Old Rig

[zulu53]

Albert,
I would appreciate if you could please share coil details on the transmitter included in this thread.

"Albert H
proppa neck!
proppa neck!
Re: Stable VFO with BF245
Post Sun Oct 20, 2019 10:47 pm

FETs are wonderful devices!

I don't like etched VFO coils for two reasons - they're microphonic and they have lower Q than a conventional coil. However - if your VFO is working well for you, that's great.

I used to use a dual-gate FET oscillator:
40wvfo.png
40wvfo.png (25.87 KiB) Viewed 176 times

[Albert H]

Which transmitter?

[yellowbeard]

viewtopic.php?t=2337
That one. I searched the BFW16A from it for divelment, one of the links that came up for it was at radiomuseum.org - you can still get them but I bet they are in the dark dusty corner of the component house...

[zulu53]

Albert
Please find the transmitter schematic.

[Albert H]

The oscillator coil was a yellow S18 (4½ turns). The oscillator drain coil was four or five turns of 28 swg enamelled copper wire through a ferrite bead (FX1115). The collector coil for the second stage was sometimes a tapped 9-turn 6 mm diameter 18 swg job, or two separate coils with a common "land" in between them for the 56p into the base of the third transistor.

The Collector supply RFC for the third transistor was the same as the oscillator drain choke, and the 10 turn coil was 6mm diameter 18 swg stretched to about 2cm long. The base coil going into the 2N6080 was 2 turns at 6mm dia stretched so that the gaps between the turns were the thickness of the 18 swg wire that it was wound from. The 2N6080 base to ground choke was either the same as the oscillator drain coil or a 0µ22H moulded choke (I had a boxful of them!). The collector coil for the '6080 was 6 turns at 8 mm diameter stretched to about 1 cm long.

The input coil to the SD1272 was a single 6 mm turn, but if I used a '6084, it would benefit from two turns at the lower end of the band. The final collector to supply coil was the same as the one for the driver stage, and the 2 turn output coil was 8 mm diameter 16 swg about 1 cm long. The two output filter coils were 5 turns at 6 mm diameter, stretched so that the turns were spaced roughly a wire diameter apart. The base damping resistor for the final had to be non-inductive (I used ancient carbon ones) and the output trimmers were either the clear "blue spot" types with 22p in parallel with them, or Dau 80p (grey) trimmers that were about 1 cm in diameter.

The BFW16 could be changed for a 2N2219 (at a pinch) or a 2N4427. Whichever you use, it'll probably need a push-on heatsink.

The oscillator circuit and its following transistor were housed in a little tinplate box, or a box made from PCB material, to shield it from the rest of the board. The output filter was also screened away from the rest of the board in an effort to keep things clean. The regulator for the oscillator supply could be a 78L09. The 10p, 56p 33p and 68p to ground on the gate of the FET osillator were either NPO ceramics or (preferably) silvered-mica parts (if you could find them). This went a long way to helping the stability of the oscillator. A few of them had their oscillator box above the main PCB with a piece of expanded polystyrene ceiling tile material between the box and the main board to thermally isolate the oscillator and stop it drifting when the PA heated up.

It wasn't the world's greatest rig, but it worked pretty well, didn't drift too much, was surprisingly clean and made 35 - 50 Watts when run from a car battery / CB supply. We made them on a single PCB (for the amplifier stages) and a separate oscillator / buffer board that was over one end of the main board. They were mostly built in folded aluminium boxes with a heatsink on top. The box was 32cm X 20cm X 15cm high, with a phono socket and pot on one edge for the mod input, and an SO239 socket at the opposite end for the RF out. Power went in through a grommet on one of the edges, and there was a power LED. A few of them had an extra hole near to the mod input with a multi-turn trimmer behind it, for fine adjustment of frequency. Most of them were spray-painted with either black or cobalt blue "Hammerite™" paint. Many of them were pop-rivetted shut!

If I come across one of them (I haven't seen one for several years), I'll get some photos.

[zulu53]

Thank you for taking the time to explain so thoroughly.

[Shedbuilt]

The 10p, 56p 33p and 68p to ground on the gate of the FET osillator were either NPO ceramics or (preferably) silvered-mica parts (if you could find them). This went a long way to helping the stability of the oscillator.

Interesting. All the silvered mica caps I used back in the day, were quite strongly PTC. I found if I used those, along with all the other PTC components, the result was horrendous downward drift with increasing temperature. I found a mix of NPO and NTC caps were generally needed to compensate the other PTC factors.

[Albert H]

The silvered mica I used back in the 70s and 80s were (generally) pretty close to NPO and slightly NTC, which would largely compensate for the PTC of the inductance and the semiconductors. I remember spending ridiculous periods of time with a hair dryer and a can of freezer spray, heating and cooling various oscillator designs to try to get them stable.

I remember the joy of discovering how to do PLLs, and the satisfying "thump" of an oscillator locking on to frequency!

Back in the days of 200kHz Droitwich, I used the off-air signal as the reference for my shortwave rigs - accurate to 13 digits. That struck me as a bit obsessive, and so I reverted to using cheap crystals (sometimes in an oven)! That obsession began because of the really nasty heterodynes that I used to get at night on mediumwave.

[1608cc]

Some time ago I tried to explore history of "Veronica" excitier. It was popular in Poland (and people were convinced that Veronica is norwegian) because of "pink.art" website about pirate radio from the turn of the millenium - still accesible via archive: LINK

obraz_2025-02-20_231134506.png

Eventually I came across Gateavisa (from wiki: "Norwegian: Street Newspaper, is a countercultural magazine. It was first produced by an anarchist collective in 1970 at Hjelmsgate 3 in Oslo"). It turned out that it contained a part dedicated to electronics. Nr. 4 from 1995 "Svadatron" and 1 from 2000 made public two schematics of FM transmitters, and the 1/00 was Veronica! Graphic design of pink website and gateavisa are almost the same. Also time of publication. So now I know that "Veronica" come to Poland from Norway. The next question is how it came to Norway from UK

I've attached PDF's of both in zip archive bellow - Svadatron and Veronica.

obraz_2025-02-20_230046672.png

obraz_2025-02-20_230144304.png

obraz_2025-02-20_230229177.png

[Albert H]

Stephen Moss (the originator of the circuit) exported lots of transmitters to various parts of the world. The design was quite simple, and very reliable, and if built neatly would give reasonably stable results and plenty of power from cheap components. It was difficult to align wrongly, and wouldn't burst into spurious oscillation if misaligned.

I had one example - free running - with components carefully chosen to minimise thermal drift, that would stay within 12 kHz of its set frequency over a 25° change in temperature! It's a remarkable circuit.

It's based on a circuit developed (with valves) in the 1930s by a French radio engineer Mesney. He used it to generate VHF signals from valves incapable of VHF frequencies on their own, but paired in this kind of push-pull circuit would get there!

In the early 70s, a variant of this oscillator was used in a general coverage receiver circuit published by the RSGB. They praised its purity and stability and that it was capable of resolving SSB signals for long periods without drift.

Stephen saw a version of the circuit, and wondered if he could replicate it with low power transistors for a "bug" circuit. He was amazed by the stability he achieved, and continued to experiment and refine the circuit.

In the PLL Pro boards it confers a number of advantages - as long as rigorous symmetry is retained. If it's unbalanced (like in the clueless A****F attempt), it makes plenty of spurs! The biggest advantage is that the actual RF generation is at half the required frequency, preventing the possibility of instability and hum caused by RF feedback from nearby higher power stages. It's an elegant solution - even if it does take up more board area than a simple oscillator!

[Frequent Lee]

Stephen Moss

Could you provide any information about the early limiters and stereo encoders that Stephen used to build? The first ones I ever bought from him in about 1989 were built on veroboard and housed in the old Altai CB power supply style enclosures. As far as I can remember they were all built with old style toco cans and capacitors and resistors etc and didn't have any ic's or crystals in the circuits.

The limiter used to have that pumping effect if it was driven really hard and didn't have any user controlled potentiometers etc to change the gain and so on.

At a time when every other pirate in Leeds wasn't using any form of limiter and were all in mono my station was head and shoulders above them and had modulation the same level as the ILR stations unlike the pirates that were either really quiet and humming in the audio or totally over modulated and farting when anything bassy was playing along with splattering the band about 500khz each way

Any info about them would be great

Thanks

[radium98]

Wow wonderful . by the way in the old era 10 years and more i built semi pcb of the vero 5W version and not complited .if i can find some pictures i will post here .
i would like to ask a question about a replacement of a tl072 , can i replace with tl082 or lm358 ?

[Shedbuilt]

Stephen Moss

Could you provide any information about the early limiters and stereo encoders that Stephen used to build? The first ones I ever bought from him in about 1989 were built on veroboard and housed in the old Altai CB power supply style enclosures. As far as I can remember they were all built with old style toco cans and capacitors and resistors etc and didn't have any ic's or crystals in the circuits.

The limiter used to have that pumping effect if it was driven really hard and didn't have any user controlled potentiometers etc to change the gain and so on.

At a time when every other pirate in Leeds wasn't using any form of limiter and were all in mono my station was head and shoulders above them and had modulation the same level as the ILR stations unlike the pirates that were either really quiet and humming in the audio or totally over modulated and farting when anything bassy was playing along with splattering the band about 500khz each way

Any info about them would be great

Thanks

The first encoder I saw of his, would have been in '92. He was using PCB by then, but probably the same design. As you said, no crystal, and no chips. It used a variable frequency oscillator, and analogue mixing. To my mind, analogue mixing is the "proper" way to produce a stereo signal.

[Shedbuilt]

Wow wonderful . by the way in the old era 10 years and more i built semi pcb of the vero 5W version and not complited .if i can find some pictures i will post here .
i would like to ask a question about a replacement of a tl072 , can i replace with tl082 or lm358 ?

TL072 and TL082 are produced as the same device, but selected based on specifications. The lower noise examples become TL072, the TL082 has lower offset specification. LM358 is also pin for pin equivalent.

[jvok]

LM358 isn't really great as an audio opamp, its designed for DC accuracy not low distortion. Theres a trick you can do to put a resistor from the output to VSS pin which knocks the distortion down a bit but honestly I wouldn't bother

TL082 (or TL062) will work fine

[Albert H]

Wow wonderful . by the way in the old era 10 years and more i built semi pcb of the vero 5W version and not complited .if i can find some pictures i will post here .
i would like to ask a question about a replacement of a tl072 , can i replace with tl082 or lm358 ?

TL072 and TL082 are produced as the same device, but selected based on specifications. The lower noise examples become TL072, the TL082 has lower offset specification. LM358 is also pin for pin equivalent.

The'358 isn't meant to be used for audio - it's an unusual op-amp in that its output can get really close to the rail voltages - I've often used them for DC amplification and filtering (for example) in loop filters for PLLs.

The '082 is just a (slightly) noisier '072, and in practical use you're unlikely to hear the difference.

[Albert H]

Lee - Stephen's earlist limiters were transistor-based, using FETs for the attenuators. He managed to get very fast attack, and a level-dependant release, which sounded pretty good. He added the Toko passive filters a bit later on. These were originally meant for tape recording, stripping off the 19 kHz stereo pilot when recording, to prevent beating with the tape "bias" signal, causing whistles to be recorded. He realised that their specification was ideal for broadcasting purposes, and they were relatively cheap (at the time).

The next iteration of the limiter used op-amps, and one run of them used the MC3340 attenuator IC, because this really simplified the circuitry.

The last range of them retained the Toko filters, used TL074 quad op-amps, and was (eventually) available in mono and stereo versions. At the time of his death, he and I were working on a new Limiter, based on PWM attenuators, and using active, op-amp based filters (using gyrators). I have the original circuit sketches somewhere here.

His first switching stereo coder - using a 4060, TL074, and a 4016 - didn't know its left from right - the audio inputs couldn't be labelled "Left" and "Right" because the coder would start up at random (just like the A****F one still does). I upgraded the circuit to ensure "Left" and "Right" inputs, and added Mono / Stereo switching (that enabled both channels simultaneously, and killied the pilot tone).

The final coder - the Pro IV - was my design. It still uses the same 4864kHz crystal to generate the various signals needed, but I gave it oversampling of the audio to minimise distortion and digital generation of the pilot tone, using a sequence of 16 steps giving a rough sinewave, cleaned up by subsequent filtering. The results from this relatively simple circuit were remarkably good - certainly better than anything else available on the DIY market at the time.

I have the next iteration of that design here somewhere. I doubled the number of oversampling steps, and added a second multiplexer at 114kHz (the third harmonic of the 38 kHz switcher), which was added (in vastly reduced quantity) to cancel the harmonic products, reducing the output filtering requirement, thereby enhancing the phase accuracy of the various components of the stereo multiplex signal. The spectral purity, noise, separation and distortion specifications of the output from this "Pro V" design is as good as the very best commercial designs, but retained the simplicity and ease of construction of the earlier versions.

[radium98]

Thanks.

[Shedbuilt]

The silvered mica I used back in the 70s and 80s were (generally) pretty close to NPO and slightly NTC, which would largely compensate for the PTC of the inductance and the semiconductors. I remember spending ridiculous periods of time with a hair dryer and a can of freezer spray, heating and cooling various oscillator designs to try to get them stable.

I remember the joy of discovering how to do PLLs, and the satisfying "thump" of an oscillator locking on to frequency!

Back in the days of 200kHz Droitwich, I used the off-air signal as the reference for my shortwave rigs - accurate to 13 digits. That struck me as a bit obsessive, and so I reverted to using cheap crystals (sometimes in an oven)! That obsession began because of the really nasty heterodynes that I used to get at night on mediumwave.

Haha, yes, I spent a ridiculous amount of time heat cycling rigs; monitor cold, warm it up, monitor again, change one or more capacitors, sometimes inductor core material etc, let it cool down, repeat. That included the times where we were offering PLL. Market pressure was such that the buyers were telling us what they were prepared to pay for a rig, which didn't cover much more than the cost of the components - especially if we had to pay full price for an output transistor and / or other major components. We decided to suck it up, offer PLL at extra cost, with a bit more margin, and try to sell PLL. You can guess how that went.
My experiments with silver mica caps in oscillators, were in the very early days. I'd read quite a lot in RSGB books, and "knew" that silvered mica caps would be "the best". They were larger, capacitance for capacitance, than the modern ones, and differently encapsulated. Didn't take long to start using the cheap ceramics, like everyone else !

[reverend]

I remember one TX that I made which drifted terribly as it warmed up - I've no idea what the problem was with that particular one. It was meant to be used on 103.9 but would fire up on 102.4 and only after half an hour or so of heating up would it reach the right frequency. It was pretty stable after that though. I did end up with a bit of a reputation for making rigs that worked best if you took the lid off and let some air flow around the insides.