What's new

I've failed to bridge a 5Ghz signal with any good thruput... I need help.

  • SNBForums Code of Conduct

    SNBForums is a community for everyone, no matter what their level of experience.

    Please be tolerant and patient of others, especially newcomers. We are all here to share and learn!

    The rules are simple: Be patient, be nice, be helpful or be gone!

SledKnight

New Around Here
Looking to see if anyone can give me guidance on my Wireless thruput issues (5Ghz bridge is REALLY slow ~ 100Mbps):


WD DX4000 NAS --> (Gigabit ports, CAT6) --> ASUS RT-N66U Router --> (wireless, 5GH, 30ft through two sheetrock walls, 2nd to 1st floor) --> TrendNet TEW-680MB Bridge --> (Gigabit ports, CAT6) --> Win 7 64bit PC


All devices on latest, non-beta firmware (3.0.0.3.112, 1.0.1.0, etc)

DX4000 transfer specs are 85 MB/s
PC is also capable of those speeds


If I run a CAT5e from RT-N66U to PC directly, I can transfer a 175 MB file at ~55 MB/s (per windows file copy).
When I connect them via the 5Ghz bridge, I can only transfer at ~8 MB/s (per windows file copy). [same speeds I was getting on the old WRT300N and 2.4 Ghz USB Adapter]


I used inSSIDer to determine I am the only 5GHz signal in my area.
I used TutuSoft LAN Speed Test 2 and LAN Speed Server to eliminate HDD speeds as issues, and when I connect Router to PC via CAT5e, LAN Speed Test shows 65 MB/s rates, but via 5Ghz bridge I only get around 9 MB/s. I disabled QoS and got that up to 12 MB/s. I moved the bridge 3 feet from the router and ran CAT5e from bridge to the PC (to eliminate signal strength and interference) and got it up 14 MB/s. Even at that, file transfer only shows 12 MB/s. Speeds seem as I expect, except on the bridged signal.


What I think I know:
1) DX4000 and PC are both capable of transfering files, when hardwired, at ~55 MB/s.
2) Network, when wired, can transfer ~65 MB/s.
3) 5Ghz wireless speeds are dropping to around 12 MB/s (even when it only hops 3 feet to the bridge)
[all devices in the house are set to use the 2.4 Ghz SSID, and ONLY the bridge is set to use the 5Ghz SSID]
[WPA2 is enabled, but I removed that for tests and saw no improvement]

What I expect:
With the Router and Bridge each capable of 450Mbps (~55+ MB/s) I think 30-45 MB/s is a reasonable speed to shoot for.

What I get:
No matter what I do (channel bandwidth 20, 20/40, 40 Mhz, auto/manual channel selection, QoS on and off, moving bridge closer, etc.) I can't seem to get anywhere NEAR that speed.

Any ideas would be greatly appreciated.

The purpose of the 5Ghz Bridge was to keep me from running wire down to living room (support wall, have to go a roundabout way, etc) while still maintaining ~40 MB/s access to network to give a speed boost to the 4 Ethernet devices in the living room that were connecting via 2.4 to WRT300N (or not connected at all).

So my questions are:
1) Shouldn't I be able to hit 30-45 MB/s with 5Ghs 450Mbps bridge at this distance?
2) Is there a better bridge than what I've picked?
3) Would a second RT-N66U work as a Bridge (there is a "bridge" setting in the menu, but it requires a MAC address and appears to only join to an open AP - So not sure that will work without DD-WRT or Tomato)?

NOTE: I decided to do another test with an open network, but after I changed the Router, the Bridge lost connection (as expected), but I have been completely unable to get back into the TEW-680MB settings. It is completely bricked. I've done several resets and can not access the device at all... so... it's a good time for alternative solutions (but I prefer trying to bridge the 5Ghz SSID [that's why I picked the RT-N66U] or going hardline rather than using the power outlet type network options) :eek:

Anyway... I'm a little green at networking, but I've exhausted everything I can think to try at this point. Help is appreciated.

Thanks! :D
 
In the 5.8GHz band, do both the router and the bridge promise in their specifications to support 40MHz mode? I read the more than terse user's manual and specs for the ASUS and couldn't determine.

No doubt you know that the net yield is about 60% of the indicated WiFi link speed ("burst rate"), e.g., 11g with an ideal signal has a burst rate of 54Mbps = 6.75MBps. Your question about exceeding 30MBps (240Mbps) net would be more than I'd expect to get with WiFi, even in 40MHz mode, and with an ideal signal strength and no interference.
 
Last edited:
You're expectations are way too high for wireless, you're currently hitting about the max for 3 stream @ 40MHz. Highest speeds recorded for 3 stream @ 40 MHz by SNB is in the 75-105Mbps range or just 9-13MB/s.
 
Last edited:
Your lucky to get what you can out of the 802.11n @ 2.4GHz and @ 5GHz. The next standard promises a lot more but with the way the WiFi is you really can't hope for so much. Most everyone operates in mix mode. Forcing 802.11n @ 40MHz always seems to go back to Auto/20/40MHz. These manufacturers seem to test at only Auto and Mix mode. The other options are requested but never recommend. Also most like a fallback method. What you get in MBPS even the highest you get is better than nothing lower or higher.
 
So... how exactly can they advertise 450Mbps when pulling less than 1/4 of that is considered "normal"? That seems like false advertising... I mean, what's the point of routers/bridges that are "300Mbps" and "450Mbps" when all they can actually do 100Mbps?

I knew I wouldn't be able to expect a full 450Mbps, but geez, I'm getting the exact same speeds as I was on the old non-N 2.4 Ghz setup from 6 years ago.

It seems my best option is to run CAT6 down and get a gigabit switch for the living room. I'll send this TEW-680MB back, as it has become completely unresponsive anyway. I'll keep the RT-N66U, because it looks nice, performs well, and I needed the Gigabit port.

Very frustrating, however. Seems like marketing has gone over the top in the 6 years since I last upgraded, but technology hasn't really advanced much...

Oh well...

Thanks for your help.
 
The data rate is the PHY layer and it is a fair number. There is MAC layer overhead, and by the time you get up to the application layer, 35 MBps is about what you can see... take away protocol overhead for SMB, and on a dual stream wifi connection, 11 MBps is a reasonable number...

also consider that WiFi is half duplex... this is largely why Wired will be faster, even at 100BaseT :cool:
 
300mbps = 150mbps

TX = 150mbps
RX = 300mbps

Most you get 65mbps to 270mbps and that depends a number of factors.

450mbps = 225mbps

TX = 225mbps
RX = 450 mbps

I've noticed about 1% spike up on the 450/450mbps
 
Since the beginning of WiFi, the non-geek consumer is duped by air link bit rate versus net yield bit rate. Like 60%.

Cars: engine horsepower vs. road horsepower, where the latter suffers the losses in the drive train and friction.

But these losses are much less than the WiFi air link bit rate vs. net, mostly because WiFi is a half-duplex medium. Transmit a burst, wait for a received-OK acknowledgement, transmit again. And oh yes, before each transmission, delay and ensure no other system is already transmitting on the shared medium.
 
Howdy,
First, it sounds like your confusing MB/s and Mbps.

Second, products from different manufacturers always lose something in the connectivity.

Third, a more compatible bridge will be the Asus EA-N66, it should be showing up for sale soon.:cool:
 
Looking to see if anyone can give me guidance on my Wireless thruput issues (5Ghz bridge is REALLY slow ~ 100Mbps):

100mbps at the application layer is "REALLY" slow? In what world? I think that's excellent, and you're seeing great results!

Sorry to have to say it that way, but as has been said a couple times, 450mbps is a theoretical PHY layer speed under perfect radio conditions. You'll never achieve that, and even if you do, due to overhead, you'll see a fraction of it at the application layer.
 
Realistically sustaining 100Mbps at the application layer is not only excellent, it's hard to do with consumer PCs and Windows file overhead- for realistic use cases rather than disk-less or big-file disk transfers.

For entertainment, take two fast PCs and do two concurrent transfers. You'll see that for two, the sum is greater than for one, due to windows overhead. This continues to a break point. I used disk-less transfers and I recall that the break point was about 4 or 5 concurrent. This will vary by PC speed and goodness of switches (assuming all transfers are inter-LAN).
 
Last edited:

Similar threads

Support SNBForums w/ Amazon

If you'd like to support SNBForums, just use this link and buy anything on Amazon. Thanks!

Sign Up For SNBForums Daily Digest

Get an update of what's new every day delivered to your mailbox. Sign up here!
Top