Post
by Albert H » Mon Dec 26, 2016 10:50 pm
The problem with 2.4 GHZ is that everything else is there.
Years ago, before 5GHz became used much for wi-fi, I used a video link up there for about 450m between blocks. It worked OK, but the signal to noise ratio wasn't great because I didn't use much of the 6MHz wide video channel. Trying it again, I'd use very wide deviation.....
The best - and most secure - links I've ever done were in Band IV and V, with TV Yagis at each end. The gain from a TV aerial is huge, so you can go a long way with milliwatts! A TV aerial doesn't look out of place anywhere, so it's not easy to decide which of the dozens in a street - or even on a single rooftop - are being used. I linked 22 miles with two 48-element Yagis with just 100 mW going up the coax. It was fully noise-quietening stereo, and had a return telemetry path using the same aerials!
Look on the interweb for the Dutch 60cm stuff. They used to have an unofficial broadcast band up there! Modern UHF transistors make the electronics fairly simple - way back, I used endless stages of multiplication to get there, and used a pair of 2N3866s in grounded-base parallel to get my output. It wasn't easy. Nowadays, common devices like the BFR96 will give you an easy 250mW at ~500MHz and just cost pennies. Small trimmers aren't easily available these days (at sensible prices) so I use Lecher lines with capacitors to ground at appropriate points. An absorbtion wavemeter is essential for aligning these stages since trying to watch the spectrum analyser whilst soldering on parts can be tricky. I have a wavemeter with an audible output - as the signal strength rises, so does the pitch of the tone from it. This makes alignment really easy.
The link receivers can be based on TV tuner front-end modules. These are really cheap, and often include a PLL chip! The IF is normally 35MHz, so I mix that down to 10.7MHz so that I can use my normal PLL IF. The receiver sensitivity can be spectacular - and then you use a high gain aerial as well. The technology isn't too difficult, but some of the clowns throwing together rigs these days wouldn't have a chance. Most of them just copy old designs - even old layouts! I still see rigs using my 1982 exciter with the Colpitts oscillator, buffer (with the PLL take-off from the emitter) and a double - tuned doubler, into a BSX20 then 2N3866, with the voltage regulation between the 3866 and the last BSX20. Built properly, it's a clean and stable circuit, but it's certainly possible to cock up the tuning and generate crap.
I had one clown actually phone me up a few years ago to ask how to re-programme a 74HC4040 for a different frequency. Talk about clueless. I told him that if he couldn't work it out for himself, he'd better find something less technical to do, like flower arranging or train spotting!
"Why is my rig humming?"
"Because it doesn't know the words!" 
