[FAQ] Using two routers together/ Extending Wi-Fi

It's "Big Wi-Fi Myth Number 2" that client devices are always "hunting for the best signal." They do not and some clients need it to get pretty grotty before they initiate a roaming assessment.

With many clients, there's nothing you can do about their roaming behaviour and it's in the gift of the client device designer. (If/when to roam is not prescribed in standards.)

To stand the best chance of roaming, SSID names and passphrases need to be identical, ideally all AP's offering the same protocols (A/B/G/N/AC,) though that is not absolutely essential and I'd ensure the AP's are on differing radio channels to ensure there's less "near-far" masking occurring and/or mitigating the chance of AP's & client transmissions in different cells "interfering" with each other.

I prefer not to use "auto" tuning functions on small (unmanaged) deployments. In the 5GHz waveband, I just ensure the AP's are tuned to different channels. In the 2.4GHz waveband, the channels need to be different and at least "5 apart," for example, choose from the set [1,6,11] If you wanted to go the whole hog, you could assess any neighbouring Wi-Fi using tools like InSSIDer and try to avoid the neighbours if you have any.

"N" Wi-Fi in the 2.4GHz can also be problematic if one uses "fat" (40MHz) channels as there is insufficient frequency spectrum to deploy multiple calls with "fat" channels without them interfering with each other. In big deployments, it's not unknown to restrict N in 2.4GHz to "thin" (20MHz) channels, accepting that so doing effectively halves the maximum link rate (speed) but could yield a more stable infrastructure and more usable radio channels. You could have a play with that if your routers offer such choices.


Cheers mate, but have to say before i done the reset and that it was working ok and not like this
 
I installed a new BT Smarthub today so had a go at configuring the old HH5 as an AP.
Connected it to Mac by wired and wireless disabled on Mac.
Changed the DHCP Range by reducing it to .252.
Reboot and changed IP address to .253.
Reboot and disable DHCP function. Turn off HH5.
Reboot laptop.
Reboot HH5 and plug into LAN wired.
Its sending out an SSID but no internet access. Any ideas?
 
If you have literally just changed routers, it may be that your client devices are using old DHCP leases from your original router which (now) have incorrect IP addressing. This usually sorts itself out in 24 hours or so as the old leases time out an clients acquire new ones and the automation in DHCP does it's thing.

If you are beyond 24 hour, try some ping tests and report back the results:

Connect to both routers in turn and ping both router and something on the Internet (BBC usually answers, not all web sites do) and see what can be pinged from what.
 
After some help, I need to get WiFi in the garden, I currently have a plusnet router in an old house with thick walls, WiFi signal suffers towards the back of the house and is rubbish in the garden which is around 30-60 foot from router. What would be the best repeater to use? Home plugs are hit and miss due to the old electrics. Any advice appreciated plus I do t want to spend a fortune, I do have power in my shed as well.
 
There's no magic router/AP/repeater with "much better signal" than everyone else's so don't waste any effort looking for one. Wi-Fi is two way "conversation" like walkie-talkies, not a one way "lecture" like television: All the Wi-Fi clients (phones, tablets, laptops, etc.) are transmitters too and AP needs to hear" clients just as well as client need "hear" the AP.

Just like sound, if two communicating peers cannot "hear" each other then you need to get them closer together and/or remove any obstructions between them. For Wi-Fi, that almost always means installing additional AP's (Access Points) nearer where the clients are (I'll presume we're not going to knock down the walls of the house!)

You might try positioning an additional AP in a window overlooking the garden if you have one - glass attenuates Wi-Fi much less than masonry. Beware of any foliage, that's a Wi-Fi killer too.

You might try looking for an AP with separate antenna (sometimes styled as "external" AP's) so that the antenna can be mounted on the outside of a suitable external wall. Though of course, than means getting the drill out a knocking some holes through your wall.

Finally, you need to consider how best to establish the "backhaul" link between you outpost AP and the rest of your network. In order of preference (reliability and capacity) wired ethernet is best, HomePlug is probably next best and Wi-Fi backhaul (such as repeaters and "mesh" links) is do-able but probable least fast - though the latter is becoming an assertion that's less of a concern as Wi-Fi is getting faster and things like "tri-band" kit mitigate the bandwidth (speed) clobbering that Wi-Fi repeating causes.

Before spending any money, you can experiment a bit with AP physical positioning and see if it makes much difference. Using either you existing router (or an old one if you've got one lying around) try physically placing it where you are thinking of deploying an additional AP and go see how that effects the "bandit" area. Don't worry too much about the actual value of the signal strength/quality numbers - what we're looking for is trend, is it significantly better, worse, or the same. Hopefully that will give some idea of whether and "repeater in the back bedroom" so to speak, is going to make much difference.

If you use your incumbent router to conduct such as survey, it will of course loose Internet connection, but that's OK, it's Wi-Fi will almost certainly still work fine - you just won't be able to do anything online whilst testing/surveying.

Of course, a HomePlug/Wi-Fi combo in your shed would seem ideal, but if the mains isn't up to it then that's a non-starter, but if you can beg/borrow any Homeplugs to try it out, I'd give it go.
 
hi all,

thanks for this thread - very helpful and now i have seamless wifi over my 4 floors!!!

i have a netgear R7000 i am using as the 'secondary'. What is the recommended configuration for the R7000?

It can be configured as a Router, AP, Bridge or Repeater.

I am using it for WIFI and also the 4 ethernet ports.
 
I don't know that kit, but there are others here that do. However, if you've got a cabled backhaul between the outpost R7000 and the rest of your network, then I'd suggest "AP" mode is probably the most appropriate. "AP mode" is probably a "one click" solution that automatically disables the DHCP Server and (possible) sets it up to get it's IP address automatically by DHCP (acting as a DHCP Client) but you'd have to check the manuals to be sure. If you are unlucky, it might disable some of the ethernet ports, but I would be surprised if it did so.
 
now i have seamless wifi over my 4 floors!!!

i have a netgear R7000 i am using as the 'secondary'. What is the recommended configuration for the R7000?
I'm slightly confused. If everything is working as it should, why do you need to change the settings on the R7000?
 
I was using a Superhub 2 with DNS turned off, connected to my Superhub 3, and both having the same SSID and passwords, but my new phone just does not seem to like it.
I keep finding the phone has lost WiFi and then goes around in a loop desperately trying to reconnect.

Perhaps the WiFi chipset in the new phone is different, but it's not a happy bunny.
If I turn the superhub 2 off, it seems to be fine.

I'm wild guessing, as it's the same SSID and password on both, something signal wise is getting confused.

What is annoying though and I've only learned this recently is WiFi devices won't actively go out and look (much) for a stronger signal, but will hang onto the weak one almost as long as they can.

I did some test and I walked away from my Superhub 3, and stood in front of my SH2 and it was still sticking on the SH3 did not wish to let go.
If I disconnected and reconnected wifi, then it jumped to the SH2, but again, I walk back to the 3, with a weak signal and it won't jump to the 3.

Only if it's literally dying does it want to let go and use the better signal.
Seems a bit silly, but hey....
 
Welcome to our world (of pain.) What you describe is known in the Wi-Fi business as the "sticky client" problem. I often cite that it is "Big Wi-Fi Myth Number 2" that clients are "hunting for the best signal" and as you have experienced, they do not and some clients need it to get pretty grotty before they instigate a "roaming assessment." I've done tests on large sites with lots of AP's where I've seen a client bind to one AP (called "Associate" in Wi-Fi speak) and I've then walked past a dozen other AP's before my client decided it's service was sufficiently poor it should look for a better option.

In newer Wi-Fi standards there's a kind of "hint" mechanism where the AP's (there are Wi-FI AP's inside the "routers" in your case) can suggest to a client that it might do better using a different AP. But it's still up to the client as to whether it wants to "take the hint" so to speak.

However, this only used to be available in enterprise scale systems where the AP's "talk" to each other to swap information about which AP can "hear" which client the best and/or is more lightly loaded and often "pre-stage" the roaming handoff (credentials and all that) to expedite the roaming process.

A heterogeneous mix of effectively stand alone AP's lacks such functionality and thusly is all in the hands of the client device (and therefore that devices designer) as to how to handle roaming - it's not mandated in the standards.

The "good" news is that some of the enterprise system functionality is beginning to trickle down into the SOHO market and is often included in some of the newer "whole home" and "mesh" systems (though that is becoming something of an overloaded and meaningless term,) though of course, you'd have to go out and buy one and replace your SH's to gain the function.
 
Last edited:
Out of interest, say you have two devices pumping out WiFi.

Would you recommend they both use the same SSID or not?

I was originally told, yes, make them both the same, SSID and Password.
And then, the other day I found a YouTube video saying No, you should name them different, else you will never know or be able to specify which device uses which item.

Any rules?
 
There's no "right" or "wrong" way to do it, both approaches have their strengths and weaknesses.

With SSID's the same, you may get automatic roaming as/when clients feel like it, but you'll never know for sure which AP you are associated with. With differing SSID's, clients will never roam until they completely loose connection and start again as if you'd just turned it on. You have to explicitly choose which AP you connect to (and flip them yourself) but you will know for sure which AP you are associated with.

On a big site with dozens/hundreds of AP's, I wouldn't dream of making all the SSID's different and compelling my user base to pick their AP as they wander around the site. I make them all the same, but then I've got enterprise scale solutions with all the fancy steering etc.

At home I've got two AP's and only one user, so I have them on different SSID's so I can pick and choose depending on which room I'm sat in. And of course, I'm a relatively sophisticated user that knows what's "up" - I wouldn't fancy trying to explain it to my sisters kids, so would perhaps keep them the same.

Similar arguments prevail over whether one names the SSID's differently in dual-band capable devices - does one name the bands the same or differently..? I'll just leave that one hanging.
 
Last edited:
So, you are saying in effect that having 2 items with the same SSID will allow a device to "jump" more easily to a stronger signal than if they are different names?

Weird, as the incorrect image I had in my mind, in the past, was that, it would all be very fluid and it would simply pick the best signal at all times.
In reality it seems that's not what happens, as they said, they want to hang onto a weak signal despite a stronger one being there for the picking.
Seems lie a design flaw does it not?

Like trying to scrape the last bits out of a peanut bowl, whilst you ignore the full bowl right next to you, as you don't take any notice of the full bowl, unless the old bowl is totally empty ;)
 
Indeed: But consider the contrary situation - if one by chance happened to be exactly half way between two AP's and we're constantly "hunting for the best signal." Due to natural fluctuations in the signalling conditions, which AP is "best" might vary moment to moment and the client spends more time flipping between AP's for no gain.

If you are already in session with an AP that is providing "good enough" service, one might as well stick with it and put up with occasional speed variance, (which happens more often than you might think) and the odd re-transmit and noise blips and so forth. It's not just about "strength" of the received radio signal - it's more complicated than that.

But there's the rub - just what is "good enough" service? It's not prescribed in standards and is largely up to the client device designer. Some laptops I've owned offer a degree of control and have a "roaming aggressiveness" in the NIC drivers setting, not that I ever thought it made much difference. But a lot of more "appliance" type devices aimed at consumers (phones, tablet's, etc.) don't give the user any control and you are completely in the hands of the devices designer.

The basis of the paradigm with SSID names is that with SSID names the same clients will believe the AP's to be part of "the same" network and thus eligble for roaming without user intervention. Whereas with SSID names different, clients consider these to be "different" networks and hence won't roam between them.
 
Last edited:
Agree,
I guess the ideal scenario would be where a device knows the WiFi that's available for it to use, and in some way monitors levels, so when the one it's on drops a significant amount below another it just flips over.
Of course I'm speaking without any care of how hard this would be to implement such a system.
If you took your phone around the home/garden, then of course in an ideal world this would just happen as you went from room to room.
Probably eating battery as it constantly did this! ;)
 
If it were me, I'd want my clients to be looking for trend, rather than just responding moment to moment and across several metrics such as the signal to noise ratio, the amount of packets received in error, re-transmit rates, actual versus nominal link rates, and probably some others if I sat and thought about it. But so doing would make the client device implementation more complex of course.

You are correct about the constant hunting eating up your batteries. It also eats into the performance. For example, to do a check of what's currently available, the client need to stop what ever it's doing, tune in to each of the possible radio channel (13/14 of them in 2.4GHz waveband, dozens in the 5Ghz waveband at time of posting) listen out on each for any AP's advertising themselves & collect some data from them (including signal strength) then return to what it was doing. Again, if one considers it from the perspective of a client getting "good enough," (whatever that might be,) service, one can ask "why bother" until it needs to.

Of course, if one can have "the system" manage all this sort hing for you, it frees up the client from doing so (and makes them simpler and cheaper and less battery hungry) which enterprise scale systems managers such as I lobbied the standards bodies to avail and we got the "hint" process whereby the system can suggest to a client it would be better off roaming to a different AP.

We also like to use it for load balancing: Imagine I've got (say) 4 AP's that can all hear 40 clients thereabouts. If all clients simply bind to their "loudest" AP, I might (for example) end up with 30x clients on AP1, 7x on AP2, 3x on AP3 and nothing on AP4. It would be preferable to steer a few clients away from AP1 to the others and balance out the numbers nearer to 10x clients per AP. It may thusly be the case that client No. 12 (say) thence "only" gets nominal 150mbps link rate on AP4 versus 300mbps on AP1, but because it's now getting less air time contention, (10 way versus 30 way) it may well be able to get more bits shifted over the airwaves per unit of time - ie it's "faster" even though the nominal link rate is slower!

Some of this sort of thing gets really deep and enterprise scale system give you reports up the wazoo so the system manager can try and tune the deployment "just so."

It's interesting to see some of these techniques trickling down (albeit mostly automated) into the home marketplace with the "whole home" type systems where we are now seeing "fleets" of managed AP's and sometimes a management app. Though if I were cynical, I'd observe its also a way to get buyers "locked in" to a particular vendor's product range if/when they want to expand their systems.
 
Last edited:
Thanks for the interesting post.
You are obviously really intelligent, and clever when it comes to an understanding of this stuff.
Myself a mere mortal, just wonders why it does not work better ;)

Sometimes however, in defense, is can help to have someone who does not have background on things to ask the simple questions, and find out things work the way they do, because that's the way they always have worked, and extra has been piled up upon something simple in the beginning to get where we are today, and a totally clean sheet, with a fresh design for what's needed today, would be ideal, though due to current hardware that needs to run, it's simply not realistic to start from scratch.

Part of the reason PC's and software for them are not where they could potentially be, if the scrap it and start from scratch approach we had in the 1980's / 1990's was still in place, when we went from ZX80 to Amigas kind of thing.

Do you think there could we a way to improve this, so devices could pop on and off from one signal to the next with more ease than is currently practical?
 
Thank you for your kind words.

It does indeed make for an interesting dialogue when people question why things are as they are - not least because sometimes it causes "insiders" to take a pause, have a fresh look at something that has "always been done this way" and ask themselves, "are we working this way simply because we have never thought to do otherwise..?" And it can make for an interesting narrative for subsequent readers.

As you describe, a lot current IT solutions are as they are because they have evolved from other things and/or been re-purposed for uses that were never contemplated when they were originally designed. Look at Internet email for example - if we knew spam was going to be such a big problem, SMTP (the email protocol most widely used) would have been designed at the outset to deter and/or prevent spam. (There have been some attempts retro-fit improvements, but they've not gained much traction.)

Similarly, if we'd known streaming media was going to become a thing, there are things that would have been built into network protocols (all of them, not just Wi-Fi) to make streaming more reliable and simple to use. As it is, media streaming has been reverse engineered on top of network tek that was designed for short "bursty" interactions instead of constant, continuous time critical delivery. It's kind of like trying to deliver television using the postal system!

Of course there are numerous other exemplars of "if we knew this is where we were going to end up, we'd have started from somewhere else."

I am not aware of anything new in development to improve roaming, but that doesn't mean the boffins aren't beavering away on it in some darkened room somewhere. Newbies do tend to rather obsess over "signal strength" as if it's the only important factor, but it's only one factor and (arguably) not even the most important: Signal "quality" can be argued to be much more important - a quiet but high quality signal can allow more data to be encoded than a loud but noisy signal. Viz: When I go to a Motorhead gig with my mate, I can stand next to him and yell at him and he will bearly be able to hear me and be saying "what" a lot - he's definitely getting a "loud" signal, but it's a very noisey (corrupted) one (and I combat it by speaking slowly, dropping the pitch of my voice. and moving closer to him.) Cut to next morning when we're sat in an empty pub having lunch (and our ears have stopped ringing) I can more or less whisper to him from across the room and he'll "hear" everything just fine.

Wi-Fi is much more complicated than a lot of people realise - even people in the IT business who sell, maintain and install it. I could describe the basics of wired ethernet (and ethernet switches) in a pamphlet - my reference texts on Wi-Fi are getting on for a thousand pages!

Keep asking the questions though if you have any more: As my Grandad used to say, there's no such thing as a "stupid question," only those too stupid to ask. You'll generally get a sympathetic response in this corner of AVF and are less likely to get the sort dismissive, disparaging remarks one sees in some other forums.
 
I'm currently looking at purchased a reasonably priced WiFi router, so I can place my Superhub 3 into Modem Mode and use the new device for the WiFi signal.
With much general reading from many 'normal people' there seems to be quite a good level os success/improvement in going down this route.
 
Incidentally, you don't need to buy a new "router" just to change your Wi-Fi: all you need is an Access Point (AP.)

Most SOHO routers will let you simply disable their in built Wi-FI AP in a couple of click and leave everything else (switch, router, NAT, firewall, DHCP, DNS, etc) intact. If your SH is facilitating all these other function OK, then you don't need to place SH into "modem mode" just to disable it's Wi-Fi.

Indeed, you could even contemplate leaving you SH as is (Wi-Fi and all) and deploy you shiny new AP "somewhere else" and run multiple cells which will give a guaranteed coverage improvement compared to a forklift replacement of your SH. But then you are back into roaming and you'll need to find a mechanism to establish the "backhaul" links between any outpost AP's and your router.
 
Last edited:
Just purchased an 'Archer C9' which should arrive mid next week.
Reading reviews, despite it not having 78 antenna's and looking like an Alien Mother-ship, from what I'd gleaned from user reviews it gets a pretty strong thumbs up for actually doing a nice job when it comes to sending out a nice strong WiFi signal in comparison to the Virgin Superhub does on it's own.

I have my WiFi analyser app on my mobile so will be easy to see how strong a signal and how far it reaches in comparison.

I'll post my findings up here, in about a week after I hopefully have it, and gather actual real world results :)
 
Just purchased an 'Archer C9' which should arrive mid next week.
Reading reviews, despite it not having 78 antenna's and looking like an Alien Mother-ship, from what I'd gleaned from user reviews it gets a pretty strong thumbs up for actually doing a nice job when it comes to sending out a nice strong WiFi signal in comparison to the Virgin Superhub does on it's own.

I have my WiFi analyser app on my mobile so will be easy to see how strong a signal and how far it reaches in comparison.

I'll post my findings up here, in about a week after I hopefully have it, and gather actual real world results :)
Don't forget the signal is two way. The Router might shout, does the recipient have good ears and a matching voice?
 
Don't forget the signal is two way. The Router might shout, does the recipient have good ears and a matching voice?

I have plans.....

fbca85e0dc473db233f5262a7497b854.png


322fcd670837b75eed0097ee89e3b8b0.png


7316e7631ba65fe4f19f0de0046baff8.png


:thumbsup:
 
Most of my static equipment is now hard wired ethernet. Due to the flakey mobile data signal I have multiple APs for me and my phone really. I hate dark spots.
 
Most of my static equipment is now hard wired ethernet. Due to the flakey mobile data signal I have multiple APs for me and my phone really. I hate dark spots.

I like wires :)

Want to run some cables to two security cams, trouble is, I've made the holes, but getting the wire from the hole, around a 90 degree and up bend and into the loft have yet defeated my many attempts.

:(
 

The latest video from AVForums

TV Buying Guide - Which TV Is Best For You?
Subscribe to our YouTube channel
Back
Top Bottom