Need help with laptops connecting to WRT160N

[wally33]

I have a WRT160N and 3 laptops connected wirelessly and 1 desktop wired. Internet access is cable modem with 12+ Mbps connection speed.

Laptop #1 (vista, wireless N) connects at no more than 2 Mbps

Laptop #2 (vista, wireless G) conects at roughly whatever desktop is getting (10-12+ Mbps)

Laptop #3 (windows 7, wireless N) connects at no more than 2 Mbps

All laptops are within 5 feet of router.

I gather something is set differently on laptops #1 & #3, but I can't figure out what it is.

I have tried updating drivers on both laptops, with no improvement.

I'm only trying to get all 3 laptops operating at a roughly comparable speed (i.e., like #2).

Any suggestions?

 

[thiggins]

Make sure you are using WPA2/AES encryption and that WMM is enabled in the router.

 

[wally33]

WMM was already enabled, so no change there.
Changing to WPA2/AES (both the router and laptop adapters) seems to make things worse. Laptop #3 still only connects at less than 1Mbps, but now #2 connects at roughly the same (slow) speed.
Other thoughts? Thanks for your help.

 

[thiggins]

How are you getting the speed numbers that you are quoting?

 

[wally33]

It's an internet based speed test, using my ISP's nearest server. But, please understand I'm not so much interested in absolute transfer speed as I am in the relative speed between the 4 devices. I understand that introducing the internet into the equation brings a host of unknowable variables. But if laptop #2 has a speed that is close to the wired desktop connection, that is good enough for me. Likewise, I'm just trying to get laptops #1 and #3 reasonably close to laptop #2. Make sense?

 

[thiggins]

You are introducing a LOT of variables with your test method.

First thing I would do is to connect them all via Ethernet run the same test and see what your speeds are to rule out OS, disk, etc. issues.

Next, I'd need to know what all three laptops are reporting for their wireless link rate. That's the "Speed" value reported in the Wireless Network connection properties.

I assume you are running the test on each laptop one at a time, correct?

The encryption method used must match in the router and the clients.
Are all three laptops now set to use WPA2/AES? What were they using for wireless security before?

 

[wally33]

Via ethernet, all 3 laptops test at roughly the same speed as the desktop (this time of day, ~20 Mbps).

Yes, testing one machine at a time.

Link rate showing as 65.0 Mbps on #1 and # 3 (the slow ones). Link rate of 54.0 Mbps on #2.

All three (and Router) set to WPA2/AES. All were using WPA Personal before.

 

[thiggins]

I would try forcing the router to G mode only. The 65 Mbps link rate sounds like you have cheap single-stream N adapters in the two problem notebooks.
They could be having trouble with a mixed network.

If you can force the N wireless adapters to G only mode, I'd also do that.

 

[overdrive31]

Why does the only wireless G device test higher? My guess is you're using 2.4GHz @40MHz, too noisy. Reduce it down to 20MHz channel width(not auto 20/40) and do a site survey to find a free channel to use that other nearby networks are not using, or if you can't find a open one, to choose a channel that consists of networks with the least signal strength(farthest away). Forcing the clients to use wireless G is basically the same thing as reducing channel width to 20MHz using 2.4GHz N, but you do have the speed advantage using N. Link rate(65Mbps) would suggest 20MHz channel width and a single stream radio, but it could also be the best rate considering how crowded the spectrum is in your environment had they been trying to operate @40MHz(130Mbps). If not already, change the channel width to 20MHz only and if it makes no difference, try changing mixed mode output to wireless G only to force the N clients to connect at G.

 

[thiggins]

overdrive. Setting 20 MHz mode on an N radio is not the same thing as setting G only mode. G mode will limit the link rates to the B/G set only, which does not include 65 Mbps and will prevent use of N protocols, which is my theory of the cause of the problems.

 

[overdrive31]

Yes, it's not the same, but G is similar in that it also operates @20MHz(basically the same.. were my words) without the benefit of increased speed from N and it's enhanced protocols. I was only addressing the possibility of 40MHz use and a crowded spectrum being the cause. Many hate to find they purchased an "N" router only to find they cannot use it, but for just G. I was merely stating the option to try and fix a common problem with 2.4GHz N, before jumping to the forced G speeds. What if the N radios were MIMO? Is it still the mixed(N & G) nature causing link rates(link rates were on par with single stream, but not MIMO) and internet speedtests to slow? So many folks complaining about wireless N, I blame it solely on channel bonding in 2.4GHz, should never had been implemented for the radio band.

I have personally noticed some, but not all radios with mixed clients(2.4GHz G & N), cause even the MIMO N clients to have wireless B like reception throughput, yet also have MIMO N transmit performance. This could be what he is experiencing, would like to know why a mixed network would impact only the wireless N clients and their reception speeds. You might be right about falling back to G being the fix, but I wanted to express an option to retain wireless N in the interest of the wallet(why did I buy this if I can't use N?? GRRR!). Can remove the G client from the mix, and either upgrade it to N, or use an old router as an AP to serve it wireless G.

 

[wally33]

Tim, you get my vote for genius of the week.
Forcing the router to G mode fixed it.

So, do I understand that all 'N' adapters are not created equal?
Seems like I'd always be better off with a good 'G' adapter than a cheap 'N'?
For my future reference, how does a non-technical person tell the difference?

Thanks

 

[overdrive31]

Tim, you get my vote for genius of the week.
Forcing the router to G mode fixed it.

So, do I understand that all 'N' adapters are not created equal?
Seems like I'd always be better off with a good 'G' adapter than a cheap 'N'?
For my future reference, how does a non-technical person tell the difference?

Thanks

* no SGI/with SGI
(A) Up to 65/72Mbps = single stream @20MHz

(B) Up to 130/150Mbps = single stream @40MHz

(C) Up to 130/150Mbps = mimo 2 stream @20MHz

(D) Up to 270/300Mbps = mimo 2 stream @40Mhz

(E) Up to 195/225Mbps = mimo 3 stream @20MHz

(F) Up to 405/450Mbps = mimo 3 stream @40MHz

You will most likely see advertised as highest capable speeds using 40MHz, so the only numbers you notice on packaging are rates listed with SGI from B, D, and F. Also, if your router is set to transmit only N(not mixed) @20MHz only, just match your reported link rates with A,C, and E to determine which it is, but this would require the router to be as fast or faster to make the determination or else it becomes the limiting factor.

eg. Your WRT160N is MIMO 2 stream capable of either 130Mbps@20MHz or 270Mbps@40MHz using mimo 2 stream clients. Your link rates were only 65@20MHz or could have been 130Mbps@40MHz using single stream clients. Thus they must be single stream clients, unless having 40MHz enabled caused them to reduce their link rates to 65Mbps from interference.

 

[thiggins]

I don't hold single-stream N in very high regard, particularly on the router side of things.
Router Without A Reason: Belkin N150 Wireless Router Reviewed
Not-So-Brilliant Disguise: Linksys By Cisco WRT120N Wireless-N Home Router Reviewed

It looks like implementations on the client side aren't up to snuff, either.

Seems like I'd always be better off with a good 'G' adapter than a cheap 'N'?

Not necessarily. Properly implemented, single-stream N adapters should play nicely with G adapters. But it looks like your notebooks have adapters that are not good citizens.

You have to dig into your notebook specs to find the wireless adapter used. Then you need to do homework on it to see if it is full N or single-stream. Not a task for the "non-technical person".