Dawson PLL transmitter

[sinus trouble]

Good point Mr shed!
i didnt realise the oscillator section was running at half frequency!
Ive never used the tsa5511 but the saa1057 is very sensitive and needs only a very small sample signal to get it to kick in!
when i built my version i used a resistor on the feedback to PLL i tried many different values to get a compromies between output drive and PLL needs!

[sinus trouble]

Another good suggestion Mr Pjeva!
The saa1057 will operate direct from the output too!
Mr pjevas method will also add some decoupling avoiding shunting the output significantly!

[rettimsnarT]

The first time I ever saw PLL added to VFO circuit it was connected to the RF output.

[sinus trouble]

Yes Mr snart!
That does work but control of the oscillator where possible is the best option!
amplifier stages can cause complications with tuning!

[teckniqs]

Yeah I think the old Venture Electronics 1997 'frequency synth' had the RF output connected to it??

[rettimsnarT]

When PLL first came about it seems to be the old way it was being done. Oh how times have changed since the good old days. The John Dawson PLL was originally designed and built by Dave Martin and Steve May back around the middle of the 90s and John/Keith copied it and slightly changed the layout. They were very good for their day but for the price but it would cost more to build one these days than to buy one ready built on eBay.

[YellowFM]

Veronica (nrg pll)
I do not want to do because at the forum also had a lot of its deliberations , it creates a lot of noise in our TV range , about 50 MHz ..
There after generator consists of two separate generators on the 50 megahertz frequency is apparently one of them passes the same to the output, and amplifies the power amplifier , thereby creating interference tv

[Shedbuilt]

The Veronica / a****f boards are not quite the same as the older NRG boards. Both use the twin half frequency oscillators you mentioned, but some of the details of the implementation are different. Stephen Moss of NRG, paid a lot of attention to detail, and this included suppression of the 1/2f component. This will always be present, to some extent, and to some extent, depends on the matching of components (between the two oscillator circuits); this includes matching the values of mirrored components, transistor gain, keeping inductor lengths as close to identical as possible etc. The NRG oscillators are symmetrical by design. The Veronica / a****f are not quite as symmetrical, and the amplifying stages are different. On the whole, I would expect the NRG design (or designs); if carefully built, will quite likely perform better than the Veronica / a****f. Some 1/2f will get through, but perhaps less. I imagine the biggest problem, is probably with broadband amplifiers, which are probably only low pass filtered at the output. If they amplify the 1/2f, it will go straight through the low pass output filter. The 1/2f could be suppressed quite a bit by a simple lc trap filter. You could put one on the output of the driver board, and one on the output of the amplifier. This would need to be tuned (to the oscillator frequency), however, as they are fairly narrow-band, so you would need some means to monitor the strength of the ~50MHz signal. I can give you theoretical values of L and C if you like, with theoretical dimensions for L, but it will definitely need tweaking in the real world.
Going back to the design you were talking about in the other thread, the schematics for some of the BW "at frequency" driver boards are easily available on the internet, but without a track layout. If you have some experience of PCB design, and some knowledge of RF techniques, it shouldn't be too hard to design a PCB for the RF section (if you wanted to use that RF section with a different PLL design).

[thewoodstarr]

There is no software used, it's uses Counters. it is a half frequency doubler, it's a logic PLL. I do have the codes here, cos 20 years ago we used the same lock, ask me nicely. It's a terrible design, you need at least an Oscilloscope to set it up. Really I don't know why you would bother? It was Crap 20 years ago, it's still crap today!

[Albert H]

It wasn't a bad design as it was, but Keith decided to try and make it smaller, and tried to cut down on the parts used. The original circuit was a BDHK design, from the early 80s, but Myers used three counter ICs (like the old NRG) rather than the 4059.

My version of the synthesiser used a 4MHz crystal (cheap as chips), a 74HC4060 reference oscillator and divider, 74HC4024 as the "prescaler", dividing by 4, a 4046 as the phase comparator (and 76kHz phase-locked oscillator in one version!) and a diode-programmed 74HC4040 as the pre-settable divider. Back in the 80s, when I started building boards with this synthesiser, the whole part cost for it was around £1.50. I still (sometimes) use the same circuit today!

Other engineers were agonising over VFO stability, with Pyers buffering his VFO with three FET stages, Myers picking particular temperature-curve capacitors, Michael and others using dual-gate FETs and temperature compensation.... I was building a simple, clean, stable, reliable PLL that added almost nothing to the cost of the rig, and ensured that it stayed on frequency.

When these rigs were getting sold to "certain" people, the PLL and the VCO were cast in a resin block. This stopped rig thieves changing the frequency without putting a whole new exciter into the thing, and there were various other "counter-measures" to stop "re-use" of stolen gear, including a rather cunning electronic "lock" that would prevent the PA powering up unless a certain plug was in a particular socket at switch-on.

Keith's re-spin of the board, turning it into the infamous "Dawson" exciter, cut out a few crucial parts - including part of the loop filter and lots of the supply filtering. He just wanted as much power out of the 4427 as possible, with little concern for spectral purity, so the driver stage was hammered! Not great, but better than most of the cr@p around at the time!