Band 1 Channel 1 Rod Pentode modulator

[Murphyv310]

Hi.
It was last December that I lashed up a 405 line Band 1 Channel 1 vision modulator. Where has the time gone?
After thinning down my collection and having decided that I am no longer repairing Vintage Radio & TV for others It's now possible to have some "me" time, like us all we are not getting any younger.
I've now rebuilt the vision stage onto copper laminate board and have tested it.
As before I'm using a 1ZH42B dual grid rod pentode, the HT is 80v and the filament is running 1.12v from an Ni-Mh AA battery via a 1Ω series resistor.
The video input is fed to one control grid via a 4.7kΩ pot (set near max). The interesting thing is the result was a negative going picture. I didn't have this issue in December.......Then it came back to me that the grid needed to be driven positive between 0.7 & 1.0v, the result then flips positive.
The second grid has a crystal connected to it and the other end to the G2, which gets its HT via a 470μh choke, the second grid is bled to ground with a 1mΩ resistor.

There is more info on this video. https://youtu.be/CXAPpi9cFYI

Bias adjustment added since video.

  rod pentode .pdf (Size: 73.17 KB / Downloads: 15)

[Freya]

Would like to get one of these together soon, found the idea fascinating when you first made the project.

[Murphyv310]

Hi Stephen.
Indeed a really simple modulator.
I eventually lost patience with your modulator, used the Nano VNA to get the coils right and still nothing!!! One reason for getting back into the project was to test the crystals as they sit there waving at me even on my tester but work in this! Odd to say the least, mind you I've not made the sound modulator yet so who knows what the 41.5 MHz one will do.

[Murphyv310]

Hi.
More work done today and some changes to the circuitry.
The 45Mhz crystal is low and doesn't always start and the 41.5Mhz crystal is as dead as a dodo. So I have found a 44.3Mhz crystal for Vision and a 40.815Mhz crystal for sound, both a bit low for channel 1 and of course not exactly a 3.5Mhz spacing but it works extremely well.
It has been pointed out to me on post 1 I have quoted the wrong valves they are 1ZH37B NOT 1ZH42B (there is no such valve as a 42B anyway its a 1ZH42A) Just to clear up any confusion.
I have contacted Vincent Jakomin re 45 & 41.5Mhz crystals.
I have also tidied up the bias arrangements.
I'll upload a revised diagram in due course.
Two further Videos here: https://youtu.be/CiT2tFXA-4M
https://youtu.be/7rQrSisX10E

[Mike Watterson]

A centre tapped choke/coil between g1a & g1b fed at centre with video (or audio) and crystal to g1a or gib. The best for signal linearity is signal is common mode (equal on g1a and g1b) and differential (or feed only one half of grid) with LO
Or feed g1a OR g1b via cap and RFC. Tie g1a & g1b via small caps to crystal. This will give better Osc gain, but poorer input linearity
Make sure feed when oscillating doesn't overdrive crystal or it will drift with age and fail. Maybe drive crystal via G2 biased with a potential divider and RF tune the Anode, a tap on tuned Anode to drive crystal.

The 1j37b is not a dual grid in sense of regular valve, nor like dual gate FET. The g1a and g1b shorted (common mode signals) is g1 action. Any differential signal steers the beam and reduces the signal, because a g1 is a pair of plates parallel to the two beams, very close either side of the filamentary cathode. Even the dual grid Wunderlict valve is actually two interleaved grids, so doesn't work like this. The 7360 beam switching modulator is more like the 1j37b in operation if you only used one anode, except the differential LO signal on the steering plates is fed to g1a and g1b, and the signal to 7360 signal grid should be equal on g1a AND g1b. Of course on any modulator (ring diode, dual gate fet, Wunderlich, pentode, hexode, octode etc) you can swap the LO and signal, but normally you want the more linear and sensitive connection for signal, and the less linear or switching action can be done with a higher amplitude LO.
https://www.radiomuseum.org/tubes/tube_7360.html
https://www.radiomuseum.org/tubes/tube_wunderlich.html Read Joe's explanation. Unlike the 1j37b and 1j42a it doesn't easily act as a frequency doubler because it's really two grids. The 1j37a g1a and g1b are NOT two grids. They are a single grid to common mode signals and like the X or Y plates in a CRT (or modulation plates on 7360) for differential signals. So the differential g1a and g1b will frequency double as well as modulate the common mode g1.

When the theory of operation is fully understood, the best performance is achieved. See also ring diode mixer theory, where the best performance is when the LO is making it act like a pair of change-over switches.

Best stability always uses a separate oscillator. A single device mixer-oscillator always has an FM component in the output due to the input signal changing the oscillator operating point.

Working isn't the same as reliable or best performance. One signal must be equal on g1a and g1b and the other must be differential.

[Murphyv310]

Hi.
Michael, the whole reason of this project is to make a usable modulator with the minimum of components, easily reproducible and simple enough for those that have limited skills and equipment.
I challenge you to make one in the way you think is better. I also remember you criticising my simple MW pantry transmitter using a single 1ZH42A, this has been running for years with no problems, a single PP3 has been its HT source for as long as its been built.
Perhaps you can explain how the phase reversal occurs.

Also running the G2 via a potential divider doesn't work, yes on low frequencies it does but above around 20Mhz the Crystal's don't start, I have found the G2 needs to be close to the anode volts for reliable starting.
I sincerely hope others try this simple approach, perhaps not what you think is the correct way but the proof of the pudding is in the eating, I await your modulator, its stability, reliability, and performance.

[Murphyv310]

As an aside.
I'm not criticising what Michael has said, I also realise there could be better methods, like fixing the roads there are many ways, if this works which it does and what I want from it then why go the complex route?

[Mike Watterson]

Yes, it works and I don't deny it. But you might as well use an 11mA 1j24b as a 1j37b at 60mA, because you are wiring the 1j37b as if it's a dual grid equivalent of a dual gate FET. A single npn transistor and crystal (or resonator) can also work as osc and modulator. I have one. Your 1j37b circuit will work with a 1j24b because you are not actually utilising the unique operating mode of the 1j37b. Re-read the theory.

I'm explaining the theory of the 1j37b & 1j42a used as a modulator, detector, doubler or mixer. That's the starting point, rather than a circuit that happens to work OK.

I'm not suggesting it needs to be more complex, but if simple is wanted, then use a triodised 1j24b as mixer osc*, or if in Pentode connection, drive g2 or g3 with video and osc is Anode to g1. There is no point to a 1j37b at all unless one signal is common to g1a and g1b and the other signal (osc feedback) is on g1a OR g1b, or differential drive to g1a and g1b.
[* see ANY EF80 based VHF tuner, the 1st EF80 just buffers the aerial, video or audio drive replaces that on a modulator-oscillator. Or DC90 based VHF tuner head (first Philips battery VHF in Germany) or Triodised DF97 VHF front end on almost all battery VHF valve sets]

Ability to startup and running level are two different issues with crystal oscillators. So if doubling component count is not an issue you can use either a 1j37b as mixer and 1j24b as osc, or use 2 x 1j37b (one with g1b& g1a tied as osc) in series filament, or 2 x 1j24b using g2 of one as LO in and the other a triodised osc. 11mA off a 5V supply via ballast resistor.

The 1j37b is filament hungry (60mA vs 11 mA) niche part that only gives an advantage over a 1j24b if used as intended (g1a + g1b = signal and LO ONLY to one of g1a & g1b, or differential)

[Murphyv310]

Funny thing is the source of the Crystal's is a modulator from Freya that didn't work. I've replaced transistors and wound coils that were missing. The Crystal's didn't oscillate in it nor in a counter that tests crystals but the 45Mhz does work in my project but a bit lazy. The 41.5 Mhz one is dead as a dodo so we at least know now that the crystals are the issue.

[Murphyv310]

Vince has two older crystals that are at a good price so I've ordered them.
I may eventually try to modulate the 1ZH24B with Video and see how it performs in comparison with the 1ZH37B.
Perhaps in a week or two.