Wireless N is VERY Dissapointing

Aznboy1993

2[H]4U
Joined
Nov 9, 2009
Messages
2,078
Using my recently bought hardware consisting of the D-Link DGL-4500 router and the D-Link DWA-160 usb adapter I did the following tests using LAN Speed Test version 1.1.0 build 97 (latest version at time of writing - 3/21/2010).

  • The filesize (in megabytes) that I used for the test was 100MB.
  • The directory of transfer was to another PC I have connected to the router using a wired gigabit connection.
  • The destination computer was idle with a WD Black 1TB 7200 PRM drive and a core 2 duo at 3.6ghz with 2gb of ram running Windows XP.
  • The source computer was idle with a SCSI 10k RPM drive and a xeon dual core at 3.2ghz with 1gb ram running Windows 7.
  • My G settings I was using were as follows: 2.4GHz, G only, Channel 11, Best (automatic) transfer speed, and WPA2 AES security.
  • My N settings I was using were as follows: 2.4GHz, N only, Channel 11, 40MHz, Best (automatic) transfer speed, and WPA2 AES security.

The results were very dissapointing... my first run was using G, the second using N.

Wireless G
wirelessgtransferspeeds.png


Wireless N
wirelessntransferspeeds.png


So according to my tests N is only (at most) twice as fast as G. IMAO it is very dissapoint considering the amount of money I spend on the router and adapter. I might be returning both in the very near future or trying another adapter (D-Link DWA-556) since maybe it's just the USB adapter that plain sucks. But if that fails then I'm giving up wireless altogether and am just going to stretch a 100ft. Cat6 cable to my desktop downstairs... SIGH :(
 
USB 2.0 is 480 Mbit/s, which is around 60MB/s. Taking into account the overhead of USB/IP/etc, 43MB/s is pretty good. What were you expecting, gigabit?
 
USB 2.0 is 480 Mbit/s, which is around 60MB/s. Taking into account the overhead of USB/IP/etc, 43MB/s is pretty good. What were you expecting, gigabit?

Looks to be 5MB/sec there chief.

And yes, N pretty much sucks. You get no additional range, only a little extra speed at moderate ranges. Get a wire run if at all possible.
 
Looks to be 5MB/sec there chief.

And yes, N pretty much sucks. You get no additional range, only a little extra speed at moderate ranges. Get a wire run if at all possible.

Oh you're right, well wireless is for Wiis, coffee shops, and waits at airports.
 
Based on experience, I'd firmly put the blame on the D-Link USB adapter ...

He's not getting 43MB/s, his test shows 43Mb/s, which is actually 5MB/s.

Realistic sustained USB 2.0 transfer rates are closer to 320Mb/s or 40MB/s. 480Mb/s is just the theoretical burst limit. That 480Mb/s maximum assumes quite a bit, including nothing else going on for that particular USB host. Your results fall way short of that, so it's not USB per-se that is limiting you.

When I upgraded my wife's PC to Windows 7 on the day it released it would not talk to the Linksys Wireless G card that she was previously using, and Linksys didn't (and I think still do not) offer drivers for it. So I had to find a locally available device that did have Windows 7 drivers right then and there.

The only such device I could find was a D-Link USB adapter. It was worse than useless.

The network adapter properties panel in Windows showed full signal and claimed a 130Mb/s link speed (driver lying it's arse off?), yet actual throughput was closer to 4Mb/s (about 500KB/s). That was significanly slower than the Wireless G card she had before and definitely pointless.

Since then I've ditched the D-Link POS and replaced with a Linksys Wireless N PCI card. That gets, in the same location, about 70 Mbp/s. Interesting is shows lower signal strength than the D-Link USB stick did, despite having proper antenna, AND the driver seems to show a link-speed that corresponds with what I am actually getting in terms of transfer rate.

Incidentally, the router in question there is an Apple Time Capsule (not the fastest router in the world by ANY means), and the wife's PC is the wrong side of a 4,000+ sqft house and on a different floor.

I can't comment on D-Link's other products, but the USB wireless N adapter (sorry, don't know the model number, don't see it on the site anymore - and I returned it for being useless) was utter crap.
 
  • My N settings I was using were as follows: 2.4GHz, N only, Channel 11, 40MHz, Best (automatic) transfer speed, and WPA2 AES security.

For me the only way to get decent speeds out of N was to run in 5GHz N only mode. Both your USB adapter and router are dual band, so you should be able to do that without any issue.
 
Have a powered USB hub handy? Plug the USB adapter's power cord into that, and see if that helps.
 
Ya I've pretty much narrowed it down to the stupid USB adapter. EFFIN gay man. I'm picking up a Linksys dual band PCI adapter as soon as I can get my hands on one :D

BTW, I've tried 5GHz and speeds seem to be better but the signal totally BLOWS (~50% with one wall in between that is about 1 ft. thick and like 20ft. of open air)... WTF?!
 
If performance is a concern you shouldn't be using a half-duplex shared medium that is subject to interference, i.e. don't use wireless.
 
Did a little configuring and rearranged the router and now I get rougly 50% signal strength using 5.8GHz and the speeds went up to around 100Mbps (~10MB/s).
 
If performance is a concern you shouldn't be using a half-duplex shared medium that is subject to interference, i.e. don't use wireless.

Amen. WiFi is suitable for non-data intensive laptops and nothing more.
 
Did a little configuring and rearranged the router and now I get rougly 50% signal strength using 5.8GHz and the speeds went up to around 100Mbps (~10MB/s).

You might get another 25 Mbps max out of that under optimal conditions due to the fact that actual data throughput on wireless connections tends to be less then half of the advertised rate due to network overhead.
 
Amen. WiFi is suitable for non-data intensive laptops and nothing more.
is so would but it isn't economical for me. my desktop is in the middle of nowhere x.x i would need a 100ft, cat6 cable and it would be UGLY :p
 
is so would but it isn't economical for me. my desktop is in the middle of nowhere x.x i would need a 100ft, cat6 cable and it would be UGLY :p

Nope. Try a HomePlug adapter.

I can vouch for the quality of the NetGear ones. My dad's had a 11Mbps one for the last 3 years, and hasn't had a single problem. And that house was wired for electric like 70 years ago!
 
Wow, don't hate on N, just go and get some real gear. I have tested a Cisco 1252 AP in the 2.4ghz range non channel bonded and got a consistent 10MB/s transferring large movies via SMB. That was while using a really crappy Lenovo ABG 1x1 NIC as well. The limitation is in your gear, not the standard. The standard is great IF you have the gear.

Looks to be 5MB/sec there chief.

And yes, N pretty much sucks. You get no additional range, only a little extra speed at moderate ranges. Get a wire run if at all possible.

Yes, you get no additional range since the power and frequency requirements remain the same. However, only a little extra speed at moderate ranges is an understatement. Where you'd get 1mbps in ABG, you'd get 7.2mbps in N. Over 7x the throughput is not "little" by any means. Heck, in the real world test linked below, one .11a station was getting 1.87mbps while the .11n station was getting 20.81mbps.

Here are some actual test results:

http://www.cisco.com/en/US/solution...394/ns348/ns767/white_paper_c11-492743_v1.pdf
 
Last edited:
Wow, don't hate on N, just go and get some real gear. I have tested a Cisco 1252 AP in the 2.4ghz range non channel bonded and got a consistent 10MB/s transferring large movies via SMB. That was while using a really crappy Lenovo ABG 1x1 NIC as well. The limitation is in your gear, not the standard. The standard is great IF you have the gear.

This. Channel bonding is stupid awesome, but the gear needs to support it. Is that Dlink stuff up to the N standard, or is it Pre-N? It's hard enough trying to get stuff to behave when it adheres to a standard, all bets are off using the 'close enough' standard trash.
 
You get range or speed, but not both. You either get better speeds at the same range, or similar speeds further away, basically.

Also much of the benefit does rely on fairly high-end MIMO gear. Last I read there were still some MIMO improvements to be made that are in the standard but no gear currently implements.
 
Antenna orientation too, I forgot about that one. Polarization FTL, You can get crazy losses with a complete horizontal/vertical mismatch.
 
Yes, you get no additional range since the power and frequency requirements remain the same. However, only a little extra speed at moderate ranges is an understatement. Where you'd get 1mbps in ABG, you'd get 7.2mbps in N.

And for any large file it'll be even faster if I walk to a location with a gigabit ethernet port and plug in, transfer the file and then return to the couch. I see wireless as a convenience, not as anything even remotely bearable for large file transfers.
 
it's ok now i guess but not as good as gigabite. my files transfers are comparable to those of 100Mbps networks. pretty good IMAO since it's wireless :D
 
Back
Top