Low Wifi signal

Recently we have been having problems with weak wifi signal at the far end of the house from the router. For ages the signal was great and just as good as if you were sitting near it.

Our Fire TV stick has recently been reporting weak signals from the wifi.

We are using the original Netgear WNR1000 we had when we took out our fibre. The white Thomson box never had great signal.

Is it worth getting a new router or updating the FW on the WNR1000?

You might just need to change the wireless channel, someone nearby might have setup a connection on the same channel that could be causing some interference.

A router with more physical antennas would give better signal strength.

Most likely a close neighbour has just got a wifi router.

Try a wireless channel change. Otherwise, get in touch with PlusNet and request their HubOne, it is a rebranded BT Homehub 5A and the range will be a lot better than your existing router.

If that fails, get a wireless repeater such as this https://www.amazon.co.uk/TP-LINK-RE355-Universal-Ext...

Rather than just guessing about channel sharing, install an app on your phone or laptop so you can see channel usage. I use this on my Android phone: https://play.google.com/store/apps/details?id=com.fa...

I have sorted out several people's wifi problems using this. The "auto" channel assignment that is the default on most routers isn't very effective. I would have hoped it sniffed around and used the freeest channel but judging by what I've seen that just isn't true.

I assume that is to the OP as I already know that.

Often this is misleading, since devices will choose 1, 6 and 11... People get an app such as this, and choose channel 3 as it appears to be free & screw up the spectrum for everybody following the standard convention.

Edited by ukhardy07 (Fri 24-Jun-16 09:57:39)

The problem is wifi access points tend to choose busy channels from my experience. The problem then with manually setting is that you set it to a channel that is currently clear and 5 minutes later a neighbours router may jump to overlap it.

I have been asked to go over to a friends because their wifi is performing poorly (and someone they know told them it was because of their ADSL filter!!!). I will hunt down a free channel and set it on that which will probably fix it but what's the betting a neighbours device will flip over to it automatically and the problem will be back.

has the aerial on the netgear been moved? maybe folded flat instead of upright?

Have the same problem with a Talk talk router next door, constantly follows my wifi signal, which then annoys my son on his laptop in his room.

manually change the wifi channel using a wifi analyzer as a test tool.

In reply to a post by blfamily:

Have the same problem with a Talk talk router next door, constantly follows my wifi signal, which then annoys my son on his laptop in his room.

????

I don't understand what you are saying there.

Take a good look in your own house for anything, usually substantial, that has been added, moved etc.

One example particularly is mirrors, due to the metallic "silvering" required to achieve the reflections.

It might not be in the same room.

My WiFi Repeater stopped working recently when a metallised crisp packet was placed in the waste basket alongside.

Taking that packet out of the basket, without moving or disturbing anything else, let the repeater start working again.

I wonder if anyone has experience of the metal foil backing on plasterboard etc.

In reply to a post by blfamily:

Have the same problem with a Talk talk router next door, constantly follows my wifi signal, which then annoys my son on his laptop in his room.

manually change the wifi channel using a wifi analyzer as a test tool.

Having 2 wifi routers on the same channel is generally not an issue. When you say it annoys him, is there really any noticeable issues or just the son notices they're on the same channel and becomes annoyed?

Often times, the routers on AUTO do choose the same channel as one another, and often this is intentional. There can be worse interference on the other channels such as baby monitors, microwave ovens, wireless doorbells, cordless landlines, wireless HDMI transmitters, bluetooth keyboards, bluetooth headphones, bluetooth mice, bluetooth fitbit fitness trackers etc. You do not see these devices in operation. If 2 routers are close by and are selecting the same channel, chances are there is other interference in the area causing that. You are probably just forcing your own device to use a channel swamped with those devices you are not seeing & all of these apps inssider, android apps etc are generally ignoring, as they only look at the wifi being broadcast.

wifi analyzer will not map out every device using 2.4ghz, it only maps out the wireless routers, as such it often causes you to choose a bad channel. Routers on the same channel do not cause the major issues at 2.4Ghz, it is the often analog devices such as a baby monitor which cause the mayhem. AUTO generally does a decent job at avoiding this, but the milage varies router to router.

Edited by ukhardy07 (Sun 26-Jun-16 00:19:07)

To follow on from UKHARDY's excellent explanation.

There are only 13 Channels in the 2.4 GHz WiFi Band.

Whilst the range or distance covered is relatively short, if you take a laptop or suchlike with inSSIDer or some other WiFi monitoring program to the business area of any town, you will see that by name/SSID there can be hundreds of separately identifiable WiFi circuits in operation, in very close proximity, clearly limited to those 13 channels, presumably successfully.

If there was any great amount of interference of any sort, would there not be a great outcry?

Even in metal-bodied trams and busses, travelling through the streets, hundreds of WiFi circuits can be seen waxing and waning as the vehicles move along, with the vehicle's own WiFi being the only "fixed" one.

Edited by deleted (Sun 26-Jun-16 07:49:38)

The Gadget Show did a test with microwaves, baby monitors and Christmas tree lights and found little or no effect on wireless signals.

I suspect with microwave cookers, it would need dirty door-seals to create a "leakage" path.

Having assisted in the development of cavity magnetrons, the RF source in such cookers, back in 1954 and 1955, I have a particular interest in them.

I am surprised how clean the microwave cooker door seals remain over the years, compared with their interiors, when things go awry.

Have any forum members experienced this particular form of interference and been able to examine the source cooker in any detail?

Yes it is usually a rogue device that causes the issues, not a baby monitor etc in general. 9/10 it is fine, but too often these devices do perform poorly. You see much more often things like wireless TV transmitters causing issues than routers themselves at 2.4Ghz. By no means does a baby monitor, bluetooth device running etc equate to a poor wifi network everytime, if it did there would be a lot of issues, it is just a small amount of the time. Needless to say the point is that choosing a channel yourself often misses interference. Then again, some routers also do not account for this interference either on AUTO (and some do).

I concede eckidoo's greater experience of microwave technology. However it's still notable that you see clusters of routers on a few channels, usually 1 and 6 in my experience. I have also definitely improved wifi by switching to apparently clear channels.

Without a spectrum analyser (or at least a scanning receiver) it isn't possible to see what else is occupying the same RF spectrum. I doubt it's possible to make the wifi radio in a smartphone do this job but happy to be proven wrong. Perhaps a software defined radio device that covers the 2.4GHz band might give a fairly low cost solution.

The wavelength of 2.4GHz is about 12cm. So bits of metal that are that size, half that size and possibly a quarter of that size could have profound effects on the wifi signal if close to either the router or the phone/tablet etc. One day i'll take my router out of the cupboard where it's close to the consumer unit, burglar alarm and see if makes a difference.

As an electronics engineer with much experience (admittedly not much with RF and antennas) I still find it pretty mriaculous that you can put multiple radios and their antennas in a tiny box, hold it at a random angle with proximity to conductive elements (humans!) and get them to work fairly reliably. But that's exactly what our phones, tablets etc do.

Thank you.

I generally agree with your technical comments, particularly the need for a genuine spectrum analyser or a software one to run on a PC. Must try the latter again, unsuccessful in previous attempts.

Agreed particularly with your final paragraph. I wonder how many realise the miniscule scale of most WiFi Tx/Rx "units" - about a very tidy wrist watch; and Bluetooth ones being even smaller.

Combine with the very unusual design of the typical WiFi aerial usually hidden inside the screen part of laptops.

Again being of the size of a very thin wrist watch; and many resembling a squared version of the "Isle of Man" symbol.

------

My Bright Box 2 is set to AUTO Channel Select and to extend DOWNwards if necessary, on both WiFi Bands, 2.4 GHz and 5 GHz.

At home, there are relatively few other WiFi systems to compete with; and I am uncertain as regards other potential interference sources, other than the reasonable assumption of one microwave cooker per household; and there may or may not be a baby monitoring system "two doors along".

Occasionally, busses equipped with WiFi have shown up - very rare as although there are about 3 services passing the house, they are all local, so generally less likely to have WiFi on board - although the facility is clearly increasing.

Using ACRYLIC on an older laptop but which also has 5 GHz band facilities, the situation is-

Chan SSID
11 My Brightbox 2
11 My WiFi Extender, one of the BT range
40+44 My Brightbox 2

There are two other SKY WiFi systems on Channel 6

Plus a third SKY, 9 BT and 1 "Direct DB" all on Channel 1.

Two of those BT have "with FON"

All of the 2.4 GHz have a "Max Speed of 144.4 Mbps, the Direct DB being 72.2 Mbps

TalkTalk has just appeared on Channel 3, 270 Mbps.


So clearly not approaching the quantities I have noted in town etc.

Strictly speaking, a microwave oven should use 2.450GHz while wifi uses 2.412 GHz to 2.472GHz. But microwave ovens aren't designed for frequency stability so even fairly small leakage could muck up wifi. Even if it's spot on frequency if there's a lot of leakage from the oven I suppose it could saturate a wifi receiver where the front end selectivity won't be sufficient to reject 2.450GHz

Useful information about usage of the 2.4GHz ISM band:
https://en.wikipedia.org/wiki/List_of_2.4_GHz_radio_use

Thanks for the link.

One of the aspects we were investigating was whether there was any relationship of the "natural frequency" of the machined anode block, immediately before final assembly with cathode, glass seals etc, to the eventual working frequency of the assembly.

The equipment sounded like SONAR/ASDIC pinging away all day.

In reply to a post by ppppenguin99:

Strictly speaking, a microwave oven should use 2.450GHz while wifi uses 2.412 GHz to 2.472GHz. But microwave ovens aren't designed for frequency stability so even fairly small leakage could muck up wifi. Even if it's spot on frequency if there's a lot of leakage from the oven I suppose it could saturate a wifi receiver where the front end selectivity won't be sufficient to reject 2.450GHz

Useful information about usage of the 2.4GHz ISM band:
https://en.wikipedia.org/wiki/List_of_2.4_GHz_radio_use

I thought 2.450Ghz was channel 9 on the 2.4Ghz WiFi spectrum?

You're right. 2.45GHz is between channels 8 and 9

Wifi chanel frequencies: https://en.wikipedia.org/wiki/List_of_WLAN_channels

As you have indicated, individual magnetrons can vary and do, given that the cathode voltage is directly mains derived, so their working temperatures etc can vary significantly.

This alters their physical dimensions sufficiently to move their operational frequency to a fair extent, as their frequency of operation is highly dependent upon the "rising sun" anode cavity. I forget the description of the other type of anode, with circular apertures at the ends of the slots.

In other words, very distinctly different from, for example, crystal-controlled oscillators, and even typical LC tuned circuits.

It also meant that tuning the equipment involved highly risky procedures.

I have left ACRYLIC running for a couple of days, in another part of the house where it is likely to pick up more WiFi circuits.

There are now 26 circuits showing up.

My own three remain as was.

The other 23 are-

Ch - Quantity
1 - 6
2 -1
3 -1
4
5 - 2
6 - 8
7 - 1
8
9 - 1
10
11 - 3
12
13


Channels 1, 6 and 11 not unexpectedly have most.

Again not unexpectedly, there are no others in the 5 GHz Band.

One SSID started with AUDI.... on Channel 1, is this likely to be an in-car WiFi?

Don't forget that although the wifi peaks at the chosen channel, it spreads to some extent over the two each side. Also if anyone triggers the 40 setting rather than 20, then two separated channels are used.

Agreed as with any modulated radio transmission, whether AM or FM etc.

I am tipping the ones using non 1, 6 and 11 channels are talktalk routers. They really mess up the entire wireless spectrum by defying all standard conventions. They are also the ones set to use 40Mhz as well.

EDIT: Audi do have in car WiFi yep.

Edited by ukhardy07 (Wed 29-Jun-16 21:13:56)

Sorry if i've missed it somewhere... what exactly is the issue with using channels other than 1, 6 & 11?
From the spectrum usage shown here https://1drv.ms/i/s!AnOT_guNf7hGiKRJtvBt0b-iBiPJ9g , by using channel 13 to me it appears that i'm less likely to pollute others airspace and them mine.

Two are identifies as TALKTALK, Channel 9 with 270 Mbps and Channel 2 at 144.4 Mbps, resoectively.

Both have HUAWEI as "Vendor" and there are another two with that Vendoe; but seem unlikely to be TalkTalk..

Reloaded inSSIDer on the same laptop as ACRYLIC.

Restarted both.

General agreement of results.

-----------------

Not having used inSSIDer for a couple of years, today I noted two things-

Very good training lessons-

http://www.metageek.com/training/resources/


On the working inSSIDer, at the top-left, there is a line-

"Your Link Score would be improved if you move to Channel x"

No idea of how effective at present; and don't want to disturb my present satisfactory settings.

No the aerial is pointing straight up.

How much will the HubOne cost me from PlusNet?

Another aspect is basically the aerials in his laptop or such-like should ideally be facing "square-on" to the local source of his wifi signals.

Over the past two days, I have been mapping the signal strengths separately for my three wifi circuits.

I was using NetSpot mentioned in this issue of "Computeractive" on an older laptop which basically is the main user device that has 5 GHz as well as 2.4 GHz.

During some of the measurements, it was obvious that the Signal Strength varied with respect to the angle of the screen, the aerials usually being fitted at and behind the "top" of the screen.

Typically sitting with the screen at a comfortable angle, the WiFi symbol at the bottom right, was full and solid.

If I simply angled the screen away from me towards the horizontal, it could reduce to being the "centre dot" only.

So positioning is relatively critical.

As expected, walking one pace could also change the signal strength deastically.

In reply to a post by 10forcash:

Sorry if i've missed it somewhere... what exactly is the issue with using channels other than 1, 6 & 11?
From the spectrum usage shown here https://1drv.ms/i/s!AnOT_guNf7hGiKRJtvBt0b-iBiPJ9g , by using channel 13 to me it appears that i'm less likely to pollute others airspace and them mine.

http://www.metageek.com/training/resources/why-chann...

Hope this helps

It is likely to be free if you renegotiate the contract.

Wouldn't it be better to use 1, 5, 9, and 13 ?

In the US there are only 11 channels available, so it makes sense to use 1, 6 and 11.

In Europe there are 13 channels available, so yes, 1, 5, 9 and 13 may very well make more sense. That said, it would only make sense if everyone stuck to it. Unfortunately, Cisco, BT, Sky et al will choose 1, 6 or 11 on AUTO channel selection, hence why 1,6 and 11 is the best choice here.

So I got in touch with PlusNet about upgrading our router to the Hub One. I was told that it was 99.99 or if i recontracted I could get a better deal and get the hub for free.

Im on Fibre up to 78 unlimited at £19.99 but get £3.00 referral which brings it to £16.99. The deals on the website start at £22.99 so I fail to see how i could get a better deal.

Hope you jumped on it - the router is worlds better.

At £100, I didnt.

Might have to wait a while longer.

Did you not negotiate the price if you recontracted? It is often better than the new customer deals.

You can get the hub on ebay for £35 also.

What is the model of the hub?

PlusNet Hub One

Would the BT branded one work? Can get that cheaper

Yes BT HH5A and 5B will work

Connect to HH at 192.168.1.254
Login to the hub using the password on the card supplied by BT
Click settings
Click enter username and password
Username: <whateveryourplusnetusernameis>@plusdsl.net
Password: Whatever your plusnet pw is.

I bought a BT HH5B for £30

How do I connect it up to the phone line? When we had FTTC installed we had the extension kit installed to bring it closer to the computer desk. Do i plug a CAT5 cable into the extension port and the other end into the WAN port?

No you need an RJ11 modem cable

In reply to a post by ConorMK:

I bought a BT HH5B for £30

How do I connect it up to the phone line? When we had FTTC installed we had the extension kit installed to bring it closer to the computer desk. Do i plug a CAT5 cable into the extension port and the other end into the WAN port?

You need to get rid of absolutely all equipment you currently have, modems, routers etc.

The Homehub plugs straight into the telephone socket, no modem required. Ignore the WAN port, you do not need that, just plug the router into the telephone socket with a standard RJ11 cable.

If you have a modem at current, the cable which connects the modem to the phone socket, is the one you plug into the back of the homehub.

Edited by ukhardy07 (Tue 26-Jul-16 17:43:01)

OK so its the one that goes into the existing white modem?

That is the one.

It depends why you got the HH. If you have an unlocked HG612 and still want to see the full stats you just plug the ethernet cable from the HG612 into the HH WAN port.

If you want a single-box solution then as ukhardy says.

Okay back online. HH5 is working well. Just have to go around the house now and rejoin all our devices to the new wifi.

RobertoS, I got the HH because the wifi on the WNR1000 wasnt all that good and our Fire TV stick wouldnt connect. I have tested it now and its working well. Thanks to you, UKHardy and Batboy for your help.

Bet the signal is a lot better!

A quick fix, if you are feeling lazy - just set the same SSID, password and encryption as before. Your devices wont even realize a new network is setup.

EDIT: Or keep it as BT, everyone can think you're flashy with the most expensive main-stream provider

Edited by ukhardy07 (Tue 26-Jul-16 18:52:38)

I have noticed the wifi signal is better.

The wifi i need to change anyway as its been the same since we had fibre installed. About time for an update

I don't know what options you have on the Home Hub 5 interface, but whilst you're working on your Wi-Fi, it's worth turning off some of the compatibility options that you likely no longer need that will be hurting security and, quite possibly, performance.

Try disabling WPA and use WPA2 only. This might involve turning off a "'WPA Mixed Mode" setting somewhere - it depends on the router. WEP is hopelessly broken and can no longer be considered any sort of security whatsoever.

Try disabling support for 802.11b, and operating the 2.4GHz radio in g/n only mode. There are few useful devices left that do not support 802.11g.

Ideally, you should set a WPA2-Personal passphrase of 63 random characters unless you are in a position to run your own RADIUS server and switch to WPA2-Enterprise.

Try disabling WPA and use WPA2 only. This might involve turning off a "'WPA Mixed Mode" setting somewhere - it depends on the router. WEP is hopelessly broken and can no longer be considered any sort of security whatsoever.

It comes out of the box like this

Try disabling support for 802.11b, and operating the 2.4GHz radio in g/n only mode. There are few useful devices left that do not support 802.11g.

Having wireless B support will not slow anything down unless a wireless B client connects.

Ideally, you should set a WPA2-Personal passphrase of 63 random characters unless you are in a position to run your own RADIUS server and switch to WPA2-Enterprise.

Seems a huge overkill in my view. Imagine entering that on every device. I understand it does help with a brute-force attack, but a 10 to 16 digit random password should suffice in the real world.

The biggest security aspect to consider is WPS, which can be compromised relatively easily. Even with a 63 digit password, with WPS enabled, I am pretty sure I would be able to get the full 63 digit password overnight, possibly in a few minutes depending on the version of WPS in use.

WPS is an 8 digit pin number, hard coded into the device, so vulnerable to brute-force.
Some manufacturers made the WPS PIN a combination of the routers MAC / Serial number which is one issue...

But the biggest issue:

The 8th digit is generated based on the first 7 digits, so it is predictable
There are two stages to authentication, M4 and M6.

Say my PIN is 12345678, M4 is 1234, M6 5678

WPS will drop the connection immediately if you make a mistake after M4. That is, the first 4 digits must be correct before the AP will process the second 4 digits.

This greatly reduces the time required to brute-force. 10^4 is only 10,000, so 10,000 possible combinations to figure out the 1st four digits.

The 2nd half, well the 8th digit does not count as it's not random, so we only have 10^3 = 1000 combinations.

Then we can get the 8th digit, based on the first 7 digits.

So at absolute maximum we have 11,000 different combinations. This would take 3 hours to brute force at 1 attempt a second, but often the AP performs fine with 1 attempt every 0.5 seconds, ie 1.5 hours, and this is assuming you brute force it at the 11,000th attempt, when actually it may happen on the 1,000th attempt ie 15 minutes into running.

Once you have the WPS PIN, the AP will just hand over the full WPA-2 password, it does not matter if it's 63 digits or not.

Change the password - it does not matter, I have the hard coded, unchangeable WPS PIN, your router just hands me over the new PW.

Edited by ukhardy07 (Tue 26-Jul-16 19:54:20)

Agree fully with your final paragraph - it is miraculous!