Question Router Throttling my speed

[esarhaddon]

My Netgear 6400 router is throttling my ethernet speeds. I can get 900+ Mbps connected directly to the modem, but going through the router it is throttled down to just over 100 Mbps. THIS IS ETHERNET ONLY! I thought routers were pretty much passthrough. I don't think my old ASUS had this problem. Have I set it up wrong or is oit just ready for the trash

 

[kanewolf]

My Netgear 6400 router is throttling my ethernet speeds. I can get 900+ Mbps connected directly to the modem, but going through the router it is throttled down to just over 100 Mbps. THIS IS ETHERNET ONLY! I thought routers were pretty much passthrough. I don't think my old ASUS had this problem. Have I set it up wrong or is oit just ready for the trash

Start with a factory reset of the router. Configure only the admin password, SSIDs and WIFI password. Retest.
Is the speed above or below 90Mbit?

 

[lantis3]

Did you check WAN port LED status? Was it showing it's running at 1Gbps speed? Have you swap another cable?

What's your subscription speed? and https://www.speedtest.net test result ID?

 

[bill001g]

A router actually is doing quite a bit. Every packet that passes through in both directions it must replace the internal IP and port number and the recalculate the checksums in the header. When packets come back it must reverse this as well as keep all the traffic separate for all the different machines in your house.

Routers have tiny cpu so it is easy to overload them doing this..most routers get about 300mbps. The router manufactures though have a trick where they moved this function to dedicated asic in the silicon. It can now pass traffic even on routers that have 5gbit internet connection.

The reason it was recommended you factory reset the router is using some features in the router will disable this hardware nat accelerator and the cpu must do all the work.

 

[lantis3]

Hmm, fellow forum member R6400 speed

Question - 1gb connection, only getting 300mb when using my own router netgear R6400v2

Using firmware 1.0.4.102_10.0.74 I have a 1gbps internet connection. When i run a speed test I'll usually get around 500-600mbps down and 800mbps up. However, other then the speeds i get from speedtest.net, everything else seems to max out at around 300mbps down. I don't think i've ever...

forums.anandtech.com

 

[esarhaddon]

This has nothing to do with cables. I have all the latest updates and it has been rebooted a million times. I did get about a 25 Mb boost when I turned off the Child protections but that still leaves 750 Mbps unaccounted for.

 

[lantis3]

This has nothing to do with cables. I have all the latest updates and it has been rebooted a million times. I did get about a 25 Mb boost when I turned off the Child protections but that still leaves 750 Mbps unaccounted for.

You still did not answer what's your subscription download/upload speed. Showing speed test ID also will not reveal your identity.

 

[esarhaddon]

A router actually is doing quite a bit. Every packet that passes through in both directions it must replace the internal IP and port number and the recalculate the checksums in the header. When packets come back it must reverse this as well as keep all the traffic separate for all the different machines in your house.

Routers have tiny cpu so it is easy to overload them doing this..most routers get about 300mbps. The router manufactures though have a trick where they moved this function to dedicated asic in the silicon. It can now pass traffic even on routers that have 5gbit internet connection.

The reason it was recommended you factory reset the router is using some features in the router will disable this hardware nat accelerator and the cpu must do all the work.

 

[esarhaddon]

You still did not answer what's your subscription download/upload speed. Showing speed test ID also will not reveal your identity.

My sub speed is 800 Mbps and my UP line is the same both ways at 22 to 25 Mbps no problem with that. I like that response suggesting a factory reset. I was not aware of the NAT problem. I think you all this is the fastest response I have ever had.

 

[kanewolf]

This has nothing to do with cables. I have all the latest updates and it has been rebooted a million times. I did get about a 25 Mb boost when I turned off the Child protections but that still leaves 750 Mbps unaccounted for.

You say you turned off child protections. Did you actually do the factory reset ?

 

[bill001g]

The hardware nat feature allows the traffic to bypass the cpu part of the chip. Using any feature even simple lists of what sites are being accessed need the cpu and to be able to see the data. Pretty much you can't use a lot of the fancy features on a router when you have a fast internet connection.

 

[lantis3]

Do a iperf3 test between LAN devices first to make sure at least you have no problem on the LAN side

View: https://www.youtube.com/watch?v=3JQg76nqmps

By the way, do you run any sort of VPN?

 

[esarhaddon]

RESOLVED. After I did a factory reset it turned into a boat anchor. I did all the recovery suggestions online to no avail. Now I have a new TP-Link router and will never buy another Netgear as they have NO SUPPORT.
Also for the person who said that flat cable is smaller. I haven't measured yours but I did measure the wire in a spare of mine and with a calibrated micrometer they were within a .002 of a round cable. Just because they are laid out flat is no call to bad-mouth products. i am sure that there are cheap junk ones out there but I know of plenty of Junk ROUND cables also. I expect better from this forum. Let's keep it civil people.

 

[Grobe]

I haven't measured yours but I did measure the wire in a spare of mine and with a calibrated micrometer they were within a .002 of a round cable.

How did you measure and with what device ?

What are that "0.002" number supposed to mean? Ohms?

 

[esarhaddon]

A MICROMETER! and .002 in is 2/1000 of an inch. I believe I stated that already in my statement. "calibrated micrometer"
I mean if you look at the text was referring to a person who stated that flat CAT wire is substantially smaller than round CAT wire. So I don't understand you switching to electrical measurements OHMS, which haven't been brought up, NOT LOGICAL.

'Calibrated' is a system of proving or verifying and if needed adjusting some measuring device to a known standard. Usually a National Standard.

'Micrometer' See link for a pict.
https://www.globalindustrial.com/p/fowler-1-inch-digital-counter-micrometer?adlclid=acf44755c9f414053b0b229ba5fa8af6&msclkid=acf44755c9f414053b0b229ba5fa8af6&utm_source=bing&utm_medium=cpc&utm_campaign=[ADL] [DSA] Tools & Instruments {High}&utm_term=Tools & Instruments Measurement and Layout Tools B&utm_content=Tools & Instruments - Measurement & Layout Tools - B
Micro = extremely small
Meter = a unit of measurement or a measuring device.

If you want that in MM (milimeters) I can provide that for you also. .0508 mm

 

[esarhaddon]

A sample piece of Bleden CAT5e measures .0215 in.

If flat CAT wire was significantly smaller they would be forced to manufacture custom RJ45 plugs as the slot on those has to be very accurate to cut the shield and not cut the wire. I can't see a manufacture cutting costs on the size of a wire and then having to increase the cost in a major way by having to source custom RJ45s

 

[bill001g]

I am not sure where someone bad-mouthed your cables in this thread but now you temped me to respond.

When you measured your cables with your micrometer did you first strip off the insulation.

Wires is not really measured in fractional inches it uses a standard called AWG...which I guess you can convert if you really want to.

Key here what it REQUIRED by standards to be meet to be called a ethernet cable.

The standard says the wire size must be 22-24 AWG. In addition it has very specific requirements on how the pairs are twisted together and how these pairs are then twisted with the other pairs.

Almost all flat cable uses 33 awg wire and does not properly twist the pairs. So it really doesn't matter if it might work for someone or not it is not a ethernet cable. You can call it junk or fake or whatever you choose but you can not call it a ethernet cable unless it meets the standards.

The reason there is a standard is you know that it will work with equipment that was designed to function with those cables. If you make up your own wire type that just happens to plug into a ethernet port there is no guarantee it will work.

 

[lantis3]

Comparing flat and round ethernet cable size or width doesn't make any sense at all because twisted threads make much less noise.

There are requirements how many twists must have for cat5 cat5e cat6 cat6a, etc. per inch for a cable.

Flat ethernet cable has never been standardized, It works just because it's short length. All other standard require you can keep the specified speed at 100 meters.



Applicable to ethernet cable too.

View: https://www.youtube.com/watch?v=P7WfY9P2uNY

 

[esarhaddon]

I am not sure where someone bad-mouthed your cables in this thread but now you temped me to respond.

When you measured your cables with your micrometer did you first strip off the insulation.

Wires is not really measured in fractional inches it uses a standard called AWG...which I guess you can convert if you really want to.

Key here what it REQUIRED by standards to be meet to be called a ethernet cable.

The standard says the wire size must be 22-24 AWG. In addition it has very specific requirements on how the pairs are twisted together and how these pairs are then twisted with the other pairs.

Almost all flat cable uses 33 awg wire and does not properly twist the pairs. So it really doesn't matter if it might work for someone or not it is not a ethernet cable. You can call it junk or fake or whatever you choose but you can not call it a ethernet cable unless it meets the standards.

The reason there is a standard is you know that it will work with equipment that was designed to function with those cables. If you make up your own wire type that just happens to plug into a ethernet port there is no guarantee it will work.

AWG is not an actual measurement but a gaging standard. I know that sounds the same but there is a difference. and I am a Quality auditor and I do believe that I am smart enough to neasure the bear wire. I also studied Electronic Engineering at the USC, was an Avionics tech in the Navy, and also worked at Bell Labs.
The flat cable I cut open was nicely twisted and as stated was less than .002 in of round cable. even more important is the shielding, and all of this combined to meet standards for CAT cable it MUST fall within their specified measurements. Possibly I just have good stuff lying around but if I can just grab a piece of wire from under my bed and come up with acceptible measurements, I believe Most reputable manufacturers would be able to prove their product is within parameters specified by standards and code. If you have something that is 33 AWG and is being labeled as a category that requires 22 AWG you might have grounds for a lawsuit.
Next my measurement of .0215 in is good with your AWG standards and your claim of 33AWG is hard to believe as that is only .007 in. Not even realistic in the real world. see the chard on AWG here:https://www.rapidtables.com/calc/wire/wire-gauge-chart.html

 

[esarhaddon]

Cat6 Cable Specification | R&M Electrical Group

Cat6 Cable Specification Chart - Specification for Hyper-Data 2000 CAT6 UTP Stranded LSZH Lan Cable

www.rm-electrical.com

 

[esarhaddon]

Comparing flat and round ethernet cable size or width doesn't make any sense at all because twisted threads make much less noise.

There are requirements how many twists must have for cat5 cat5e cat6 cat6a, etc. per inch for a cable.

Flat ethernet cable has never been standardized, It works just because it's short length. All other standard require you can keep the specified speed at 100 meters.



Applicable to ethernet cable too.

View: https://www.youtube.com/watch?v=P7WfY9P2uNY

Nice data but you completely leave out anything about flat CAT cable other than your remarks and NO intrinsic data. My flat cable has twisted pairs. SHOW ME THE FACTS!

 

[bill001g]

Unfortunately the official standard are locked behind paywalls. I used to have access where I worked but I am not going to pay that big fee just to prove forum posters wrong....couldn't even post copies of the documents if I had access. They are also massively complex with way more math that I am willing to read when referring to the things like twist radius.

You have to be very careful about what you find about cables some of it is outright wrong and posted to make cables sales.

One of the documents you post talks about stranded cable. That too has a requirement as to how many strands. I don't remember but it clearly says the wire size is 24awg even though it made up on multiple smaller stands.

It is not just the 2 wires them selves that must be twisted there is a requirement for how they are oriented next to each other.

There is massive amounts of this fake cable being sold everywhere espeically on amazon. Even amazon themselves sell cable and they clearly state it has AWG 30 which automatically invalids it as ethernet cable.

https://www.amazon.com/Amazon-Basics-Ethernet-250MHz-Snagless/dp/B089MD7FZM

This is the garbage 99% of the people posting here are buying.

Maybe you have something different. Would be interesting if the vendor actually publishes a fluke meter certification test of a 100 meter run.

 

[lantis3]

While I was still studying in the college, we have to learn what's Xerox CSMA/CD ethernet ,IBM's token ring and cable types and calculate how much time it will take from one end of cable to the other end with different protocols.

When I got my first job, the company was still using coaxial 10Base2 with 20 PC. Next year we have 100 PC completely running on 100Mbps ethernet.

 

[esarhaddon]

Unfortunately the official standard are locked behind paywalls. I used to have access where I worked but I am not going to pay that big fee just to prove forum posters wrong....couldn't even post copies of the documents if I had access. They are also massively complex with way more math that I am willing to read when referring to the things like twist radius.

You have to be very careful about what you find about cables some of it is outright wrong and posted to make cables sales.

One of the documents you post talks about stranded cable. That too has a requirement as to how many strands. I don't remember but it clearly says the wire size is 24awg even though it made up on multiple smaller stands.

It is not just the 2 wires them selves that must be twisted there is a requirement for how they are oriented next to each other.

There is massive amounts of this fake cable being sold everywhere espeically on amazon. Even amazon themselves sell cable and they clearly state it has AWG 30 which automatically invalids it as ethernet cable.

https://www.amazon.com/Amazon-Basics-Ethernet-250MHz-Snagless/dp/B089MD7FZM

This is the garbage 99% of the people posting here are buying.

Maybe you have something different. Would be interesting if the vendor actually publishes a fluke meter certification test of a 100 meter run.

You touched on something that I found. You get what you pay for. And if you insist on purchasing from some Chinese CCP affiliate, you probably won't like the product you get. Now Cramazon is not inherently evil, but just allows almost anyone to sell there. You mentioned stranded cable! I believe this has a lot to do with the 33 AWG versions. Yes, the individual strands might be .007 in but in a stranded bundle, they make up a total that closer approximates 22 AWG. I don't know how far that goes, but I am sure it is true in some cases. BUT if a person had extremely high-quality OFC wire why would there be a significant loss in a smaller wire? These are all 'IF' scenarios.
Now I did find on Cramazon multiple ads for 33-34 AWG CAT wire, but it didn't take me 30 seconds to ALSO find multiple ads for flat CAT# cable much more in line with round cable standards. See sample here for 24 AGW CAT6 flat cables from Monoprice: https://www.amazon.com/Monoprice-Cat6-Ethernet-Patch-Cable/dp/B09HNF36Z4/ref=sr_1_3?adgrpid=1332608665032402&hvadid=83288255967271&hvbmt=bp&hvdev=c&hvlocphy=93546&hvnetw=o&hvqmt=p&hvtargid=kwd-83288400873195:loc-190&hydadcr=946_1014950857&keywords=ethernet+flat+cables&qid=1708032097&sr=8-3&th=1

BUT even at that what is the real story? Digging down in the description they admit 30AWG stranses 'UNSHIELDED' I would agree that this could not qualify as CAT 6. So it only shows to be a 'Buyer Be Ware' situation. ANY cable that is not twisted pairs would have serious degradation and crosstalk and unshielded would have serious iductive leakage even in say a 1 ft patch cable. My desktop is connected with a quality flat Cabel that has .022 single-twisted pairs and some level of shielding. I get flawless 900+ download speed with NO evidence of packet degradation. This is on a 25 ft cable.
Concerning my original comment, I just want to sow this up by saying that when people just start tossing out a bunch of unsupported derogatory remarks, it does no one any good. There might be a case for warning someone to be aware of some products and caution them to be diligent in their research before purchasing something but to just make a blanket statement impugning the integrity of a general design is as bad as those dealers that sell the junk products.
Also, uneducated people are talking that just make ludicrous statements like the guy who asked me if I was stripping off the shielding before I measured the wire. GO BACK TO KINDERGARTEN. I you have a serious question find out the person's credentials before you start slamming them. Then if you hear some data or statistics, find out if it is credible and if they have repetable results. During my quest to resolve my problem, I ran across a tier-two CS Agent who spent nearly an hour trying to convince me that WiFi was stronger, more reliable, and had more bandwidth than Ethernet cable. Yes, these people do exist and get jobs in places of authority. Notice I was able to mock this particular UNNAMED person Without naming the company she worked for, which would have defamed them. That is even with a company I literally hate.

 

[Ralston18]

And with that this thread is ended.