Valve frequency multipliers

Mine oscillate, sometimes at lots of different frequencies.

Spice usually works well with its lumped part models. Leads and wire bonds and PCB traces usually hide the subtleties.

Spice can model thermal systems too, but badly. Thermals really need

It's usually easier to build a thermal thing and play with it, than to simulate it with some expensive software. I guess if I was designing a rocket engine or something I'd go to the trouble to simulate it.

Simultaneously!?!

Spice is really good at operating outside normal ranges.

Some might call that an advantage.

RL

Q: What is the difference between squegging and a blocking oscillator?

A: A blocking oscillator is when you want it to happen.

Sure. Usually some GHz burst thing on top of my desired 50 MHz sine wave. Adds jitter. Bypassing and trace routing matter, so it sometimes surprises us. I've given up on phemt Colpitts oscillators because they are too unpredictable that way.

Blocking oscillators used to be popular tube-digital things, like for frequency division and such.

Phantastrons were fun too.

You were right: with a crystal frequency of 16.656 Mc/s and two triplers the results are much better.

New block diagram:

New oscillator circuit:

The stages are: EF91 Reactance, EF91 Oscillator, ECC91 Tripler+ tripler.

The new crystal necessitated changes to the component values in the 'pulling' circuits, but the basic concept remain unchanged. Over large frequency swings the reactance valve characteristic appears curved but with 0dBm audio input at 400 c/s there is no audible distortion and the frequency swing is around 15 Kc/s pk/pk, which is three times larger than the permitted deviation in the amateur bands.

I tried putting a parallel-tuned circuit in the anode circuit of the oscillator, to resonate at the third harmonic, but there was very little signal, so I decided to use one triode of the ECC91 as the first tripler. I then resonated the anode tuned circuit of the oscillator at crystal frequency but discovered that this made it unreliable at start-up. Eventually I found that just an aperiodic [untuned] anode choke gave plenty of drive to the grid of the first tripler and allowed the oscillator to start reliably.

The parallel-tuned circuit of the first tripler and series-tuned circuit between the second tripler and the output cable both have comfortably gentle tuning, which is an advantage because they should hold their settings without adjustment for a long time. The output voltage is only around 300mV rms into 75 ohms, but this can be stepped up when it gets to the mixer grids by a further resonant circuit.

Are you sure that the crystal lets you pulll it by 11 kHz and add angle modulation to it?

Please design and verify the 144 MHz transmit and receive filters associated with the mixers for enough attenuation on the 150 MHz band.

It pulls on receive and is modulated on transmit, the two never happen together. According to a frequency meter I am getting the correct shift on receive and the correct centre frequency on transmit. The disadvantage of pulling a crystal is that it becomes less stable the further you pull it. This is less important when receiving, which is where the largest 'pull' occurs.

As a modulation test, I connected the output of the multipliers to a length of wire to act as a transmitting aerial and fed 400 c/s at 0dBm into the audio input. I then tuned an Eddystone 770R to 150 Mc/s and connected it to a panoramic display unit, Eddystone EP14, which can be adjusted so that the sidebands cover a marked width of the display screen. My accurate signal generator doesn't go up as far as 150 Mc/s. so I set it to 75 Mc/s and used the second harmonic. I adjusted the F.M. modulation until its sidebands covered the same width as the VXO signal, then read-off the deviation from the signal generator's scale. Because of the effect of frequency doubling, the actual deviation was twice the scale reading.

I am nowhere near ready to make the power section of the transmitting chain, but when I do, I shall be very careful to check it for spurious emissions before using it with an aerial.

You seem to have got all your ducks in a row.

Tripling within the oscillator seems to have been pretty popular back in the day. I'm not sure what kind of stress they were putting on the crystal to do this, but they wanted >5V from a substituted VCO to get similar downstream responses.

Unable to find RAH'69 on the web, I've copied two related articles:

If ever you do fool with LTspice, the oscillating HV PSU file should run 'out of the box'. Tube models were typed into the schematic to get around software revisions and modeler's individual file handling pecadillos. Schematic legibility obviously suffers as a result:

I'm beginning to have my doubts now: a search receiver suggests that something funny is going on at about 200 Mc/s, but I can't work out which stage is generating it.

I think you have to make the oscillator work very hard to get much third harmonic and most crystals in amateur gear were around 7 Mc/s or lower. At16 Mc/s things begin to get much more difficult. My main concern was getting the crystal to respond reliably to the pulling and frequency modulation, even if it meant losing the tripler effect.