Russian rod pentode circuits

[Murphyv310]

You'd be better using a suitable number of PP3 batteries in series during experimental developments. CPC do cheap zinc carbon ones for not a lot of cash, saves blowing up inverter types.

[Kalee20]

It's also well worth while putting a couple of series-connected 1N5402 diodes across the LT supply. If you short HT to filaments, these'll clamp the voltage to 1.5V or so. Which is above the operating voltage, but not so much as to fuse them instantly.

I'm using a load of CR2032 cells from Poundland. With a drain of a few mA, I haven't had any issues yet. You can get holders from eBay sellers.

Then you really can work with power on. (But I do have a little on/off slide switch. which breaks LT AND HT).

[BusyBee]

A good idea with PP3's as it would likely be easier than the 'lash up' I often have at present. It would be an idea to wire them together in a box with just 2 leads. I do have several 12V 'gel' batteries so could use those, although would need fusing because of potential current flow if shorted. I can see the possibility of HT on the heaters so I will give it some thought (bearing in mind that the diodes would effectively short HT) The convertors are cheap so not too big an issue, but fused gel batteries would sort both issues.. The past few day's experiments have given me food for thought anyway, and as I have learned quite a bit in the time I have been using these valves I may change my approach (at least try).

The blown module problem came along after I decided to tidy things up and mount the anode tuned circuit on my module board. At that point things were much tidier but the oscillator failed to work. Any subsequent lash up of wiring brought it back into life (sporadically). After twenty minutes or so of tinkering, the one thought that came to me was that I had no shielding on the valve, and I new from previous things I had read that lack of shielding invited problems (Joe Sousa and Mike). Yesterday I bought some tinplate and today made a shield. My oscillator module is now outputting stabily on the 7th Harmonic. I have since checked and the nearest station of any note seems to be a Spanish one around 5kHz higher.

I have made rough checks for a buffer amplifier, but found similar stability issues with the amplifiers (funnily less so with the oscillator in it's initial state), so now plan to fit shielding to all valves and fully fit tuned circuits before testing, hence hopefully getting more reliable results. With far less croc leads the prospect of a wrong connection should be far less likely. I will do this to them all over the next few days so they are more ready for tests.

See attached photo of oscillator board. The metal screen is close, but I have fitted internal rubber strips, the intention of which to dampen any mechanical vibration of the screen which may cause unwanted modulation of signals.

[Murphyv310]

May I ask the reasoning for using the 7th harmonic? Do you not have a crystal of the fundamental frequency you wish to use?

[BusyBee]

Yes it may seem a little strange to be going a more complicated route, but there are several reasons which influence me going this way (in no particular order and probably not exclusive):

1: I have not got a great many crystals, especially around these frequencies. I know they are readily available but other reasons deter me from going this way. Most of those I have are on fixed MHz points, except for a few oddities.
2: I have a number of these glass low frequency crystals. I tend to collect them. They have a more vintage look about them .
3: Part of the reason for doing this project is not just to get a useable pantry Tx, but to learn quite a bit about different types of circuitry.  I feel it more sensible to do this at low powers particularly with these valves than at a possible future higher level. Multipliers and harmonic issues, to me, are a big thing to learn. It is Ok in theory, but my brain tends to wander, multiplying and adding frequencies almost like a mathematical puzzle to get to a particular (at the time) frequency in mind. Practical experience will help me to understand what is reasonably possible and how to do it. To some extent I have a mathematical mind, and have realised it is possible to design with mathematics and implement the result. It might be a bit 'nerdy' but it's powerful and I have found, in at least one occasion, almost the only possible way.
4: Perhaps less of a reason, and maybe one against, is that I am thinking about microcontroller timer generation of squarewave frequencies and filtering the squarewave to extract, principally the fundemental, but possibly a higher harmonic.  I know there are chips specifically for this but I do tinker with programming as well. Although not strictly crystal controlled maybe an Arduino Nano with suitable output filtering would be more useful as programmable for a wide range of frequencies, especially when the lower (higher amplitude) harmonics are taken into account?
5: With the components I am using the 7th harmonic is within the range of tuning and is also within the MW band, from my initial checks, at a quiet spot. I could obtain a crystal of this or nearby frequency and probably avoid the need for a buffer but I am happy at the moment with progress. I have other experiments I have tried which, if implemented and successful, may be interesting.

[Murphyv310]

Hi.
All very good reasons indeed.
Thank you for explaining. I certainly am enjoying your input and the way you are doing things. Like you I like experimenting with these little valves and finding out what can be achieved.
Good luck.

[BusyBee]

It's about time I updated on my progress.

I had found that using more than one of my little modules with a 'lashup' of wiring got very complicated. With this and also feeling that I had generally sorted my circuit power supply I decided it was time to use my most usual approach to design of larger projects. In this I decide how much space is likely to be taken up by circuitry (often visualising), then obtain a suitable box / enclosure. The circuits, as I produce them, are then fitted into place. This often means working within limits, but, usually, it works.

With the pantry Tx design I bought a plywood box from The Works (£4). I have varnished it internally, and will decorate the outside when the internals are sorted.

Apart from the power supplies, I have done little beyond sorting a buffer amplifier for the RF oscillator. One of the photos shown is of the unit with output of the buffer amplifier shown on the oscilloscope. The oscillator is fixed in position, but the buffer is not fully accepted at present. It uses 1J29b.

The other photograph shows the box containing the oscillator, plus 3 other modules, which shows the layout I am thinking of. Final details will be sorted as time goes on, as will any additional shielding etc, needed.

Looking at the box, apart from the valve modules, to the left of them are the upper boost converter and negative bias battery packs (NiMh for convertor and standard alkaline for negative supply). To the front of these is the on/off switching and audio I/P socket box (contains fuses). Under that is the convertor box. To the right of this is a 6V gel battery for heater and lower boost convertor box, There is a board in front of this for distribution and negative bias setting for valves which require it. To the right of this is the 1.24V heater regulator board (with heatsink). Fine setting of 1.1V is done on each valve module.

My next task will be to sort audio input / amplifier for audio drive.

I have yet to fully decide on modulation. There are specific ways as mentioned here for rod pentodes, plus there are recomendations for 'standard' techniques. I have been reading around and found a few pointers.

The only modulator using the 1J37b I have found is one by Joe Sousa, which appears to use the method described by Mike, but the valve is used in Gammatron mode with anode being g1b. I don't intend to go this way as this, although maybe efficient, is 'off at a tangent' from common design. I have noticed a similarity of feeding grids differentially with 1J37b to similar looking circuity for VHF double tetrode circuitry like with QQV03-10. It seems to deserve a closer look, later if not now.

Traditionally (just from what I have read so I am not 100% on this), combined anode / screen modulation appears the most efficient, but needing high AF drive, and grid modulation, rather less efficient but likely less issues of distortion and also lower drive needed. With the intended application, high output power is not required (and would be bad), but audio quality and RF purity is. With grid modulation the valve impedance varies so matching is not as precise, but at the low power envisaged this may be less of an issue. The choice appears to me to be between simple grid modulation and the differential application described. The screen modulation described being similar to grid modulation. The differential approach may well be far better, and easier, at higher frequencies with easier wound coils etc. As I said, I have yet to finalise so anything may come to pass.

More to follow!

[Murphyv310]

Hi.
Very neat work indeed.
Well done. Nice to see people doing projects, something that is lacking these days.

[BusyBee]

Well after a few hiccups I have made further progress.

One of these was very straightforward to sort as my use of harmonics made things obvious. As soon as I fed the oscillator stage into the buffer amplifier and checked the output I found that the oscillator output frequency had jumped from the 7th to 9th harmonic. It didn't take long for me to realise that the oscilloscope probe capacity was to blame. Although I don't know its capacity I was able to roughly work out it's capacity to around 45pf (see attached maths). Addition of a 51pf silver mica capacitor across the tuned circuit sorted. The issue with the oscillator stage was the low output, requiring the probe on x1. Not an issue further down the line. My lack of experience showing here! (see attached calculations)

I have constructed more modules, 2 x 1j29b (as one required for audio stage) and one 1j37b. During  this process I decided to replace my original 1j37b. I think I may have flashed the heater somehow (probably when using original lashups) and destroyed its emission. It seems to work but low gain. It needs a thourough check so I replaced it. While doing so I found that it was connected wrongly, with g1b and g3 interchanged. I am not 100% on this, but if so, my original modulator experiment was using g1a for rf and g3 for audio.

Following this I sorted the audio stage then assembled things. As yet I am not certain on the final situation with the modulator, but I need to get something together for test purposes, then steadily work toward optimum (at least as required for this application). Shown are photos of output waveform (input audio 1kHz sinewave from a signal injector to input socket) and the current working circuitry. The valve layout (from left to right) are oscillator (1j24b), rf buffer (1j29b), modulator / rf output (1j37b) and audio amp (1j29b). I have only done a little testing and optimising of settings. With a short lead for aerial I have received a signal across the house, on frequency, as checked using my comms receiver (digital frequency readout).

I can see yet quite extensive work needed as yet, but am happy with progress so far.

[BusyBee]

Another busy couple of days. I had been thinking about what sort of aerial would work best with the  transmitter. I was wondering about using a frame aerial. I do have a large one I constructed (80cm square) which I use sometimes for reception. I visualised a much smaller example in the lid of my box.

While out last weekend I called into our local secondhand shop. They had a pack containing three embroidery hoops with silks for fiftty pence. I do embroidery at times (when I have time   ) so thought I could use one for an aerial frame, and the others for their intended purpose.

I took things a bit as they come, and wound 9 turns of 24SWG enamelled copper wire on an 180mm diameter inner wooden ring. This was about all that would easily fit (see photograph). With a bit of work using my dip meter I determined what I thought was about right and put together a tuned circuit comprising this ring, a 47uH inductor in series and 440pF plus what I think is an 100pF trimmer in parallel with the combined inductance. Later work leads me to believe that the original coil was about 50uH (checked from a coil calculator Coil32 – probably around 40uH). This gives around 100uH in parallel with around 550pF – around the right frequency) .

I coupled the constructed assembly to the anode output loosly via a 6.8pF capacitor. This effectively gives two tuned circuits loosly coupled.

My initial tests have shown that tuning is fairly sharp and output appears to be somewhat higher than with my short wire previously used. I still have quite a bit of work to do in checking and optomising (maybe even changing things), but it seems more logical to me to have a specific load (and indeed the intended one) for checks. As the transmitter is intended to be used on one frequency only I am hoping only minor adjustments will be required in use. The tuning is faily centralised on the trimmers and initial tests seem to show things are faily stable. I have got quite a bit to learn here so do not intend to pull things about too much unless I understand why I should.

Next checks – harmonic output and then try other than sinewave input. For theory, as I am not hot on aerial therory I will look at the coupled double tuned tuned circuits theory as well for issues with coupling. One thing I am learning is that when testing it is tricky not to detune things with probes etc.

Not much on the rod tubes this time, but I will descibe more fully my modules at some point.