Very slow connection with CAT6 connection between buildings

[MFortie]

Are your ports set to auto-negotiate? Back in the day (from what I remember), that seemed to slow a network down. I always set the port config manually instead of auto...

(It's been a lot of years though! )

YMMV...

 

[wyliesdiesels]

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.

Youre using hubs or switches?

 

[wyliesdiesels]

Are your ports set to auto-negotiate? Back in the day (from what I remember), that seemed to slow a network down. I always set the port config manually instead of auto...

(It's been a lot of years though! )

YMMV...

If the switches are non-managed there is no way to change the negotiation speed.

 

[eegger]

If you connect you PC using the same port with a short cable and speeds are normal, then your router is not the issue. Like you've tested.

If you re-terminate the Cat6 cable does not resolve the issue, then the only solution is to replace the cable.

 

[Innovate1]

Youre using hubs or switches?

Got sloppy there. They are switches. Don't think it makes any difference for this issue.

 

[Stuff]

Assuming not all wires are bad in the cable and you are okay with 100mbs.... Try with only pins 1, 2, 3, and 6 eliminating half the wires. The old 100Base-T should work. If doesn't work then use the other two pairs on 1,2,3,6.

Also verify that you are not color blind and followed the T568B order at both ends. (T568A for government/military)

 

[jeepxj]

Assuming not all wires are bad in the cable and you are okay with 100mbs.... Try with only pins 1, 2, 3, and 6 eliminating half the wires. The old 100Base-T should work. If doesn't work then use the other two pairs on 1,2,3,6.

Also verify that you are not color blind and followed the T568B order at both ends. (T568A for government/military)

for most general internet use 100mb will be plenty. streaming video, music, general web. you will barely notice.

 

[Innovate1]

Wire is solid copper - someone had asked if it was CCA and it is not.

 

[haveissues]

There can really only be one of 3 things wrong-the devices on either end, your terminations, or the cable. If you are sure the first 2 are fine pull a new cable.

 

[wyliesdiesels]

Got sloppy there. They are switches. Don't think it makes any difference for this issue.

Hubs actually could make a difference but those are rare these days as they've all been discontinued...

 

[American Locomotive]

It could be possible that the transmitter/receiver (the PHY) on your router or PC is degrading. Maybe It can do gigabit with a short cable, but not the long cable.

Do you have another router, or another PC (laptop?) where you could try different combinations of devices with the same cable?

 

[HenryAZ]

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.

The "basic" level of negotiated port speed is only that, what the port can do. Transferring data over the wire will reveal the true speed of the point to point link. You can have GB port speeds and still have a <GB data transfer speed. Use a tool such as iPerf to check actual transfer speeds.

 

[Innovate1]

Hubs actually could make a difference but those are rare these days as they've all been discontinued...

Please elaborate on the difference it would make. Just interested in the details as I am not a networking guru...

 

[Innovate1]

The "basic" level of negotiated port speed is only that, what the port can do. Transferring data over the wire will reveal the true speed of the point to point link. You can have GB port speeds and still have a <GB data transfer speed. Use a tool such as iPerf to check actual transfer speeds.

Ok. So as I understand it the negotiated speed is the max the port can do. I know that is bad. So the file transfer will be that bad or worse. It doesn't seem like much is to be gained by doing a file test. What additional information is actual transfer speed going to give? Seems pointless to me but maybe I am missing something.

 

[eegger]

Please elaborate on the difference it would make. Just interested in the details as I am not a networking guru...

it wouldn't, 10/100/1000Gb on a hub is the same speed as it would be on a switch, the difference comes into play when you have a large number of devices.

A hub broadcasts all traffic to all ports, a switch does not, it broadcasts to specific port to port.


If you know what the max a port on a switch is capable and what the client side is capable, ie Say switch port support 1Gb and laptop supports 1 Gb, that they should show as connected as 1 Gb. If not, there is an issue with the cable and or the cable termination ends.

If you already verified that another shorter cable can connect at 1gb, any other test is pointless. Re-terminate the cat6 ends can make sure it's done correctly and if it still doesn't work, the cable would have to be replaced.

 

[Innovate1]

Here's the cable I have. Wireandcableyourway.com said they changed systems about 2 years ago and old orders didn't transfer. I was just over that time but they sent me my previous invoices so happy with their service digging this up.

https://www.wireandcableyourway.com/cat5e-pe-outdoor-gel-type-db-1000ft-spool

Don't see any reason why this shouldn't be working. Maybe my termination technique is poor although it seems good?

 

[eegger]

One thing I see from the link , that is an unshielded cable. A shielded cable helps reduce volume and intensity of all kinds of electrical noise and EMI, along with the adverse effects it has on signals and transmission.


I would do the following, remove the cable from the conduit, you can attach a pull string to one end, if you don't already have a away to pull another or the same cable back in.

Test the connection with that cable, as you would a patch cable. This lets you do 2 things, inspect the cable and if it works fine outside the conduit run, then you know it's interference.

 

[jeepxj]

Here's the cable I have. Wireandcableyourway.com said they changed systems about 2 years ago and old orders didn't transfer. I was just over that time but they sent me my previous invoices so happy with their service digging this up.

https://www.wireandcableyourway.com/cat5e-pe-outdoor-gel-type-db-1000ft-spool

Don't see any reason why this shouldn't be working. Maybe my termination technique is poor although it seems good?

did you redo the ends yet?

 

[RAYJAY]

one thing i had a problem here on my whole house install , first thing did you direct connect cable to modem and check speed ?
my problem was a a fire tv was dragging down my system,

 

[Innovate1]

did you redo the ends yet?

Yes. I redid them once switching from male ends to keystones. And I terminated a spare run with keystones. hard to believe I did that bad of a job several times...

 

[jeepxj]

Yes. I redid them once switching from male ends to keystones. And I terminated a spare run with keystones. hard to believe I did that bad of a job several times...

with new ends did you bypass switches on both sides? laptop at far end. router/modem on home end?

 

[Innovate1]

with new ends did you bypass switches on both sides? laptop at far end. router/modem on home end?

used a switch on one end and PC on the other. Also tried router on one end, PC on the other.

 

[jeepxj]

used a switch on one end and PC on the other. Also tried router on one end, PC on the other.

When you do the PC on the one end what does network connections show as the link type?

 

[jeepxj]

if you wanna get super nerdy:

but this is getting deep under the covers.

 

[shade]

Just run a new cable from point a to point b.
Not through the conduit.
Does it work right?
This will rule the cable in or out

 

[Innovate1]

When you do the PC on the one end what does network connections show as the link type?

10 Mb

 

[jeepxj]

10 Mb

both sides set to auto negotiate in settings?

while not unheard of looks like a cable failure.

 

[Innovate1]

both sides set to auto negotiate in settings?

while not unheard of looks like a cable failure.

I would agree but TDR doesn't show much loss or any other issue that I can see and both runs behave the same. Could the whole spool be bad but in a subtle way that isn't obvious? I do have a most of a spool of CAT6 (not direct burial) that I could string out above ground for same length to test that the end hardware doesn't have any issue with that length.

Both sides auto negotiate.

 

[CoogarXR]

Just for shits'n'giggles, can you show a picture of your terminations?

 

[jeepxj]

I would agree but TDR doesn't show much loss or any other issue that I can see and both runs behave the same. Could the whole spool be bad but in a subtle way that isn't obvious? I do have a most of a spool of CAT6 (not direct burial) that I could string out above ground for same length to test that the end hardware doesn't have any issue with that length.

Both sides auto negotiate.

you have 2 cables in the same conduit run that are only connecting at 10mb after redoing all 4 ends? assuming nothing else changed physically, ie installing a nice 7200v transmission line on them or something i'm going with its in the hardware on the ends.

can you try setting both sides to 100mb half duplex. does it connect at 100?

 

[jeepxj]

Just for shits'n'giggles, can you show a picture of your terminations?

what its fineeeeee. jk

 

[RPH]

I’ve done a lot of work on industrial communications systems for machinery. Bad comm cable does show up on a regular basis.
Ford had bought a full run of cable for a large product. Vendor was told to toss it once it was discovered to be bad.
Vendor didn’t, got used on another ford project. The **** hit the fan 2 days later. It wasn’t pretty.
But try new sourced cable between the two points. Best advice so far that I’ve seen.

 

[jeepxj]

hell even just use a random cable you already have. test it out with both in the same room kinda deal. do they connect at better speeds. then take the same devices and use the long troublesome runs.

 

[Innovate1]

hell even just use a random cable you already have. test it out with both in the same room kinda deal. do they connect at better speeds. then take the same devices and use the long troublesome runs.

Already done. Equipment connects with shorter (premade) cables at Gb speed.

 

[jeepxj]

Already done. Equipment connects with shorter (premade) cables at Gb speed.

sounds like just bad cables then. you've covered pretty much everything at this point.

 

[kaffine]

Do any of your network devices have built in cable tests? I know the one I've been working with will give length of each pair. It also can provide a signal to noise ratio.

Wireshark isn't going to help with auto-negotiation issues the signals don't go past the PHY.

You can look at the autonegotiation signals with your scope. Look at them with the devices near each other and connecting at 1G and then with the long cable connected.

 

[Innovate1]

No built in cable tests that I know of. Just consumer switches and routers. I do have ddwrt on the routers so might be some way to do cable tests with that. I do have a $50 cable tester that gives length of each pair and it gives length of each pair. It shows 95.4 meters for all pairs which is high. The most it actually is about 260 feet which is about 15% error, maybe due to velocity factor?

I may try the scope on the signals just to see what I can see. Interested in the details but also tired of fighting and looking at going to fiber.

 

[jeepxj]

No built in cable tests that I know of. Just consumer switches and routers. I do have ddwrt on the routers so might be some way to do cable tests with that. I do have a $50 cable tester that gives length of each pair and it gives length of each pair. It shows 95.4 meters for all pairs which is high. The most it actually is about 260 feet which is about 15% error, maybe due to velocity factor?

I may try the scope on the signals just to see what I can see. Interested in the details but also tired of fighting and looking at going to fiber.

honestly you're so deep into this we gotta know the root cause.

 

[tsperry88]

Just doing a quick scan of this thread, I agree with those that said paralleling power, in close proximity, over a long distance could cause issues. Shielding would help, if one end is properly grounded.

 

[HeadsUp]

Shielding needs to be properly connected to return at both ends. Otherwise it's an antenna