Poor Wi-Fi - access point to blame? channel? - Linksys WRT1900AC

[yetollerey]

I have a problem with poor Wi-Fi speeds. I'm testing it via Speedtest.net and i'm getting 3Mb (from closed range) instead of 30Mb. If I connect my Laptop directly to the main router, i'm getting the full 30Mb.

Of course i'm testing the Wi-Fi speeds using the wireless main router, not the access point.

Let me explain my setup:
Main router: Linksys WRT1900AC
Wireless Wi-Fi access point: TP-LINK TL-WR841N.
Both devices with latest official firmware.

When I defined the Wireless Wi-Fi access point in the TP-Link, i did it using the build-in Survey option. Both devices are using channel 1, so I thought maybe that's the reason for the slow Wi-Fi. If I try to change the channel in one of the devices to other then 1, the Wi-Fi access point does not work. This is just a thought, don't sure if that is really an issue.

Anyway, would appreciate any help from the friends here

 

[thiggins]

Multiple APs should always be set to different channels, using 1, 6 and 11 for 2.4 GHz. Otherwise the APs will share the same bandwidth, defeating the purpose of using multiple APs.

Then check that you are using 20 MHz bandwidth on the 2.4 GHz radio. In an area with many surrounding networks, using 40 MHz bandwidth doesn't buy you anything and can degrade performance for you and surrounding networks.

Shut off the AP and try your experiment again. Also try other channels and see if that helps.

What wireless card is in your laptop? It can affect the throughput you get.

 

[yetollerey]

I thought that there may be something with the channels, that's why I already wrote that I tried to change it, but if the router and AP do not share the same channel, there is no connection between them.
I repeated the test without the AP, same Wi-Fi results. I get the same speeds using my laptop (Intel Centrino Wireless-N 1000 with driver from 26 Jan 2014) and my iPhone 5s.

 

[yetollerey]

I just replaced the WRT1900AC with Netis Beacon router - full speed in same scenario, so the problem is with the WRT1900AC. What could it be?

 

[yetollerey]

I just found that I configured the TP-Link TL-WR841N to work as "wireless Access Point to Point to Point Bridge mode": http://www.tp-link.co.il/article/?faqid=176


You can see in the article that the AP and the router must share the same channel.
During configuration there is also a warning that the channels should be the same.

How to configure proper Wireless AP?
I was not able to find proper article in TP-LINK knowledge base: http://ticket.tp-link.com/index.php?/Knowledgebase/List

 

[stevech]

Commonly, an AP has a WiFi access point side and an ethernet cable side. The latter connects by wire to the router. Or instead of ethernet wired connection, it can use IP over power line or MoCA.

A product that is both an AP and a bridge, where the link to the router is WiFi, is unusual. If it is not dual-band it has to "backhaul" the WiFi access point traffic to the router via the same WiFi channel its using for access. This is essentially a WiFi AP and WiFi repeater. Works, but halves the throughput.

Cisco calls this a WiFi bridge with WiFi clients (AP part).

 

[sfx2000]

I have a problem with poor Wi-Fi speeds. I'm testing it via Speedtest.net and i'm getting 3Mb (from closed range) instead of 30Mb. If I connect my Laptop directly to the main router, i'm getting the full 30Mb.

Of course i'm testing the Wi-Fi speeds using the wireless main router, not the access point.

Let me explain my setup:
Main router: Linksys WRT1900AC
Wireless Wi-Fi access point: TP-LINK TL-WR841N.
Both devices with latest official firmware.

When I defined the Wireless Wi-Fi access point in the TP-Link, i did it using the build-in Survey option. Both devices are using channel 1, so I thought maybe that's the reason for the slow Wi-Fi. If I try to change the channel in one of the devices to other then 1, the Wi-Fi access point does not work. This is just a thought, don't sure if that is really an issue.

Anyway, would appreciate any help from the friends here

Use the WiFi in the WRT1900ac - simply put... it's actually quite good with some minor tweaks...

2.4GHz - use the B/G/N mode on narrow channels (20MHz) - using the automatic mode deploys 802.11ac non-standard use in that band...

5Ghz - just let the defaults run there... either share or not the same SSID's - depends on clients...

Turn off Guest Networking - most folks don't need it, but it defaults to Enabled on the WRT1900ac.

Unplug the TP-Link and see where things are...

sfx

 

[stevech]

Multiple APs should always be set to different channels, using 1, 6 and 11 for 2.4 GHz. Otherwise the APs will share the same bandwidth,

defeating the purpose

of using multiple APs.

Might I add: The usual purpose of adding an AP, in a consumer world, is to improve coverage. In some professional settings, APs are collocated but on different channels to increase capacity- IF the APs are controller based so the controller can distribute the capacity demand (load) of clients properly - by "pushing" clients to a selected AP.

For consumers, with moderate capacity demands, putting all APs on the same channel makes it a bit faster/easier for WiFi client devices to discover the beacons of the APs and hopefully decide to choose a better AP based on signal strength. And do this without channel-scanning.

To be sure, there are many ways to skin the proverbial cat.

 

[thiggins]

For consumers, with moderate capacity demands, putting all APs on the same channel makes it a bit faster/easier for WiFi client devices to discover the beacons of the APs and hopefully decide to choose a better AP based on signal strength. And do this without channel-scanning.

I would love to see data that confirms this behavior.

Consumers have a LOT of wireless devices, which tend to be 1x1. I have shown that lower class clients limit the available bandwidth of higher class routers and APs. So it doesn't take much to exhaust available bandwidth on a router when clients are streaming video.

The standard practice of setting APs to different channels is still the recommended way.

 

[chadster766]

I would love to see data that confirms this behavior.

Consumers have a LOT of wireless devices, which tend to be 1x1. I have shown that lower class clients limit the available bandwidth of higher class routers and APs. So it doesn't take much to exhaust available bandwidth on a router when clients are streaming video.

The standard practice of setting APs to different channels is still the recommended way.

This could work if the wireless cross over zone was small and had minimal SNR. If the AP's were close enough that the signal strengths were close at some point; IMO that zone would be jammed.

 

[thiggins]

This could work if the wireless cross over zone was small and had minimal SNR. If the AP's were close enough that the signal strengths were close at some point; IMO that zone would be jammed.

Sorry, I don't understand what "this" in your reply refers to.

 

[chadster766]

This could work if the wireless cross over zone was small and had minimal SNR. If the AP's were close enough that the signal strengths were close at some point; IMO that zone would be jammed.

The below quote in your previous post. Odd that it is missing in mine. I'm hitting the "Quote" button as usually but it seem to be behaving weirdly.

Originally Posted by stevech View Post
For consumers, with moderate capacity demands, putting all APs on the same channel makes it a bit faster/easier for WiFi client devices to discover the beacons of the APs and hopefully decide to choose a better AP based on signal strength. And do this without channel-scanning.

 

[sfx2000]

I would love to see data that confirms this behavior.

Consumers have a LOT of wireless devices, which tend to be 1x1. I have shown that lower class clients limit the available bandwidth of higher class routers and APs. So it doesn't take much to exhaust available bandwidth on a router when clients are streaming video.

The standard practice of setting APs to different channels is still the recommended way.

I tend to agree with Stevech - depends on load and conditions, but most modern chipsets in a dual-AP environment will coordinate, and I haven't seen any impact with let's say, single stream vs. three stream devices on a single channel, as long as they're of the same 802.11 level (e.g. 11n...).

11ac, at least with the Marvell and Broadcom chipsets, seems to be remarkable with a mix of 11a/n/ac clients - very little impact.

That being said, the "rule of 5" for 2.4Ghz still is a good route to go (e.g. 1/6/11 for channel assignments). Esp with more than 2 AP's...

sfx

 

[azazel1024]

I tend to agree with Stevech - depends on load and conditions, but most modern chipsets in a dual-AP environment will coordinate, and I haven't seen any impact with let's say, single stream vs. three stream devices on a single channel, as long as they're of the same 802.11 level (e.g. 11n...).

11ac, at least with the Marvell and Broadcom chipsets, seems to be remarkable with a mix of 11a/n/ac clients - very little impact.

That being said, the "rule of 5" for 2.4Ghz still is a good route to go (e.g. 1/6/11 for channel assignments). Esp with more than 2 AP's...

sfx

I think the biggest question is, what is the operating environment? If there are relatively few clients and/or concurrent high use clients, then setting all APs to the same channel MAY improve roaming slightly.

That said, having the APs spread across non-overlapping channels increases the overall maximum possible wireless bandwidth available.

The best I can possibly see (before overhead) on my router is 1750Mbps between 2.4 and 5GHz, my AP is 600Mbps. However, setting non-overlapping channels I can have a maximum possible of 2194Mbps (assuming 20MHz 2.4GHz on the AP and 40MHz 2.4GHz on the router). If I set it all on one channel, it produces much lower maximum possible bandwidth.

Really I am looking forward to both MU:MIMO, but also fast hand-off. I also wonder how well (probably pretty well) MU:MIMO works with co-channel access points and clients, considering how it directs the different streams, I'd think it would greatly reduce interference, especially if the clients/APs are physically spread-out well. Probably going to be pretty important once 160MHz channel bonding finally shows up...as that'll ensure there CANNOT be APs set on non-overlapping channels in 5GHz (without mandatory UNI-II support for all APs and clients, which is rare currently)