[ellisr63]

I just upgraded our internet from 100Mb down load to 250Mb, and we have a problem with getting the speed throughout our LAN.

When the Cable Company came out to fix our speed problem they found out our Modem was fine, but the wires to the router were wrong (wall jacks). They fixed it, and we had the proper speed to the router, although the guy said I need to check the other wires to make sure they are correct. A friend(used to be a LAN Administrator) come over, and wired the wall jacks up by our router, and then I hooked up the switch. We had the speed back, but only for a little while. I decided to by a Klein Tools VDV526-052 LAN Scout Junior Tester to test out the wires. What I found was that most of the wall jacks were not wired properly (except for the ones my friend, and the Cable Guy wired), and I rewired them properly (checking to make sure with the tester from the router to the wall jack). Now sometimes I have less than 1Mb when tested to over 200Mb.

Another strange thing is when testing the speed sometimes it will be less than 1 Mb, and then shoot up to anywhere from 20Mb to 150Mb or so at the very end of the download test. If I run the test right afterwards...sometimes it will go to max speed where it was only less then 1Mb a few seconds earlier.

My LAN equipment is as follows:

1: In the HT... Cisco SG200-26 Gigabit Ethernet Smart Switch with 24 10/100/1000 Ports and 2 Combo Mini-GBIC Ports
2: Where the Router is... NETGEAR Nighthawk AC1900 Dual Band Wi-Fi Gigabit Router (R7000) with Open Source Support
Netgear GS108E ProSafe Plus 8-Port Gigabit Switch
3: Family Room... Netgear GS108 ProSafe Plus 8-Port Gigabit Switch
4: Master Bedroom...Netgear Powerline 1200 and Extra Outlet (other end is at the Router location)
5: All cables are Cat6, but unknown kind in the walls.

I have Jumbo Frames turned off.


Any ideas what is causing this problem? The cable guy said he has never seen wires be wrong and still get 100Mb downloads (like we had prior to the upgrade).

 

[theJman]

When you say "wires to the router were wrong" and "wall jacks were not wired properly" what are you referring to? If this is ethernet RJ45 connectors they'll either work or they won't, at least that's been my experience. I can't envision a scenario where crossed wires would lead to anything other than a total lack of connectivity.

What end point are you using when running the speed test? There appears to be a lot of different gear in that signal chain, depending upon where you are in the house. Just out of curiosity, have you double checked the results with something like speedtest.net? I wonder if the two would exhibit the same issue. Another thought, with regards to all that equipment you have... are they running the latest BIOS/firmware/driver version? You made a major jump in throughput, so it could be one of those components needs updated code to handle it.

If you've gotten over 100Mb even once then the cables in the wall have to be at least cat5e. Cat5 stops at 100Mb, so it's probably not cable grade that's the issue. Does any of the wiring go near things with a large motor like an A/C unit, refrigerator or even florescent lights? Any one of those could generate enough RF to cause wacky issues. It could also be a problem outside your house, one the ISP would have to look into deeper; those numbers fluctuate wildly, giving it the appearance of flapping on the line. If so, there's zero you could do about it at the end point - that would be on them to rectify.

 

[ellisr63]

When you say "wires to the router were wrong" and "wall jacks were not wired properly" what are you referring to? If this is ethernet RJ45 connectors they'll either work or they won't, at least that's been my experience. I can't envision a scenario where crossed wires would lead to anything other than a total lack of connectivity.

What end point are you using when running the speed test? There appears to be a lot of different gear in that signal chain, depending upon where you are in the house. Just out of curiosity, have you double checked the results with something like speedtest.net? I wonder if the two would exhibit the same issue. Another thought, with regards to all that equipment you have... are they running the latest BIOS/firmware/driver version? You made a major jump in throughput, so it could be one of those components needs updated code to handle it.

If you've gotten over 100Mb even once then the cables in the wall have to be at least cat5e. Cat5 stops at 100Mb, so it's probably not cable grade that's the issue. Does any of the wiring go near things with a large motor like an A/C unit, refrigerator or even florescent lights? Any one of those could generate enough RF to cause wacky issues. It could also be a problem outside your house, one the ISP would have to look into deeper; those numbers fluctuate wildly, giving it the appearance of flapping on the line. If so, there's zero you could do about it at the end point - that would be on them to rectify.

Thanks for responding JMan.

1: The wires to the router from the modem go through a wall jack in the modem room, and through another wall jack to the router room. The Tech found that 2 pairs were flipped. He said he couldn't believe it had been working at 100Mbps, but he saw it with his own eyes.

2: I used Speedtest.net as well as the one on Wave Broadband with the same results. I checked the firmware versions, and all are on the latest firmware.

3: Some of the outlets have cat5 jacks, and some have cat3. I am only hooking up the cat5 jacks though.

I was wondering last night if the wiring in the house might not be up to speed as the 2nd floor remodel was done in the 1980s I believe. I know the HT has cat5e or cat6 wiring as we put it in when the HT was built. If we had to rewire...we could rewire the modem to router as we have a sub floor. Our Living Room to router also has Cat5 or Cat6. Maybe I should disconnect all the other rooms (with unknown wall wiring), and see if it improves the signal??

 

[theJman]

Thanks for responding JMan.

Because of what I do for a living I simply couldn't help myself... :T

1: The wires to the router from the modem go through a wall jack in the modem room, and through another wall jack to the router room. The Tech found that 2 pairs were flipped. He said he couldn't believe it had been working at 100Mbps, but he saw it with his own eyes.

Something doesn't sound right; unless the wires were flipped at two junctions - so they essentially reversed the reversal - then I can't see it having ever worked. Swapped leads shouldn't have limited speed, it should have eliminated communication entirely.

I was wondering last night if the wiring in the house might not be up to speed as the 2nd floor remodel was done in the 1980s I believe. I know the HT has cat5e or cat6 wiring as we put it in when the HT was built. If we had to rewire...we could rewire the modem to router as we have a sub floor. Our Living Room to router also has Cat5 or Cat6. Maybe I should disconnect all the other rooms (with unknown wall wiring), and see if it improves the signal??

Cat3 came out in the early to mid 90's, so if the remodel was done in the 1980's it predates that by several years. It was also 10BASE-T, so 10 meg speeds. If you were getting 100 meg before that implies 100BASE-T, so cat5 at minimum.

Ethernet topology is point to point, meaning there's a single cable run from the switch port to each termination (generally that's a wall plate with a female RJ45). Since they're independent of each other, and therefore technically isolated, what you have plugged into any one port should not have an affect on the other ports. You can try isolating the slower devices but I can't imagine it would make any difference.

Are you using those GBIC's by chance? If so, I wonder if there's a problem during the signal conversion.

 

[robbo266317]

Just a point about the cabling being wrong. You can have straight through cabling, or swap-over, as most equipment auto-detects it.
Maybe it had been wired as swap-over for some reason? ( http://www.wiringwizard.com/primer/cables/cat5/ )

 

[ellisr63]

Because of what I do for a living I simply couldn't help myself... :T




Something doesn't sound right; unless the wires were flipped at two junctions - so they essentially reversed the reversal - then I can't see it having ever worked. Swapped leads shouldn't have limited speed, it should have eliminated communication entirely.




Cat3 came out in the early to mid 90's, so if the remodel was done in the 1980's it predates that by several years. It was also 10BASE-T, so 10 meg speeds. If you were getting 100 meg before that implies 100BASE-T, so cat5 at minimum.

Ethernet topology is point to point, meaning there's a single cable run from the switch port to each termination (generally that's a wall plate with a female RJ45). Since they're independent of each other, and therefore technically isolated, what you have plugged into any one port should not have an affect on the other ports. You can try isolating the slower devices but I can't imagine it would make any difference.

Are you using those GBIC's by chance? If so, I wonder if there's a problem during the signal conversion.

Thanks for the response... I am glad I have people that know more than I do to help me. I am doing what I do by what appears to be logical steps, but def is not working out.

This is what I did to test the jacks. I plugged in a known good store bought cat6 cable to the jack, and inserted the remote piece for the Klien tool on it. I then went to the Router room, and removed the wire from the next link in the connection (which was a switch) and put the Klien meter on it. Is this the correct way to test out the inwall connections?

GBICs... Those are the metal addon connectors for fiber aren't they, if so I am not running them.

We wired all the jacks to the "B" standard.

 

[ellisr63]

Just a point about the cabling being wrong. You can have straight through cabling, or swap-over, as most equipment auto-detects it.
Maybe it had been wired as swap-over for some reason? ( http://www.wiringwizard.com/primer/cables/cat5/ )

Thanks for the info...both ends of the jacks should be wired identically now as the tester says they are correct. I am assuming this is a good assumption from reading the meter, but I am not an expert.

 

[theJman]

Thanks for the response... I am glad I have people that know more than I do to help me.

By virtue of the fact you even know what jumbo frames are means your knowledge of networking exceeds perhaps 99% of non-IT people. Given the circumstances, that may prove very helpful.

This is what I did to test the jacks. I plugged in a known good store bought cat6 cable to the jack, and inserted the remote piece for the Klien tool on it. I then went to the Router room, and removed the wire from the next link in the connection (which was a switch) and put the Klien meter on it. Is this the correct way to test out the inwall connections?

I'm familiar with Klein Tools - and even own one of their crimping tools myself - but I don't have any experience with that particular tester. However, the steps you outlined are the standard method for tone testing so I imagine what you did was correct.

GBICs... Those are the metal addon connectors for fiber aren't they, if so I am not running them.

Not many home-owner class routers come with GBIC's, so when I saw that mentioned I wondered if your cable company might not be FIOS instead. I don't have any direct exposure to their service yet, but from what I understand they're using true fiber optics (the FIO in the FIOS name). If so, a GBIC would be necessary to do the conversion. Since they aren't in use though we can scratch that one off the list.

 

[ellisr63]

I just got off the phone with my Cable Provider, and they will be coming out on Wednesday to figure out what is going on. While i was on the phone with the Tech, my speed varied from less than 1Mbps to 256Mbps for downloading, but maintained a 10Mbps for upload everytime I ran the test. The Tech also had never heard of such a fluctuation is signal. I also decided it was a good time to sign up for the $4 a month wiring coverage. Now it doesn't even matter if my wiring is the problem...they will fix it for free. As long as the Tech doesn't mind going under our house in the crawlspace he would only have 3 runs to replace (if he wanted to replace wires...2 15' runs, and one 20'), and that would cover the switches to the router, and the router to the modem.

I will post up the results after they fix it. Keeping my fingers crossed it is something external to the house.

 

[robbo266317]

I am certainly interested in hearing what the cause is. :scratch:

 

[Andre]

I bought one of these to test if I had any bottlenecks.

https://www.amazon.com/Tonor-RJ45-Network-Cable-Tester/dp/B00OUFX38W/ref=sr_1_1?ie=UTF8&qid=1476738399&sr=8-1&keywords=cat+5+tester

 

[fschris]

i am a CCNP so i will throw out a couple things:

1. Have you tested a PC or laptop connected right to the cable modem for your desired speeds that you are looking for ?

2. that is all the cable company will care about.

Now, your network is a hodpodge of devices and those switches are not that great. You could have a STP issue, vlan issue, default gateway issues etc... even the crappiest cat 5 in your house from 1985 could probably do 100mb so i doubt you have a major wiring issue. But you never know...

Also don't forget if you test your internet speeds on dslreports (etc) you will never get those speeds in real life internet browsing. Netflix will give you about 5Mb downloads and ultra HD is around 25. I mean even trying to upload pictures to the cloud will be limited.

I highly suggest using Ubiquiti networks gear and use a Ubiquiti WAP. All the netgear stuff is kind of junk. Its kind of like amplifies sayiung they can handle 5000 watts.
https://help.netflix.com/en/node/306

 

[ellisr63]

i am a CCNP so i will throw out a couple things:

1. Have you tested a PC or laptop connected right to the cable modem for your desired speeds that you are looking for ?

2. that is all the cable company will care about.

Now, your network is a hodpodge of devices and those switches are not that great. You could have a STP issue, vlan issue, default gateway issues etc... even the crappiest cat 5 in your house from 1985 could probably do 100mb so i doubt you have a major wiring issue. But you never know...

Also don't forget if you test your internet speeds on dslreports (etc) you will never get those speeds in real life internet browsing. Netflix will give you about 5Mb downloads and ultra HD is around 25. I mean even trying to upload pictures to the cloud will be limited.

I highly suggest using Ubiquiti networks gear and use a Ubiquiti WAP. All the netgear stuff is kind of junk. Its kind of like amplifies sayiung they can handle 5000 watts.
https://help.netflix.com/en/node/306

Thanks for the info. Yes, I have tested it at the modem...when the Tech was here he did that, and got my speed up (wired and wifi, but soon after he left it dropped down. In addition now my Modem gets a latency error if I try to hook up the modem directly to the laptop. I have tried several cables, and the only way to have internet working now is with the modem plugged into the Netgear router. My Provider guarantees at least 200Mbps from the modem, and I tested it from their site with the Tech on the phone, and it varied from .3Mps to 239Mps (this was done with multiple tests from my HT with a Cisco switch).

I am not going to be buying new network switches or routers as it worked fine before we upped it to 250Mps from 100Mps.

 

[fschris]

have you ever transferred files around your network faster than 100m? i have issues doing this since my PCs and servers cant really do it. I would just not get too excited about having 250Mb download since the fastest you can download netflix is about 25 MB

 

[ellisr63]

have you ever transferred files around your network faster than 100m? i have issues doing this since my PCs and servers cant really do it. I would just not get too excited about having 250Mb download since the fastest you can download netflix is about 25 MB

I have had speeds of up to almost 300Mbps across my LAN in the past.

 

[ellisr63]

We had 2 Techs show up today to fix the problem. They found that the coax barrels were bad, and replaced them. Then they tested with their laptop to our modem, and got 250+Mbps. When they tested it with my wife's laptop the speed dropped down to less than 50Mbps. They tried testing bypassing the wires to the router, and disconnecting all 3 switches...signal came back to 250+ Mbps. Then they plugged in the switches again (testing after each one is up), and the speed was still 250+. I then asked them to run the test again just to make sure, and it dropped down again. They said that it is either my wife's laptop, our switch that is next to the router or that the wires need to be replaced in the wall. I will return the switch, and see if that solves the problem...if not then I guess it will be time to run new cat 6 to the switches to replace the current cat5 wires.

 

[ellisr63]

My friend is going to come over this weekend and see if we have a noise interference problem...since the Tech was able to get full speed from their laptop, but not from my wife's laptop a couple of feet away. Are there settings on a Win8 or Win10 laptop that could cause problems? I have noticed that her laptop doesn't even see our 5Ghz wifi, but does see the lower one. The laptop is only about 2 years old.

 

[fschris]

Is the laptop 802.11a or 802.11g ?

http://www.radio-electronics.com/info/wireless/wi-fi/ieee-802-11-standards-tutorial.php

 

[theJman]

My friend is going to come over this weekend and see if we have a noise interference problem...since the Tech was able to get full speed from their laptop, but not from my wife's laptop a couple of feet away. Are there settings on a Win8 or Win10 laptop that could cause problems? I have noticed that her laptop doesn't even see our 5Ghz wifi, but does see the lower one. The laptop is only about 2 years old.

A setting should either work or not work; intermittent issues suggest something else to me. Variations - especially wide ones like this - are often associated to a mismatch of some type. Continuing on with my driver/firmware theme... is your wife's laptop in need of an update perhaps?

 

[ellisr63]

A setting should either work or not work; intermittent issues suggest something else to me. Variations - especially wide ones like this - are often associated to a mismatch of some type. Continuing on with my driver/firmware theme... is your wife's laptop in need of an update perhaps?

I will have to check and see if she has all the latest updates.