I've been thinking about putting together a PLL design using the most basic components possible, or at least reducing the 'difficult' component count. One of the main difficulties is a pre-scalar to get the oscillator frequency low enough to feed into CMOS divide-by-N counter. In the past I've used a feed from a half-frequency oscillator (around 50 MHz) directly into a 74HC4020 but these chips are not easy to get hold of in your average electronics shop (not that many such shops exist any more).

I wondered about using an injection-locked oscillator as a pre-scalar. These are basically oscillators which free-run at a specific frequency but are then 'forced into resonance' by having another frequency injected into them. For example, a free-running oscillator at 10 MHz injected with a 100 MHz signal will 'lock' to exactly one tenth of the injected signal frequency. The difficulty is that the lock range tends to be very small. Move the injected frequency to 90 MHz and the oscillator no longer 'locks' to 9 MHz as would be hoped, but instead goes off and does its own thing back at 10 MHz, or divides by 9 and locks back at 10 MHz! If you don't mind tweaking something (e.g. a variable capacitor or inductor) to get lock, this idea works well. However, for a 'no-tune' design it's a fail.

I just wondered if anyone else had tried anything similar and had any experience with them?

I have had this project on my table. Not even started it.

Old Free Radio Magazine pll.

PCBs for that.

This thing is old, but what I have heard, works.

Guys have uset it with Stentor and Super Stentor.

The 4059 'divide by N' chip is not the easiest to get hold of. A divider made of a counter (e.g. 4020 or 4040) and a set of diodes might require a few extra components but the parts to repair it can be found very easily. The 74S112 is about the same level of 'rareness' as a 74HC4020 but does at least clock at 100 MHz.

As Albert has often pointed out, the original Veronica PLL was designed to use as many 'off the shelf' components as possible - I was kind of hoping to make an even more shelf-based design.

I like the injection locked oscillator in a perverse way but (if the goal is a kit) lots of parts and doesn't sound that easy for your average kit builder to get working? The likes of Stephen Moss considered the reliability of the design and ease of construction/alignment/setting up as important as easy sourcing of the parts.

The NRG design used a 74ALS74 as a "prescaler", it could be clocked at 55MHz and you'd then end up with half that into the divider. This was ages ago now but for an old PLL design of my own I took a leaf out of his book and used the 74F74, which is even faster. Looks like you can still get this from RS (along with the 74HC4020 actually but I get your point about smaller electronics shops and agree there aren't many left of what's left of our high streets).

In the 80's the 74S series was a life saver, could clock to 100MHz and a bit beyond that. But they were obsoleted fast because they are VERY power hungry. Mostly replaced by the 74F series.
I'm aware there are newer and faster logic families, that work only at 3.3V. I have an 74lvc74, if I remember correctly, that clocks to 250MHz.

The 74F74 will clock at 100 MHz and isn't too difficult to get hold of, so I suppose it could be used as an initial pre-scalar, though it could only manage to divide by 4 at best (to 25 MHz). A higher initial division ratio would be more useful (at least 16, say). There's the 74F193 which can divide by 16 but these also fall into the 'too rare' category. Maybe I'm over thinking it. I've managed to design the phase detector and the lock detector using one 4013, one op-amp, a few diodes and a couple of transistors. This, or simply use a 4046... hmm...

74ac74 goes to 125mhz at 5v and is still manufactured in dip package

reverend wrote: ↑Mon Jan 06, 2025 2:32 pm I've been thinking about putting together a PLL design using the most basic components possible, or at least reducing the 'difficult' component count. One of the main difficulties is a pre-scalar to get the oscillator frequency low enough to feed into CMOS divide-by-N counter. In the past I've used a feed from a half-frequency oscillator (around 50 MHz) directly into a 74HC4020 but these chips are not easy to get hold of in your average electronics shop (not that many such shops exist any more).

I wondered about using an injection-locked oscillator as a pre-scalar. These are basically oscillators which free-run at a specific frequency but are then 'forced into resonance' by having another frequency injected into them. For example, a free-running oscillator at 10 MHz injected with a 100 MHz signal will 'lock' to exactly one tenth of the injected signal frequency. The difficulty is that the lock range tends to be very small. Move the injected frequency to 90 MHz and the oscillator no longer 'locks' to 9 MHz as would be hoped, but instead goes off and does its own thing back at 10 MHz, or divides by 9 and locks back at 10 MHz! If you don't mind tweaking something (e.g. a variable capacitor or inductor) to get lock, this idea works well. However, for a 'no-tune' design it's a fail.

I just wondered if anyone else had tried anything similar and had any experience with them?

Some of us "old-timers" used approaches like that way back in the 70s. It's generally more trouble than it's worth, and over-modulation always tends to make them jump out of lock!

One cheap approach is the "huff & puff" stabiliser, which can be built with just a couple of logic ICs, but it's always a hassle to keep locked.

Another (ancient) approach is to use cascaded bistables made out of cheap pairs of transistors. The dividers have to be well screened from the rest of the circuit, but they allow the use of tthe cheap "CB" synthesiser ICs (I used to favour the PLL02). A related approach used the harmonic of the reference crystal in the CB synthesiser filtered and mixed with the lower frequency carrier generated under the control of the synth chip....

Yet another technique was to mix a crystal close to the required carrier frequency with a low frequency FM-able VFO. This required careful filtering and alignment, but was capable of good results.

The final way was to build a low frequency PLL with easy-to-find ICs, and multiply the carrier up to where it's wanted. Some of my earliest PLL rigs multiplied by up to 9 times (a tripler-tripler) to get to the output frequency.

To be honest, 74F74 and 74F160 are still easily available from the bigger component suppliers, and both work to over 125MHz. The '74 will divide by two or four, and the '160 divides by 10. At 10 MHz, you can use conventional CMOS (or a CB synthesiser IC). I used to use a '160 into an MC146106 with a 10.240MHz reference rock for the simplest two-chip PLL I could come up with.

The Far Eastern component houses have prescalers like the LB3500 for pennies....