grab a multimeter, unplug both ends and connect your ground lead to earth. Check your pins with a strand of copper wire (don't booger jab the test lead into it ) and look for induced voltages. I am worried you're too close.

You mentioned "..another run of the cable used for much lower speed low voltage connection and it is working fine.". Is this in the same conduit? Is it AC low voltage?

The OP mentioned using a scope and signal generator as a make-shift TDR. I suspect he would have seen any induced AC on the line at that time.

 

[luvtheheat]

You're 12" from a strong power line & running parallel for 230'. That worries me.

Yup!

 

[BlindViper]

is the cable rated for underground?

 

[dogdog]

The only time I have seen this happening is when I didn't use the T568-A or T568-B standard to crimp my connectors... I just chose any random matching straight color coding and it tested fine on those cheapie testers... means 1-1 2-2 3-3 etc... but the routers/switches doesn't seems to like it and I couldn't get pass 10mbps.. maybe before digging it up, cut both end and re-crimp and verify ?

that was asked earlier... Post 5.

 

[jblnut]

I could ship you my DXA2-8000 if you promise to send it back so you can test the snot out of that cable

 

[Metal-Marc]

Perfect opportunity to switch to fiber.

 

[theoldwizard1]

Yup!

Ethernet cable uses differential signalling over twisted pair. VERY NOISE IMMUNE ?

 

[Innovate1]

Try the Keystone terminators. You're 12" from a strong power line & running parallel for 230'. That worries me. Grab a multimeter, unplug both ends and connect your ground lead to earth. Check your pins with a strand of copper wire (don't booger jab the test lead into it ) and look for induced voltages. I am worried you're too close.

You mentioned "..another run of the cable used for much lower speed low voltage connection and it is working fine.". Is this in the same conduit? Is it AC low voltage?

Is that a technical term? LOL...

Reterminated on both ends with keystones - no difference.

Ethernet is isolated (to 1500V) on both ends and uses differential signals. Take a 25' piece of wire in your house and you will probably get some voltage because the impedance of a modern meter is megaohms and doesn't provide a realistic load.

I measured just to see what I would get. With both ends of the ethernet cable unconnected I get 2.6VAC. I measured another cable in the house about 25' long that goes through some joists and was routed to avoid power wires as much as possible - it measured 30VAC and it works fine. The twists of the wires cancels out any differential effects.

Just to test that I turned off ALL the electric - power and low voltage (no, the low voltage isn't AC) in the trench and to the shop. Used a battery powered laptop on the shop end. No difference.

Shorted the wires one by one to test resistance on the other end. Thought I had found a bad wire but I couldn't repeat it. Must have been a poor connection of a jumper wire.

Terminated a spare cable. Same thing.

Checked for water in the conduit by running some small tubing down about 6' into the low end. No water that I could pull out - if there is any it's very little.

Is it possible it's just crappy cable? Trying to figure out exactly what it is. Think it is Cat5e (said 6 earlier but now that I think about it I think it wasn't 6). The markings are all worn off from pulling it in. I tried to look up past orders at Monoprice and wireandcableyourway and both say I have no previous orders. I ordered from both but it's been 18 - 24 months. Seems like they should still show that. I have sent emails to both to try to find out previous orders.

So it looks like I need to get better cable or change technologies - optical fiber or radio. It's a bit too far for wifi.

 

[Innovate1]

is the cable rated for underground?

Yes. It was bought for this use.

 

[Noltz]

Ethernet cable uses differential signalling over twisted pair. VERY NOISE IMMUNE ?

I measured just to see what I would get. With both ends of the ethernet cable unconnected I get 2.6VAC. I measured another cable in the house about 25' long that goes through some joists and was routed to avoid power wires as much as possible - it measured 30VAC and it works fine. The twists of the wires cancels out any differential effects.

Yes they are. In automotive we used twisted pair for decades for this reason. But I know from personal "I did it" experience that running network cables beside 120v conductors we lost communication to an office computer <50' from the router. My cable testers (cheap Amazon ones) said all was well. I moved the cat cable 4' away from the AC lines and boom, we had connection again. That's why I brought it up.

Just to test that I turned off ALL the electric - power and low voltage (no, the low voltage isn't AC) in the trench and to the shop. Used a battery powered laptop on the shop end. No difference.

It's very clear you know what you're doing from your reply, and this would have been my next suggestion. Good call on using a laptop on battery to test. I really appreciate you posting this issue actually. I've put an offer in on my next home and I'll finally have my own shop if we can strike a deal - which mean I too will need to run a cable out to the building. I'd be interested to know what works for you!

 

[rdoty]

Do you have an Ethernet router or switch at each end of the cable? It would be interesting to connect two devices to the same switch and see if they communicate at full speed. This would help determine if the issue might lie with the switches instead of the cable.

Modern Ethernet switches can have a different speed on each port. If, for example, you have a 5 port switch with two laptops connected at 1 Gb and a printer connected at 100 Mb, the laptops will talk to each other at 1 Gb and each laptop will talk to the printer at 100 Mb.

Try doing a large file transfer over the long cable and see what you get for errors and throughput.

 

[iggy005]

I think most was covered already but if I had one around I would try a new cable. I work on outdoor wireless radios often that connect Ethernet cable. A couple of other thoughts...

Have you tried 2 different devices on each end like laptop to laptop? Could be the one device not negotiating properly at that distance.

Is it actually 230 feet including the cable going in? I have seen it work fine up to 350 but 300 feet is the standard max.

Any tight zip ties holding the cables on either side? Believe it or not I have seen seen tight ties mess up the signaling going through correctly. Also anything that could crush the cable isn't good either.

To me it sounds like a negotiating issue. 10m and 100m require 2 working pairs and if either are broken it would not work at all. With that noise would be the easy culprit such as noise from an electrical line. But that is out since you tested that. Then next most likely is a bad device on either end. It may work on a short line however the longer the cable is it may have problems negotiating at that distance. Otherwise a bad cable could relate to incorrect twits on each pair not canceling noise correctly. There is also a chance of incorrect power. I had one person that had improper voltage on the home that was causing all sorts of strange things with low voltage signalling. A laptop on battery on both ends would rule this out.

Good luck!

Also to note very rarely water in the line will do anything to an Ethernet cable unless the conductor is exposed inside or water weeps to the connector end. If you ever worked with gell it ***** but the powder version is just fine in conduit.

 

[wyliesdiesels]

The only time I have seen this happening is when I didn't use the T568-A or T568-B standard to crimp my connectors... I just chose any random matching straight color coding and it tested fine on those cheapie testers... means 1-1 2-2 3-3 etc... but the routers/switches doesn't seems to like it and I couldn't get pass 10mbps.. maybe before digging it up, cut both end and re-crimp and verify ?

that was asked earlier... Post 5.

Thats because each pair has different twists per inch.

When you dont follow the correct pin-out, you will have issues.

 

[wyliesdiesels]

I think most was covered already but if I had one around I would try a new cable. I work on outdoor wireless radios often that connect Ethernet cable. A couple of other thoughts...

Have you tried 2 different devices on each end like laptop to laptop? Could be the one device not negotiating properly at that distance.

Is it actually 230 feet including the cable going in? I have seen it work fine up to 350 but 300 feet is the standard max.

Any tight zip ties holding the cables on either side? Believe it or not I have seen seen tight ties mess up the signaling going through correctly. Also anything that could crush the cable isn't good either.

To me it sounds like a negotiating issue. 10m and 100m require 2 working pairs and if either are broken it would not work at all. With that noise would be the easy culprit such as noise from an electrical line. But that is out since you tested that. Then next most likely is a bad device on either end. It may work on a short line however the longer the cable is it may have problems negotiating at that distance. Otherwise a bad cable could relate to incorrect twits on each pair not canceling noise correctly. There is also a chance of incorrect power. I had one person that had improper voltage on the home that was causing all sorts of strange things with low voltage signalling. A laptop on battery on both ends would rule this out.

Good luck!

Also to note very rarely water in the line will do anything to an Ethernet cable unless the conductor is exposed inside or water weeps to the connector end. If you ever worked with gell it ***** but the powder version is just fine in conduit.

Max is actually 100m.... 90m horizontal, 10m patch cord

 

[jblnut]

If I read correctly the cable is in conduit the entire way. Can you just pull a new piece of wire in using the old wire as the pull string ?

If it was working and suddenly crapped out I'd say that's the next step. If it is large enough you can pull in a preterminated fiber assembly but it sure isn't necessary at that distance. You could tie a string to the old wire and pull the string in while pulling the wire out to measure the wire length and have the string in there to pull in the new fiber. Preterminated stuff is pretty inexpensive these days and switches with SFP's aren't all that expensive either.

Or just pull in another gell filled Cat6. Should work just fine.

EDIT: OP, is it CCA cable ? Copper Clad Aluminum?

 

[Innovate1]

I could ship you my DXA2-8000 if you promise to send it back so you can test the snot out of that cable

interesting option but not sure you are really serious. I plan to ask around locally to see if I might borrow a more complex tester. Would be interesting to know what the real issue is. Then there is the question of operator skill to run the equipment and know how to interpret the results.

 

[Innovate1]

Do you have an Ethernet router or switch at each end of the cable? It would be interesting to connect two devices to the same switch and see if they communicate at full speed. This would help determine if the issue might lie with the switches instead of the cable.

Modern Ethernet switches can have a different speed on each port. If, for example, you have a 5 port switch with two laptops connected at 1 Gb and a printer connected at 100 Mb, the laptops will talk to each other at 1 Gb and each laptop will talk to the printer at 100 Mb.

Try doing a large file transfer over the long cable and see what you get for errors and throughput.

Initial testing was with a switch on one end and router on the other. when I first put it in I couldn't get connection. I shortened the cable by about 25' by putting a hub right where the cable entered the house. It worked ok so I didn't do a lot of checking on the speed it was negotiating and transfer speed. Guessing that was about 18 month ago. The hub has 100Mb ports which is plenty for my use.

Recently I was using a laptop and the hub. The laptop has a Gb port. Win10. On the other end the router has Gb ports. I used the laptop with short cable to test all the ports on hub and router and all connected at their rated speeds. At one point I pulled another hub that had Gb ports to replace the slower hub but it didn't change anything.

I understand that each port may negotiate different speeds. My tests have been with just the device on each end without other things connected.

What would doing a large file transfer tell me that I don't already know? I did a repeated ping and about 10% failed on timeout. Others seems to have reasonable times.

 

[Innovate1]

Max is actually 100m.... 90m horizontal, 10m patch cord

Yes, spec is 90m cable with two 5m patch cables.

90m is 295 feet. basically 300 ft.

My building to building run is 230 ft. No additional run on one end. About 20' on the other so 250'. Still well under 90m/295 ft even without the additional patch cable length allowance.

 

[Innovate1]

I think most was covered already but if I had one around I would try a new cable. I work on outdoor wireless radios often that connect Ethernet cable. A couple of other thoughts...

Have you tried 2 different devices on each end like laptop to laptop? Could be the one device not negotiating properly at that distance.

Is it actually 230 feet including the cable going in? I have seen it work fine up to 350 but 300 feet is the standard max.

Any tight zip ties holding the cables on either side? Believe it or not I have seen seen tight ties mess up the signaling going through correctly. Also anything that could crush the cable isn't good either.

To me it sounds like a negotiating issue. 10m and 100m require 2 working pairs and if either are broken it would not work at all. With that noise would be the easy culprit such as noise from an electrical line. But that is out since you tested that. Then next most likely is a bad device on either end. It may work on a short line however the longer the cable is it may have problems negotiating at that distance. Otherwise a bad cable could relate to incorrect twits on each pair not canceling noise correctly. There is also a chance of incorrect power. I had one person that had improper voltage on the home that was causing all sorts of strange things with low voltage signalling. A laptop on battery on both ends would rule this out.

Good luck!

Also to note very rarely water in the line will do anything to an Ethernet cable unless the conductor is exposed inside or water weeps to the connector end. If you ever worked with gell it ***** but the powder version is just fine in conduit.

Most of the points/questions have already been answered.

No zip ties or crush points. Even if there were an issue in one cable it is unlikely it exists in the second one...

No issues with power and have tried multiple units on the ends.

Someone suggested file transfers through the hubs/switches. I am looking at an even more basic level - what speed is negotiated for the port. If that is bad how can the file transfer not be bad? The bandwidth of the box is not even involved in the port speed so even if the box was a bottleneck it isn't the issue AFAICT.