1П24Б 1p24b

[Mike Watterson]

They're pentodes.

They can behave like Pentodes. They are beam type devices. This means that using a resistor to set the g2 voltage is poor. The actual designs thus used either separate supply rails (such as taps on the HT batteries) or resistor potential dividers.

All the devices are a sharp cut off, no variable mu possible as that is achieved by varying the grid wire spacing. There are no grid wires at all. They also can ONLY work with a thin filamentary cathode, "indirect" heated versions are technically impossible.
The G1 is a pair of plates either side of the filament, almost gammatron type action. Then two wedges of electrons are attracted from the filament to the two anode plates at opposite sides. The g2 and g3 are pairs of rods, acting as electrostatic lenses.
The 1j29b and 1p24b and other models with a centre tapped filament are actually two "valves" in parallel, so a 3rd grid plate rather than two is used. After that the construction is similar to the "regular" rod Pentode. It's more like a wedge CRT than a Pentode.

Note that some heptode and octode battery valves use a pair of plates or rods for the "grid" that's normally the oscillator "effective" anode.

The centre tap is thus the common of the two springs at the top that tension the filament. You can actually see the spring move on some models.

Triodising:
When triodising a normal pentode it's possible to still use g3 to 0V. The triodised Rod tubes typically have g2, g3 and anode tied, the models with an internal connection between g3 and one end of the filament are not so suitable.

[Murphyv310]

Hi Michael.
Great info there on your post.
These valves amaze me no end. I've been able to get nearly 30 watts of RF into a dummy load at 280v and 60 ma on each output valve. It not been run for long at this level, about 20 mins tops. Bulb temperature is around 130c. Last evening when fiddling around the meter went to 50 watts and both started glowing. The issue was a 10n decoupler on the G2 that went O/C. The valves seem to have survived okay.

[Nick]

These tubes are tough - the so-called "missile tubes" represent the ultimate in rugged design at the end of the tube era - some were used in proximity fuses for shells...

e.g. the CK5840 has a shock rating of 450g and a continuous rating of 1000g (tested in a centrifuge) and I've seen other types as high as 2,500g.

[Murphyv310]

Thanks Nick
I looked up the specs for the CK5840 and wow the samples certainly were given a lot of abuse. I'm not deliberately trying to make them fail but really to prove they are an amazing way for us to use in projects and of course for new comers to experiment with valves that can run on lower and safer HT voltages.

What Michael Watterson, Joe Sousa and others have done is very commendable. The Russian Rod Pentodes are still cheap enough and available, hopefully after Covid-19 supplies may even be better.

[Mike Watterson]

Joe has done brilliant work.
There may be millions of some of the commonly used rod "pentodes". All military production and the factories had fixed annual quotas till the end of the of the USSR. No unemployment, in theory. So officials wouldn't cut back the quotas, because that would mean redeploying workers to the Gulags. The system was after all run by humans. A lot of stock would have been in Ukraine because they made a lot of the USSR aircraft. Not sure why there are lots in Bulgaria.

Don't forget to not just decouple g2, but use either a tapped supply or a potential divider. The gain varies so much with g2, that you can use it for AGC or DC control of power out, for 90V HT, the 1j24b uses about 20V to 50V for almost no gain to full gain. The voltage range varies with nominal anode bias and model of rod valve. I've actually one US radio where the "RF Gain" pot actually varies the HT on g2 of the two IF amps, the 1T4 (DF91). You can use the deflector pin drive of a magic eye to derive such high level positive going AGC from the normal negative detector.

Also you can vary RF or AF power out on any amp, class AB or class A or possibly class C using any of the rod tubes and a suitable voltage on g2.

The g2 is also a good input for a local oscillator to modulate. The only regular battery set I know doing that is the Vidor AM/FM model using a DF97 as 1st FM IF, but as mixer on AM. It uses a triodised DF97 as self mixer/oscillator on VHF (all the rod tubes can do that) and LO only on MW and LW. It's a good template to use to make a complete Rod Tube AM/FM radio, you just need to change the g2 resistors to potential dividers with a second resistor. All the same values will work. The DAF96 can be replaced with 1j24b and an 1N60 (1st released in 1946) and the FM detector in all battery valve radios is 2 x diodes like 1N60 anyway.
You can use all 1j24b, or do the push pull audio out (then a 110-0-110 mains transformer can be used) with 1j24b, 1j29b or 1p24b, with option to have economy mode on the 1j29b or 1p24b by turning off half the filament. You can design for 2.4V nominal LT, more series than that is difficult. With it being push pull, you use the valves in parallel filament and adjust the bias for the valve that's more positive. You need to DC filter the local oscillator for -ve bias as per the Philips Colette, as the trick of series resistor between LT- and HT- loses too much HT at Class AB. The 1j29b may need a PP3 or 3 x lithium coin cells for Class AB bias and the 1p24b may need more negative g1 bias if you go for 120V to 150V HT for serious audio power. You'd want AA cells for the HT, with a fuse, for more than 300mW audio!

[Kalee20]

All the devices are a sharp cut off, no variable mu possible as that is achieved by varying the grid wire spacing. There are no grid wires at all. They also can ONLY work with a thin filamentary cathode, "indirect" heated versions are technically impossible.

The G1 is a pair of plates either side of the filament, almost gammatron type action. Then two wedges of electrons are attracted from the filament to the two anode plates at opposite sides.

Now that's given me some thought!

The first point, no variable-μ being possible: what happens if the grid rods are slightly curved so their spacing varies along their length? I'm thinking, where the rods are close together will be easier to cut off. Then the remaining regions will be still active, but less responsive to g1 voltage changes owing to the wider spacing.

So, you've made a vari-μ device!

The second point - indirectly-heated being technically impossible: Granted that a point (or line) source of emission is what's needed. But the rod pentodes don't have a line of zero thickness. They have a filament, maybe 30μm diameter. Now 30μ cannot be a 'hard' limit, such that 30μm works but 31μm won't, so just 'grow' the filament. What happens at 50μm? 100μm? 250μm? You'll reach a point where it is possible to drill a hole through the length of the filament, and thread an insulated heater wire through.

There may be no call for doing so, and it may be difficult, But at the moment, I can't accept it to be technically impossible.

[Mike Watterson]

They are electrostatic lenses. A curved rod won't work apart from the huge assembly difficulties.
A insulated indirectly cathode is HUGE compared to a filament. About 500 um.
Joe Sousa writes: "For example, in the case of the 1Zh37B the distance from the 0.07mm (70micron!) thick coated filament to the two grid rods on either side, is only 0.12mm (120micron!)."
Re-read these two links and look at images:
https://www.radiomuseum.org/forum/russia..._id=200277
https://www.radiomuseum.org/forum/summar...todes.html
Download the translated Russian article http://www.radiomuseum.org/forumdata/upl...1_2011.pdf

A conventional variable mu or indirect version isn't possible. The USA tried making indirectly heated "rod tubes" in late 1930s due to the g2 working well as a pair of rods instead of a wound wire grid in some heptodes / pentagrid converters, the g2 being effectively the "anode" for the oscillator part and the electron source for the rest of the valve. Also RCA producing the "Kinkless Tetrode" AKA Beam Tetrode, the alternate to the Philips Pentode by replacing the wound wire g3 with a pair of beam forming plates (the same electrostatic lens function as the two pairs of rods labelled g3 in an Rod Tube). It didn't work, though many Noval Pentodes and Beam Tetrodes did develop into oval shaped grids with all the rods in a line and two anodes at either side under the outer shield.
See especially the "tear down" of  a real indirectly heated Pentode, the EF89
The cathode tube is HUGE!

https://www.radiomuseum.org/forum/ef89_anatomy.html#1






The Russians were developing these over a 20 year period and made them up till 1990. The 1j42a and 1j24b were the most advanced devices. The 1j18b and 1j37a are earlier but not the first. The 1p24b, 1j29b are incremental improvements on earlier doubled types (two filaments and three grid plates). There are essentially three generations of these. The Russians either used g2 voltage and/or a 1j37b for agc. The gain varies so much with g2 it has to be from either a LO, a fixed supply or a pair of resistors as a potential divider.

Anyone though is welcome to set up a lab and try. However both variable mu and indirect cathodes have been discussed by valve experts and they conclude both are impossible.
The DM70 / DM71 DOES have a slanted element like you suggest so that it varies the length of the bar.
However it's a very simple device with poor gain. It's also totally brilliant!
https://www.radiomuseum.org/forum/dm70_operation.html#1

[Mike Watterson]

No doubt if one had a lab and all the up to date gear and theory you might make interesting "Rod Tubes" the Russians didn't make. What about twin anodes side by side at both ends of the beam and having access to the pair of g3 or g2 rods on both sides?
See https://www.radiomuseum.org/tubes/tube_7360.html

[Mike Watterson]

I've not found yet the Russian circuit for the 1j37b or 1i42a
Here is the 1j42 construction, which uses much closer spacing than the other Rod Tubes and it's the only one known with the triangular anodes:
https://www.radiomuseum.org/forum/142_ph...ves.html#1

This explains why the 1j37b was used for AGC
https://www.radiomuseum.org/forum/russia...bes.html#4
"As illustrated in the earlier posts, the control grid rods, when driven together as one control grid, give a very pure square law control of Anode current by control grid voltage.

When one of the control grid rods is biased at a fixed DC voltage, the other rod can be used for transconductance control over a spectacular range exceeding 7 decades!

A DC sweep of this control rod from 0V to -20V, will change the Anode current from 1mA to 100pA in a very smooth logarithmic progression. The screen voltage for this sweep was 30V and the plate voltage was wired together with the suppressor grid in a tetrode configuration and biased at 10V."

Further down the page


However normally the RF in would be LO1 and LO2 pins and the Local Oscillator would be on RFIN pin, as this gives rejection to out of IF band RF or IF band at the RF input.

Normally g3 goes to 0V. This is example is suggested by the idea of the LO being 19KHz pilot and the "RF IN" being the 38kHz L-R DSBSC multiplex.

[Kalee20]

Thanks for the detailed answer, Mike!

I'd read some of Joe Sousa's notes some time ago, but thank you for refreshing me with the links.

Regarding the vari-mu aspect, yes I can see the practical difficulties of making them - though that's different from it being fundamentally impossible.

And making with an ID cathode, again, I can accept that everything would have to be scaled up. No longer 'miniature' with grid rods several millimetres in diameter - but possible in principle, it seems to me, if not, why not?

But I hadn't seen the notes on the 1j37b. That's my evening reading sorted!