Mesh Tenda MW5 setup question

[wozwebs]

Could do with some advice on the Tenda MW5’s.

My Now TV router was downstairs but we are decorating so I got the master socket moved to my office upstairs instead of the hallway as I work from home. Get a steady 71mbps in that room (I live close to local exchange and no Virgin in my area) but signal went dire in other rooms compared to what it was so I bought two of the dual packs of the MW5’s and they came today.

All linked up and I called it NOWTV2 on WiFi so on my phone, when I am linked to NOWTV2 my speed in other rooms is barely 25mps when I was expecting close to similar to what I get when in the same room as the router. Not much better than the router itself.

I haven’t touched any settings on the router though, so I still see NOWTV (router) and NOWTV2 (mesh) so do I need to disable something on the router so it all goes via the Mesh system and should speeds increase then?

 

[mickevh]

How is the link established between the MW5's an your router - wired or Wi-Fi..?

If wired, then you should just be able to disable the Wi-Fi on your router - it's rarely more than a couple of click to turn off the radios. If not, check the SSID to something different (that your clients don't know) and se a passphrase that your clients don't know so they fail if they try to connect. But that's rarely necessary as most SOHO kit will let you turn off the radios.

If the MW5 is linking to the router by WI-Fi, then by definition you cannot turn the routers Wi-Fi off as the the MW5's will have no way to talk to your router.

"Mesh" systems use exactly the same Wi-Fi protocols as everything else and are subject to the same (legal) limits on transmit power and so on - they are not blessed with some magic woo-woo that means their transmissions go further than anyone else's, no matter what the marketing BS says.

Thusly when using Wi-Fi backhaul links between nodes, (be they called "mesh" or anything else,) to avail the best performance we need to pay some attention to the physical positioning of the nodes to avail the most favourable signalling conditions between the pairs. For example, let's say you have a room getting -70dBm from your router. If you deploy a mesh node in the same place, it's getting no better signalling conditions than a client device so one would expect no better performance. For the best performance, we need to deploy our additional nodes somewhere "in range" of good signalling conditions of the (base) router/AP and the coverage hole we are trying to improve. If "R" is your router, "M" is a mesh node and "C" is the problem coverage area, it needs to be more like this R---M---C and less like this R-----M-C. You may need to experiment a bit to establish the optimum positioning.

In Wi-Fi "only one thing at a time to transmit." Which in includes the transmissions between the mesh nodes and your router. So each packet of data flying through the air between your router and a client talking to a mesh node now has to take two "hops" across the air waves - one R<->M and one M<->C. Both hops cannot occur at the same time, therefore for any given amount of data, it takes twice as long, or half the "speed." Even worse, the more nodes you have and the mode Wi-FI client devcies you have competing for the available "air time." Sorry, Wi-Fi is just like that.

Some "mesh" systems have a few trick to try and alleviate this, but they tend to be the more expensive ones. Even then, it's only a "better" solution. The "best" solution is to use cabled ethernet backhaul links between your nodes and your router if the equipment and domestic harmony permits.

FInally, be weary of you you conduct your performance test - for testing local links Internet "speed test" sites are not the best way to go about it. A better option, if you have a couple of computers locally, it to use (free) tools like NetIO or iPerf. It's kind of like running a "speed test" except you host the test server locally instead of out on the Internet thereby taking your Internet link and everything upstream of it out of the test pathway. It a much better way to "stress" test your local network. Failing that, you could try copying some largish files around locally, time it on a watch and do the maths.

 

[wozwebs]

Thanks for the reply. The primary node is connected to the Now TV router via Ethernet cable. Got up this morning and it can’t find internet connection on NOWTV2 (mesh system) but note there’s a setting to reboot daily at 3am for some reason and it did it yesterday too so think I need to disable that then I will switch WiFi off on the router itself and see. Trial and error I guess.

 

[mickevh]

If the Tenda offers the choice between operating in "router" mode or "AP (Access Point) mode" I would favour AP mode. In "router mode" you split your network in two separate sub-nets (which in the SOHO use case usually is of no benefit) and may have issues with multiple NAT's which can effect some applications. In AP mode, the Tends will "just" function as a fleet of managed AP's and you;ll have a single network instead of two.

If you cannot figure out how to turn off your NowTV routers Wi-Fi, maybe post up a screen grab of the router Wi-Fi settings page and someone will probably be able to point you to the correct bloxes to unclick.