Some of you may recall me moaning - a lot - about my suddenly non-functioning wireless network, back in April last year:
 
"Wireless connection dead almost everywhere. Neighbours the same. And no-one owning up to the kit that's killing it. EXTREMELY grumpy. May weep."

 I was demoing the issue to my brother-in-law yesterday - it's pretty clearly captured below:

This is ethernet connected to the router - and is at the lower end of what I typically get on the Sky ADSL, since I can almost spit out of the attic window and hit the exchange up the road. The router's in the attic, because - up until last April - this gave me 100% throughput everywhere in the house (and garden, for that matter). 802.11b, then g, then n, over the last 10 years odd.

This is by the front door - some degradation but I'd be very happy if I could always achieve this.  It can be a lot worse here, but rarely drops to the levels I routinely see - at best - at the back of the house...

In the conservatory at the back.  Where all the 'routine' internet use happens - children's homework, Spotify connected to the amp in the kitchen, general bits of work and leisure stuff. There isn't anywhere else on the ground floor that you can sit at a table and use a laptop comfortably. And this is as good as it gets. When the interference ramps up, the connection drops completely.

Which is all incredibly frustrating: I have several devices in the house which can ONLY connect via WiFi. I spent a lot of time last April working through all the possible causes but to no avail.  I even did a maildrop to my twenty nearest neighbours, which established I wasn't alone - but threw no light on the cause:

The rather obvious bank of phone-masts on the top of the nearby fire-station practice-tower - no longer used for practice, solely revenue - I'd mentally eliminated pretty early on: Mobile transmission is at 900 or 1800MHz, 3G at 2.1GHz. Neither, to my mind, should interfere with 2.4GHz WiFi. And they'd been no problem for the last 8 years in any case. There was no question that this was something other than another 802.11 network - the excellent inSSIDer will quickly show you what's operating around you and, in any case, the whole point of the protocol is to allow WiFi devices to co-exist and play nicely - sharing the same channels, you might see some reduced throughput, but not complete failure of the connection. More precise analysis needs at least a (£500-odd) spectrum analyser - and I don't know anyone who has access to one, let alone the expertise to use it. The only thing I've been able to do which has helped at all is hang a 5GHz AP off the router, which at least gives me a reasonable connection at the front of the house and enables the Apple TVs to do Airplay - prior to that, they refused completely, regardless of where they were sited. But it's no panacea - 5GHz doesn't have the range of 2.4GHz - so can't reach the conservatory - and several of the devices in the house can only operate with the latter.

And then I started thinking about the line-of-sight between the most interfered-with places in the house and the phone masts. And how there was fairly strong correlation between the likely number of walls in the way and stability of the signal. It had always bugged me that, bizarrely, the signal could drop to nothing just feet away from the router in the attic - until it occurred to me that the only obstruction between the masts and the device in that area are some rooftiles and a sheet of plasterboard... A bit of Googling turned up some articles penned since last year which leave me thinking my simplistic views about the frequencies involved may have led to me to look in the wrong places.

(If it turns out to be the phone-masts it would, of course, me more than a little ironic: There are four Apple devices here which get most of their usage in the worst-affected areas. They're sold by the mobile-phone companies on the basis that you can limit your 3G usage by connecting to Wi-Fi... but if the signal disappears, they'll use 3G. So the same companies that may well be responsible for knackering my wireless network also increase their revenues, as a result of doing so - at my expense.)

Back in April, I'd raised the issue on several web forums and usenet - and a number of people had opined that there was little else I could do: The regulatory bodies would have no interest in interference with unlicensed spectrum, so I was stuck with it. But, having thought through the issue again for the first time in months, I thought I'd give Ofcom a call and see what advice they could offer. Eventually talked to someone in 'spectrum' who asked me to lay out the issue in an email, although she didn't sound particularly interested. But, in for a penny... so I mailed them a copy of the letter I'd dropped in my neighbours' letterboxes with the following:

" I have a problem that has persisted since last April, with interference with my wireless broadband - the attached document is a letter I distributed to a number of my neighbours' houses at the time and will give you a detailed overview of the position as it was then. Since I conducted this 'survey' I have added a 5GHz Access Point to my house, but this seems to be impacted almost as badly.  The interference seemed to become less frequent over December/January, but I have twice completely lost the connection to the laptop I am typing on today. So beginning to feel 'end of my tether' abut the whole affair again...I'd be grateful for any advice you can offer on where to go next - my efforts to locate someone with spectrum analysis equipment (and the expertise to use it) have so far drawn a blank... "

And on Thursday I got a phone call from a very nice chap, who advised 'We've decided to take on your case', then laughed heartily at the extended, manic outburst of gratitude that ensued.

So - this Friday, Ofcom engineer Les is turning up at my house with his spectrum analyser, some directional antennae and a stack of clever software.  I'm trying not to be too optimistic about it: Sod's law says he'll turn up just as the interference drops. If he does detect it and it is the phone masts... well, I'm not sure whether that's actually going to help resolve it. But at least I might find out what the problem is.

Watch this space.