Russian Military Subminature "pencil" valves

[murphyv310]

Hi.
I tried some little circuit changes to the 1 valve radio today, instead of connecting the valve as a triode I connected it as a pentode. In this circuit I returned the reaction winding to screen (G2) then to HT but across the winding I had fitted a 390pf cap & 2 6.8uh chokes this has caused a beneficial effect of a form of AGC in the running of the radio.
On weak signals it is just at the point of oscillation but with stronger signals the gain seems to reduce. Now I am running the valve at 54v, anything above 65v and the agc effect seems to disappear and below 45v the gain drop off dramatically and reaction stops.
I don't even need a reaction control, my local station come in quite loud on some old 500 ohm phones with the ferrite rod aerial but a few turns of wire round the rod and an aerial and earth pulls in a fair amount of stations at excellent volume.

[neon indicator]

Yes apart from 1j29b and 1p24b they do seem best between 35V to 60V. 42 to 55V a good choice. The Pentode wiring is best above 22V. "Triode" wiring on for crazy very low voltage (8V to 18V) experimenting.


I think any Russain Rod Pentode alternative to the 1950s BBC radio or Elektor Radio on other threads should maybe use 6 x PP3 as HT? What you think?

Though 16V to 18V is possible. 9V is plausible but very very poor performance.

I only get any MW at night

Maybe tomorrow I will have time to test a little Radio like yours but with 1j24b (11mA heater) and 1j18b, at 18, 27, 36, 45V, 50V

Today I was creating Eagle Schematic & PCB libraries for the Rod Pentode Family.


Close-up pictures showing structure of 1j24b, 1p24b (two views), 1j17b and 1j37b here https://www.vintage-radio.net/forum/...20&postcount=5

[neon indicator]

I think we might have to stop discussing this as vintage and put it under Modern Technology for another five years!

I got a new consignment of 12 x 1j24b today and they are all '91 date code! Very clean shiny "as new" leads. OK, so twenty years old ...

I tried a different Ukrainian supplier as the previous one had only the 4 x 1j24b. He included what looks like a factory supplied datasheet. I'll scan it and upload. "Typical" ratings on it are
1.2V Filament with current 11mA to 15mA (typical 13)
Typical Anode 60V
Typical G2 45V
Typical gm = 0.9 (min 0.6, max 1.2)
gm on some other situation 0.48 (I can't figure the Russian!)

Absolute Max Filament range 1.05 (0.95) to 1.32 (1.4) V [No idea why two sets of limits)
Absolute Max Anode 120
Absolute Max G2 90
Absolute Max Cathode Current (i.e. all grids & Anode added to Filament) = 1.6mA

Various other data not seen before, but I have no idea yet what it means.

Note that with HT of +45V, your peak anode voltage with high Q tuned circuit and RF could be 60V.

[murphyv310]

Thanks for the info.
These valves are tough! My destruction project turned boring with 250v applied to the anode and eventually 1.8v to the filament! It glows like a small bulb, gets quite hot and still works perfectly!
I have now pulled the plug on it and am building a little AF amp for my one valver with the hammered valve.
I might build a P-P output stage but use an interstage transformer for phase splitting just for the 20's theme!

[neon indicator]

I wonder for how long it would live? You haven't got a 2V filament Mains valve by mistake?

Have you ever done this to DF96, DL96 etc?

The 1j18b might do 150mW to 200mW in push Pull class AB at 50V HT.

The 1j29b might do 400mW in push Pull class AB at 90V HT.

The 1p24b might do 6W to 8W in push Pull class AB at 150V HT.

The limit is the RMS "cathode" current from G2 & Anode, not really the dissipation at anode. You can see the 1j29b uses larger area "flat" on the anode plate and the 1p24b much larger, it allegedly can cope with 1W on the G2, which is just some rods. So the limit really is the filament.

I'm wiring up a 1j24b right now. Also testing with "simulated" near flat batteries suggests lowest sensible supply is 2 x PP3, though a single 9V cell will "work", in a limited sense.

Does 6 x PP3 sound reasonable for a more "reliable" beginners 1 valve Radio? (End point about 42V, fresh 54V)

I think for battery economy a high impedance transformer for phase splitting rather than an extra valve is the way to go.

See post 69 for Transformers I'll be testing. These are as cheap as 50p in quantity on line (£2 in Maplin) with Altai/Eagle branding. Many other aliases.

[murphyv310]

I suppose I am lucky in having a nice box of 20's & 30's bits and bobs, hence the idea of an interstage transformer, there is no way anything I have that will be a correct match but really who cares, its just experiments! Its funny I even got the wife to listen to the one valve set today and she was honestly impressed with the quality.
I have really enjoyed the experimentation not at your level though as I am doing a resto on a pre-war TV at the moment so fiddling about is well down the list.

I personally think these valves are remarkably robust, far stronger than the B7G "D" series valves they would never stand the abuse I have done to those CCCP valves!
What I have also found out due to excellent structural stability there is little or no micro-phony.
I think I will definitely order some other types, if you can suggest a good supplier please send me a PM.
Thanks.

[neon indicator]

Attached is graphs of filament current for 1j17b and 1HH17b.

In the lighting of my workshop filament is only obviously glowing from 2V upwards! The 1HH17b filament same characteristics at 0.6V to 1.2V after being at 2V for 5 minutes. I have no idea what a 1HH17b is (it's marked 1HH17b). It could be old non-cyrilic marking for a 1j17b (marked 1Ж18Б). It was included "free" and marking is white ink. The other valves look like etched part numbers and date codes. The purple OTK & number is quality control number.

The old datasheets all show a narrower part on the envelope tube but only the older date code ('61 to '71) valves I have actually has this. At the top it's about 1mm narrower for 6mm. Most of the newer tubes (1j17b, 1j18b, 1j24b and 1j37b) later than '77 are straight.

The 1j29b is slightly longer (2mm?) and has a slight taper at top. It's 90V rated and an output or driver for RF/IF/AF, similar power rating to DL96. It's unclear if the two base versions of the 1j29b are because one is older (middle wire + 2 spaces) or because the newer one is newer or because it's the "r" option (all 7 wires on edge in "ring of 8 with a space")

The 1p24 is ribbed tube and about 2mm greater diameter and about 5mm longer than 1j17b, 1j18b, 1j24b or 1j37b.

All the tubes except the 1p24b have top "bell" of glass black with 2.5mm to 4mm of "getter" (shiny on inside). The 1p24b is completely clear. I have ten and all identical.

Photo includes both styles of 1j17b, 1j18b, 1j24b and 1j37b envelopes
(1j17b and 1j18b only shown)
Includes 1j29b and 1j29b-r (white socks). They are same diameter but 2mm longer.
And ribbed "clear" larger 1p24b 2W P.A.
12AU7 for comparison

[murphyv310]

Interesting filament V A graph. Certainly 1.8v here was producing a well lit filament. From the level of illumination of the filament at 1.8v (no series resistor) I doubt that it would stay intact at 2.5v!

[neon indicator]

I'd guess that's a 1j18b. It's about 1/2 filament current. Later valves definitely have a very faint pale grey (etched?) two sets of marking, one part number in cyrillic script (1Ж18Б = 1j18b)and the other a date code I think with this cyrillic letter Ө (barred O) I have only one with "traditional" white ink part number.

The OTK code is quality checker ID (usually highly visible purple ink and mistaken as part number!)
A symbol in a circle or diamond is Factory code.

[murphyv310]

Yes that's all I have 1j18b's

[neon indicator]

That was my 1st purchase as there are sample circuits on the web and it's cheapest and most common. It's fairly general purpose and moderately low filament of only 23mA. Only the 1j24b is lower filament (11mA) and it's also about 1/2 the power rating. I imagine the 1j24b is less robust too!

[neon indicator]

Here is different 1j24b datasheet I didn't find online

сетка = grid
анод = Anode
экран = Screen or Shield
катод = Cathode
накала = filament
минус = Minus / Negative
плюс = Plus / Positive
ток = Current
напряжение = Voltage
первый = first
второй = second
треть = third
управляющая сетка = Control Grid
экранная сетка = Screen Grid
подавление сетки = Suppression grid

First page is "Normal" Operating, Minimum, Typical and Maximum values.
Second Page is Absolute Limits (not recommended)

So on second page ток катод is Current Cathode = 1.6mA. This is essentially the sum of all electrode currents added to heating Filament current (LT) via HT. It's an absolute Maximum.

B = Voltage

[neon indicator]

The filament + is not always pin 1. The Bias for grid is wrong if you reverse + and -

So reading datasheets
накала (плюс) = Filament Plus
накала (минус) = Filament Minus

Note some "pinouts" have "missing" or cut wires in the layout so the wires are not numbered sequentially, i.e. on 1j18b it's 1,2,5 & 6 as there are cut wires at 3 & 4 also the 1p24b is numbered to 9 though there are only 7 wires on the base (its wires are laid out as if an outer circle of 10 and a row of three in the middle, only 4 wires on outer circle.

[neon indicator]

I "blew up" a 1j24b. Somehow got +20V on the filament, at least I think that was what happened, though I was convinced I had only mis-wired G1, G2 & G3. Nice bright white light bulb for a short while. My 1st valve fatality since my teenage years in 1970s

[murphyv310]

Old age creeps up slowly neon

Here is the circuit of my little one valver, it impressed even more today as I could hear at more than perceptable volume BBC Wales!

[neon indicator]

That's similar to what I'm trying on LW on 1j24b.


Here is insides of 1j24b. I have a 2nd view later. Photo is bigger, but Forum shrunk it.

The four outer corner rods are in two up-side down U and the two bends at top connected by a "dish" under the join of the two anode rods. That's the Shield or Screen (pin 3). Not an Electrode. I have incorrectly labelled it as G2, but the test meter says those 4 x rods at corner and "dish" (visable end on) are Shield.

[neon indicator]

Fixed edit of photo labels

[neon indicator]

Here is quick reference guide.

PM me if you want Eagle Libraries (I have flat and vertical PCB versions too) or higher resolution quick guide.

I'll eventually have all the information collected as one document in a coherent fashion.

I did "joined" dashes on the grids to reflect the "Rod" structure as these are not conventional Pentodes or grids at all. I have kept the original inconsistent Russian pin numbering. Note the 1j37b has polarity on pin 1 filament reversed, not a mistake!.

P.S.
The "dish" (see previous photo) which is used to join the two sets of "shield"/"screening" rods (one rod at each corner) is likely also used for the "Getter". It's just below the "hoop" that connects the two anode rods (at opposite sides) to the top wire. The 1p24b mysteriously has no visible "getter" "bloom" at top of tube.

[neon indicator]

Examining filament currents in detail to decide on series resistor (if any) and schemes for series/ parallel.

The Philips/Mullard D series are 1.5V filamentary valves in a 50mA ('91) and 25mA and 50mA '96 family. The DL96 could be used as paralleled 25mA CT filaments 1.5V @ 50mA, or 3V @ 25mA to go in series with DF96 etc.

The Russian 1 series rod pentodes do have two parts with CT filaments, the 1j29b (similar power for audio as DL96, but rated to over 60MHz for RF) 26mA series 2.4V or 54mA parallel 1.2v and the 1p24b with 100ma series 2.4V or 200mA parallel 1.2V

Note the halves of 1p24b filaments are very closely matched and the 1j29b mis-match about 6%, so I only plotted one 1/2 of the parallel filament current on the graph. The real "full" filament mode is 200mA @ 1.2v. Philips describe a battery saver mode of using only 1/2 the filament. I will test what the 1j29b and 1p24b do wired like that.

The 1j37b is high consumption at 58mA. No doubt this is to allow the "mutual variable remote cut-off" action that having two identical but separate control grids gives.

In practice all the valves in the series appear to have 2 or 4 filaments very close to the multiple g1 rods.

The conclusion is that unlike the D... 96 series, the valves are more like E series, where each filament is a different current according to required emission of cathode. Neglecting the 1j17b (which seems an inferior earlier design) and the 1j37b (unusual mixer) the filament currents appear to be arranged such that the peak anode current is about 10%. This would be in line with filament life as well as emission needs

[neon indicator]

Three valve headphone/speaker amp + one valve Regen radio (can be used on its own with Hi-Z or crystal earpiece)

All 1j18b
HT is 2 x PP3 in series with LT.

4x AA Alkalines is too much! Running on 3x AA Alkaline. Solution is 4 x AA with a 68R, 33R or Zero Ohm in series according to battery State!

all 4 filaments in series

Coil is 3 & 2/3rd wires from 2 pairs of 2.9m 4pair Cat5e cable stripped on Toilet roll core about 65 to 70t (sprayed blue to stiffen it and keep damp out). 5th wire is 17 turns on top as feedback winding (if the pot gives no noise end to end, swap the wires from coil to pot). Gives 110uH for main coil, so may not tune all of MW.

The shiny speaker does work, but faint. I found that the 200 Ohm microphone insert made a "better" speaker (black object in photo). It's a small 5V mains transformer as O/P transformer.

Don't try running it on higher than 15V to 24V HT without changing bias and changing O/P stage to Pentode instead of Triode and RF to Pentode (separate R and C decouple for G2)
You need a slow motion drive as tuning is VERY sharp.

Yes, it's really on 2 x PP3, 530uA for HT, in series with 3 x AA (or 4x AA and 47 to 68 Ohm) 22mA LT current.

The Radio part works down to 15V HT on 1.5V battery on its own. With Ferrite Rod coils (MW & LW) or cardboard tube air core you do need an RF earth and at least 8' (2.4m) of aerial.

The 1mH coil and capacitor "isolate" the aerial a bit from the tuning coil or else it's not very stable! I'll try a frame aerial next instead of separate coil and Aerial.