Link Receiver

[Albert H]

You really don't need synthesis for a Band I receiver - we always used to use either a crystal for the Local oscillator, or use AFC by adding a varicap diode to the Local oscillator, and take the AFC voltage from the demodulation part of the CA3089 IF chip. As long as the receiver was getting a reasonably good signal from the link transmitter, the AFC would keep the local oscillator exactly on tune. It was even possible to have a VFO link transmitter, because a little bit of drift wouldn't matter - the AFC would allow the receiver to "track" the frequency of the incoming signal!

[radium98]

i dont understand that a schematic will say a 1000 words if you can , anyway thanks , just i want to know how it works nothing more , if this is true ao why we use exciter pll and not staying on vfo

[yellowbeard]

Because there are a lot more things in the FM band than band I that you could drift in to with a VFO. If you went on FM with a VFO these days your arsehole would be twitching like a rabbits nose.

[XXL]

But the receiver and transmitter both still drift causing the audio to degrade surely, regardless to weither there’s another signal near the frequency or not. Especially the receiver if it’s in a rig box getting hot. I might try that AGC though. Hopefully will be stable enough.

[Albert H]

If the AFC is done properly, a small transmitter drift is tracked by the receiver, and there's no distortion issue. I used a little VFO link rig on ~64MHz, which used to move up by about 21kHz in the daytime, and down to about 17kHz below nominal at night when the studio end was cold. The receiver AFC tracked it for months without getting either distorted or (worse yet) "losing" the carrier......

The link transmitter was a "Kallitron" self-doubling oscillator (like the NRG stuff) with a 2N4427 final for about 850mW out, and the receiver was a simple Colpitts VCO for the local oscillator, a cascode of two 2N3819 FETs for the mixer stage, a BFY90 for the input amplifier stage, a couple of BF494 transistors for the IF amplifier, SFE ceramic filters, and a CA3089 IF chip to recover the audio and provide the AFC control voltage back to the VCO. The whole set-up cost about £15 (for both ends) and linked reliably over about 2 km for several months!

[reverend]

You really don't need synthesis for a Band I receiver - we always used to use either a crystal for the Local oscillator

Depends where you are. In an area where there are Band I links every few 100 kHz ( then a crystal or PLL is needed. An 'AFC' type approach can easily lock onto a nearby signal if it drifts too far or worse... if your own TX goes off, it will happily drift over and lock on someone elses!

[Albert H]

TBH, I gave up using Band I a long time ago. I only use Bands IV and V these days, because the aerials are small and even very high gain ones are cheap. Also, a TV Yagi seldom looks out of place on a building!

With a little care, it's perfectly possible to link for miles on milliwatts on Band IV.

[3metrejim]

By now I'd have thought you'd be linking with digital. These days you'd probably be better hiding in plain sight by using an ESP32 to receive an mp3 stream and have it spit out a stereo multiplex signal to a cheap i2s dac. You can connect into it (rather than having it connect to you) and use one of those pringle-can type antennas at the sending end - receiving end needs to be lower gain to limit the excess signals about and avoid aiming problems. I doubt the signal would stand out among every other wi-fi user and you'd not be able to tell what it is either (no matter what the TV licencing guys may claim; even using a network card in monitor mode). The signal passes through glass so you can have the transmit end indoors - no one will notice something hanging from a curtain rail (as long as you don't use an actual pringles tube ).

If you link by analogue there is always the chance someone will crash the link (happened to a real station not long back, not far from myself) or you'll be traced because you're the only signal with audio that no one else will be listening to (and is coming out on a broadcast frequency) - how long would it take someone using a spectrum analyser (hardware or software) with a demodulated fm output to find you? I'm guessing under an hour (depending on distance).

[Albert H]

There are several ways of obscuring an analogue link. I used to transmit white noise in the baseband from 20Hz - 15kHz, then have two ultrasonic subcarriers (for the composite stereo multiplex). Any casual scanning would just hear white noise (with a small increase in RSSI as you tuned through the signal), so would assume that there's nothing there.

One station I was involved with in Eastern Europe used "burst mode" to fill a memory at the transmit site - a half-hour's worth of programme sent in a few seconds. This worked well, and was never traced.

In the USA, we used a mixture of microwave out of the studio to a nearby high building, then two UHF "hops" to get to the output site. This also remained unfound by the FCC despite their really wanting to target us!

I've used various digital schemes more recently, with the most effective being a little single board computer connecting over wi-fi and streaming the content off a website. The cost was low, reliability was good, and if the link went down, it would play some pre-recorded programme until the link connection was restored. I did try using those Ubiquiti transceivers, but they're insanely expensive, and their range is never what the manufacturer claims. Point-to-point digital can work, but analogue stuff well concealed works just as well and very securely.

My best analogue results on Band IV were in the guard band of BBC2 some years ago - tracking it was frustrating with a Megawatt TV transmitter nearby! We got a clean, fully noise-quietening stereo analogue link over 18km using just 350mW at the transmit end (albeit into a 24-element Yagi). That was in use for a couple of years and never found! The wide open spaces of analogue TV channels made for some great results.

However, the ultimate link - not built by me, but I was involved - used a little spare bandwidth on a geostationary analogue TV broadcast satellite. We had the major advantage that one of the TV company engineers was a "friend of the station". It used a TWT-PA uplink, with the correct dither, and could be easily received anywhere in the highest field strength contour areas of their broadcast footprint - most of Europe!
The most difficult parts of the system were the mechanical ones aligning the dish really accurately and tracking the figure-of-eight positional shifts of the satellite repeater. Reception was by any cheap dish / LNB /STB combination with a simple analogue "descrambling" system (just frequency inversion of the subcarriers). That system was in use for over 14 years!

[3metrejim]

Some interesting methods there. The burst mode link sounds like something I heard on some documentary years ago, that spies used at one time. One station years ago round these parts would switch off the link, but leave the main rig on. It sort of sounded like nothing was there but you could tell the difference between no signal and relaying link noise, plus radios would tell you tuned and analogue ones with afc would fall onto it and get 'stuck' on it. I'm guessing one of these phone apps with an sdr dongle would be very portable and give demodulation abilities for tracking.

I do think an esp32 is worth a try - real cheap and depending on the set-up, you could use an old wi-fi router as a mid point - still semi trackable as I have a piece of software that can tell you the clients of a wi-fi access point and their mac addresses that can be used to help track them down - best bet is to find a public (or business) wi-fi router that lets you communicate to other connected devices - I tried scanning internal addresses on the local public council network, but that boots you for trying... oops - a bit obvious.

There's loads of code online for things like internet radio or smart speakers that should be fairly easy to modify, but I would want 320kB/s minimum for an mp3 stream (or a good vbr if supported). I do like the idea of having an access point as the receive end though, so you can use the signal strength at the sending end to get aligned (as long as you have a medium beam pattern / plain antenna, on the access point end) - none of the left a bit, right a bit, more, more, no, back a bit that you sometimes used to hear over the air during a link alignment session.

[Albert H]

I never really liked mp3 for streaming - OGG or AAC always sounded better for a given data rate, and were much more robust. However, there's loads of hardware and software for mp3, so it's probably the obvious choice.

One station I know uses a datastream through a couple of VPNs buried into a public wi-fi system ("The Cloud" available in lots of towns), and they further encrypt their data so that it doesn't look anything like an audio stream. Another uses the wi-fi system in a nearby school for their link!

[XXL]

How would one use an esp32 to link then ?

[3metrejim]

Easiest method:

Set up the esp32 with the correct hardware - I2s audio output and feed it into a cheap stereo coder at the rig end (no rds support). Load the esp with internet speaker software. Get the esp32 to connect to a router you control (mid point). Connect to the router to obtain the IP address of the esp32 'speaker'. Send it the command to connect back to a streaming server running on your machine that is also connected to the router (or use some free home automation software to achieve the same connection). You will need some gain antenna on the router to esp and from your 'studio' to the router if you want any distance.

Hard way:

Custom software on the esp32 that runs similar wi-fi speaker code but outputs stereo multiplex and rds through I2S (assuming the dac used doesn't have audio filtering, and that the esp32 is fast enough for the calculations at the same time as decoding a stream). Program the esp32 to probe open networks and attempt to 'phone home' on the wider internet with the SSID of the bases that allow communication (if any) along with it's (local) IP address and MAC. Connect to the same base station and attempt to communicate to the esp32, then proceed as above method. You will need some website that allows anonymous communication for the 'phoning home' so you can check for messages from the esp32. You may be better off 'war driving' looking for a nearby network that allows communication between devices. You need to be able to set up the esp32 remotely (same as for the simpler wi-fi speaker), in case of unforeseen problems (it can do base station and access point at the same time). Do not hard code any networks/passwords into the esp32 code, so that they are lost on power down. It's up to you whether you encrypt the software, but shouldn't be necessary (unless you are trying to sell devices).

You can pretty much program any method of linking that you can come up with, assuming you have someone to do the programming.

Using gain antennas would technically be illegal, but you are probably joining a fairly large club, and that's probably the least of your worries.

First tests can be done using a standard wi-fi speaker to see what the maximum range you can achieve is.

Note this is hypothetical, and I haven't tried it myself (yet). Maybe someone that has already done it successfully (Albert H

One station I know uses a datastream through a couple of VPNs buried into a public wi-fi system

) can shed some more light on the subject.

Don't expect any free code for this, but all the pieces you need, are floating around on github (time for some LEGO (r) ) coding.

[sinus trouble]

I do not think Band I is the best option these days?

However, I do find the concept of "Super Het" Receivers fascinating! ALL radio enthusiasts need to be familiar with how they work!

I have a theory about the "Sinus B1 RX" which i may implement in future posts?

The premise is to use Varicaps to control the input filter and LO via a single control voltage!

As the IF stays constant throughout! This would allow the RX to be tuned by a simple pot!

Not easy? But possible?

[Albert H]

Sinus - varicap tuning is how most VHF receivers work! What you could do is use a PLL to control the frequency of your local oscillator (set it either 10.7 MHz above or below your link frequency), and the control voltage for the oscillator varicap(s) could also be used to tune the input filter varicap(s).

There are only a couple of problems you'll need to overcome - the most significant of which is to keep the digits from interfering with your receiver, degrading its sensitivity. I always used my simple 3-chip synthesiser (74HC4060, 74HC4040 and a 4046) for simple link receivers. I used to enclose the three ICs and the associated crystal (usually a cheap 4MHz one) in a metal box made from PCB material, to keep the logic switching noise out of the receiver altogether. It made for a simple, effective, stable receiver, and with little work, it could be made to work on any VHF or UHF frequency you wanted. If my LO had to be up in the UHF, either I frequency multiplied up there from a VHF start, or I used a cheap prescaler IC (usually the SAB6456) which will divide any frequency up to over 1GHz either by 64 or 128. Obviously, this made it a 4-chip synthesiser!

The three-chip synthesiser works well up to around 60MHz, and the advantage of using the 74HC4040 is that its clock input is a Schmitt, so it's quite tolerant of signal level. The loop filter I used to use was just a passive one, and worked well enough!

[sinus trouble]

Sinus - varicap tuning is how most VHF receivers work! What you could do is use a PLL to control the frequency of your local oscillator (set it either 10.7 MHz above or below your link frequency), and the control voltage for the oscillator varicap(s) could also be used to tune the input filter varicap(s).

There are only a couple of problems you'll need to overcome - the most significant of which is to keep the digits from interfering with your receiver, degrading its sensitivity. I always used my simple 3-chip synthesiser (74HC4060, 74HC4040 and a 4046) for simple link receivers. I used to enclose the three ICs and the associated crystal (usually a cheap 4MHz one) in a metal box made from PCB material, to keep the logic switching noise out of the receiver altogether. It made for a simple, effective, stable receiver, and with little work, it could be made to work on any VHF or UHF frequency you wanted. If my LO had to be up in the UHF, either I frequency multiplied up there from a VHF start, or I used a cheap prescaler IC (usually the SAB6456) which will divide any frequency up to over 1GHz either by 64 or 128. Obviously, this made it a 4-chip synthesiser!

The three-chip synthesiser works well up to around 60MHz, and the advantage of using the 74HC4040 is that its clock input is a Schmitt, so it's quite tolerant of signal level. The loop filter I used to use was just a passive one, and worked well enough!

Nice one Albert!

Yes a Receiver will need the LO and the Antenna input filter to have "synchronised" tuning!

I am also glad that you mentioned your CMOS PLL solution, As i would like to pursue a project very similar soon!

After looking at the SPI PLL and another option suggested by Rev a while back, The HC4059 seems to be a problem sourcing these days?

Revs version is an interesting concept utilising the HC4020 to replace the HC4059?

Your HC4040 alternative seems like a good candidate for consideration too!

Anyways, Its in the pipeline and i think it would be great to have a nice compact PLL that can be tailored to different needs!

[Albert H]

The 74HC4024 will divide (reliably) from about 75MHz - sometimes useful as a cheap prescaler.

The 74HC4040 is cheap and easily programmed with a 4k7 from the reset pin to the 5V supply, then diodes from the reset to the divided outputs with the anodes connected to the reset. As it counts, the reset will be held low until the programmed number is reached, at which time there will be a (very brief) reset pulse, and it will start the count again. You take the divided output from the cathode of the most significant number diode, because this will be at the required divided frequency, but will have a (very) roughly equal mark/space ratio, giving nice wide squarewaves that you can see on your 'scope, and best suited for the input of the 4046.

You can also programme the 74HC4060 in the same way - resistor and diodes - to give whatever frequency you want for your reference. It's worth bearing in mind that the '4060 doesn't have every division ratio available, but there's enough that you can find a good reference frequency. This allows the use of a cheap reference crystal (I usually used 4MHz because I had a whole boxful of them originally bought for clocking Z80 processors - bought for about 4p each!).

Remamber - keeping the digits out of the front end of the receiver is essential if you want good sensitivity.

My link receivers usually used a dual-gate FET mixer, because I could get useful gain out of them and plenty of mix product, with either a current-biased BFY90 or BFR96 front end transistor (protected by a couple of parallel diodes on the aerial input). The output of the first IF coil would feed a BF494 or BF199, then into an first SFE ceramic filter with ~180kHz bandwidth. There was then a second IF amplifying transistor (as before) with another SFE filter feeding into the CA3089 IF chip. There was an RSSI-operated relay option, and output muting on loss of carrier. There was also a small area for an NE567 tone detector - we used to use tone-controlled access - and that detector could feed into a logic bistable for tone-on / tone-off to avoid other people "borrowing" the rig!

When correctly aligned, I could get fully noise quietening reception (about 75dB SINAD!) at around 1.2µV of input signal.

The receiver was built on a double-sided PCB (etched by a company in Ealing), and could be constructed either with or without a little 8-pin prescaler (for UHF use), and the tracks on the input filter could act as Lecher Lines when it was used at UHF - with very small 10p trimmer capacitors to tune it precisely. At VHF (Band l or Band III), the input coils were air-spaced. The Local Oscillator had an option for a frequency doubler or tripler for use in Band III, whilst still using the simple 3-chip synthesiser.....

The beauty of the board was that it could be configured for almost any frequency - keeping costs right down - and at Band I it wasn't necessary to use trimmer capacitors in the front-end filters (before and after the input amplifier transistor) - I just used fixed value ceramic capacitors and squeezed and stretched the coils to tune them!

At some point soon, I'll dig out and scan the artwork (originally done with crepe tapes on acetate!), and put all the documentation up on here for anyone to use.

[sinus trouble]

Cheers Albert!

Your design would be great to see for reference!

As i mentioned the HC4059 is a pain to source these days! The 4046 and 4060 are not such a big deal!

Also there are plenty of prescalers available too!

I may attempt full SMT on this project? But that remains to be seen as yet?