Different Ethernet cable speed to multiple PCs from same router

[ScottB3333]

Have Comcast 105 with their modem/gateway. Set up four gaming PCs for grandkids. Ran ethernet from router to switch and connected four pcs to switch. Ran 88-94 mbps most of the time. My PC in office, run off different switch (which goes to a few Rokus) gets same. Then I ran an ethernet line to basement, hooked up a PC and get 126 down there. Same router. Tried switching cables from router to wall connectors and could find no difference. Thought because basement PC was connected to router without a switch that that might be it. So I unplugged feed line from each switch and plugged it into a PC and still got 88-94. I'm trying to figure out what could cause such a boost in one PC.
Thanks,
S Betts

 

[Tom Tancredi]

Honestly the extra 20 your seeking won't make much difference (all depends how your measuring and what Windows versions these are and what hardware we are talking about) unless you play to have ALOT of GB transfers from PC to PC. Your biggest issue will be if you do NOT setup a QoS table in the Router then implement QoS on all these devices. QoS is Quality of Service (ask comcast for help on that IF it supports it at all) which determines say watching Netflix would NOT be interrupted by grandkid looking at Spongebob.com website, or the Wife catching up on her Sons of Anarchy on Roku only to have Grandkid playing BattleField make her video stttttuuutttterring...

FYI Wired connection are nice but won't do better then modern Wireless normally, as Wireless runs up to 300Mbps with 802.11n and I think the last package I saw was 700Mbps for 802.11ac? Though honestly 100Mbps is what most people 'look for' if they have a Cable ISP, otherwise on DSL or such won't matter as the Internet connection is usually between 1Mbps and 15Mbps, so it wouldn't matter how 'fast' the wires/wireless is still 'stuck' on the outside pipe which is shared with ALL the computers on a first COME first SERVICE basis without QoS.

802.11AC routers are same prices as old 802.11n (the older standard) and the USB (plug and play! really!!!) Adapters for ANY computers is dirt cheap and easy to do. You would have to check with ComCast to see IF thier device is providing wireless and what 802.11 standards it follows

I mention this as a notorious other issue you MAY be experiencing you can only check with a OHM meter and quality test each line end to end and even then that won't resolve if you pull a cable accidentally around a corner that may be causing the wires to fray/break a bit and cause transmission errors (resulting in retransmits, which degrades 'speed').

 

[ScottB3333]

Honestly the extra 20 your seeking won't make much difference (all depends how your measuring and what Windows versions these are and what hardware we are talking about) unless you play to have ALOT of GB transfers from PC to PC. Your biggest issue will be if you do NOT setup a QoS table in the Router then implement QoS on all these devices. QoS is Quality of Service (ask comcast for help on that IF it supports it at all) which determines say watching Netflix would NOT be interrupted by grandkid looking at Spongebob.com website, or the Wife catching up on her Sons of Anarchy on Roku only to have Grandkid playing BattleField make her video stttttuuutttterring...

FYI Wired connection are nice but won't do better then modern Wireless normally, as Wireless runs up to 300Mbps with 802.11n and I think the last package I saw was 700Mbps for 802.11ac? Though honestly 100Mbps is what most people 'look for' if they have a Cable ISP, otherwise on DSL or such won't matter as the Internet connection is usually between 1Mbps and 15Mbps, so it wouldn't matter how 'fast' the wires/wireless is still 'stuck' on the outside pipe which is shared with ALL the computers on a first COME first SERVICE basis without QoS.

802.11AC routers are same prices as old 802.11n (the older standard) and the USB (plug and play! really!!!) Adapters for ANY computers is dirt cheap and easy to do. You would have to check with ComCast to see IF thier device is providing wireless and what 802.11 standards it follows

I mention this as a notorious other issue you MAY be experiencing you can only check with a OHM meter and quality test each line end to end and even then that won't resolve if you pull a cable accidentally around a corner that may be causing the wires to fray/break a bit and cause transmission errors (resulting in retransmits, which degrades 'speed').

Thanks, Tom,
I figured the 20 wouldn't impact Minecraft or COD (games dujour here) but I'm retired and have time on my hands so this is something to investigate. I'd much rather do wireless but, when we moved here the wireless service from Comcast was just horrendus. Couldn't hold a signal, speed up and down, etc., so I wired the house and no longer have to reboot the modem every 40 minutes. I made the ethernet cables myself and wonder if even the slightest change in crimping pressure could affect the speed? Or maybe when I was running the cables through three walls and the attic a slight tight bend could hurt? Any idea how much distance matters? One switch is 65 feet from the router and gets the same speed as a switch 25 feet away. Do mobos matter? Version of Windows 7?

 

[Tom Tancredi]

...Couldn't hold a signal, speed up and down, etc., so I wired the house and no longer have to reboot the modem every 40 minutes. I made the ethernet cables myself and wonder if even the slightest change in crimping pressure could affect the speed?

Absolutely yes. If you don't buy premade ones you are HIGH risk for error, if you don't Ohm Meter each cable set, end to end with a very expensive meter to measure all aspects, inclusive to overresistance (which would of course cause signal loss). Don't forget you can only get 99Feet before you get too high signal loss between end points that the cable is not reliable. Lastly you should be using shielded, along with Cat 6 level cabling, Cat 5 and below can't support high speeds (they were made for the older 10Mbps lines). Normal wiring meets IEEE standards for Cat5e/6 1Gb wiring.

. Or maybe when I was running the cables through three walls and the attic a slight tight bend could hurt?

uhm YEAH... more then likely you have some crimps and again depending on the quality and CERTIFICATION of the wiring you could be really hosing things up.

.Any idea how much distance matters? One switch is 65 feet from the router and gets the same speed as a switch 25 feet away.

Yeah as I noted. You can either check IEEE.ORG or Wikipedia for Ethernet Cabling standards for the exact specs. But also you need to account for the type of switches your putting (didn't mention make or model). Common ones are cheap, since no one uses them, but as I said usually they are certified for 1Gb connections. Now the way you would TEST it speed would be NOT to a website from each PC, but instead set a PC as a 'server' to run speed test of uploading / downloading random large files from one PC to another. If all your tests were based on checking Speedtest.Org (external website) the speeds can vary depending on DEMAND, as your rated speed is MAXIMUM from Charter NOT GUARANTEED speed. Honestly your hitting the marks for your class of service, and especially how 'smart' the switches are to avoid latency between points.

. Do mobos matter? Version of Windows 7?

Well hardware wise ABSOLUTELY. If your mobo of one computer is say built with a old 10/100Mbs, as compared to a newer computer with a 1Gb, yeah that will matter again FROM PC TO PC ONLY, not Internet, perse. For Windows 7 on all them, no not version of Windows 7, versions of Windows itself (Vista as compared to XP 32 as compared to a W8 etc.) does matter, but also you should run Slim Drivers on each to make sure you have the latest drivers on each PC that no 'driver' issue maybe holding one back as compared to another. Again though as I said, if your just checking internet speed your on par considering it all and your service.