Russian Military Subminature "pencil" valves

[fetteler]

Hi folks,
These little Russian valves are incorporated in a legendary Hammond organ mod, it's a pre amp to give a transistor organ a valve sound:-

http://www.keyboardpartner.de/hammon...structions.pdf

There's some nice features in this design, not the least is the simple switcher psu to provide the 1.2v heater supply.
I hope there's something useful here for you.....

Cheers,
Steve.

[neon indicator]

Put the HT- to the LT+ to get an extra 1.4V

Comparing the 1j24b, 1j18b and 1j29b.

Above 27V HT, the 1j29b is best choice. Ideally 36V to 90V HT.

The 1j24b isn't so good at low voltage.

At low voltage the 1j18b gives best gain as triode with the anode top wire connected to ground (not to G2). In this configuration from 9V to 20V it's superior to Pentode mode.

G2 is then the Anode.

Comparison of different loads on 1j18b with 2 x PP3
(best to worst. 2V pk to pk 1kHz into 1M Ohm g1 via 100nF)

1. 63 Ohm miniature speaker: Usable even at 200mV in
2. 1.7" dia 200 Ohm dynamic mic insert: Usable at even 200mV
3. 8 Ohms PC type speaker + miniature 230V/12V Mains step down. Faint at 200mV
4. 32 Ohms (or 32 + 32 Ohms L +R in series) Exterior Earphone: 2V is like 200mV on best, 2V very low, 200mV far to faint!

Simply paralleling 2 x 1j18b does double the current and give slight improvement. Likely better at higher drives.

[ms660]

Hello N I...I think what you are doing is amazing..probably be the definative research text on those pencil valves!
Initialy you said that it was for a fairly comprehensive comms. receiver...is the construction going to be modular... or stuff it all in there ?

Cheers.

[neon indicator]

Actually, the original ideas were very modest. A GDO/Wavemeter more sensitive than my Ambit/Circuit FET dip meter (Mk1 G3WPO 1.6MHz to 215MHz) for 100kHz to 150MHz with a built in VFD and PIC micro to measure GDO, frequency, RF volts etc. Also a Valve DC to RF Millivoltmeter. DMMs get upset with RF.

I have a friend who for last 4 or 5 years suggesting I build something valved. I had always fancied a "hat box" style Battery Valve radio like the one my parents had in 1960s.

So the idea is growing. The Communications TX/RX will be modular.

But the BBC 1950s and other old "one valve" radio threads and the Elektor kit thread made me think I should do a low battery voltage radio and amp.

I had the Radio working off 16V HT+ 1.2V LT on IC breadboard yesterday on MW & LW. About 10mV to 50mV audio from one 1j18b valve in Triode "regenerative" mode with a simple 10K pot to set regenerative feedback. Using old ferrite rod with LW & MW coils on it and about 8' (2.5m) of aerial wire.

I sprayed some toilet roll tubes to stiffen and will design LW, MW, SW DIY coil, with maybe option to tune with an aluminium foil covered cardboard tube instead of a variable capacitor. I'm toying with idea of solderless construction. Just screw terminal blocks.

A separate valve amp project with maybe 4 valves (gain is REALLY low at only +16V HT!) to drive speaker from the one valve radio.

Elektor cheat with 2 x transistors and an IC amp with their 1 valve. The old BBC one valve radio likely needed a long wire and earth and 2K Ohm or crystal earphones. I will have a screened loop aerial due to the high level electrical noise.

Aerial and coils to be made from cores of CAT5e cable, if that works. So only the valves, resistors, capacitors, switches and speaker need to be bought.

[neon indicator]

Originally Posted by fetteler

... not the least is the simple switcher psu to provide the 1.2v heater supply.
I hope there's something useful here for you.....

Cheers,
Steve.

Yes, thanks. I have the LT1073 as a nice cheap DIL part that can do +1.2 from 2.4V to 16V, or -12V from 1.0V to 16V or +45 direct from 1.0 to 16. Or 90V by adding another couple of diodes.

Some interesting ideas in the Keyboard pages, but initially I'm looking at the challenge of designs with no ICs or transistors

My GDO with VFO design uses 4 x LT1073 and 12V supply

You can see one page of that link how easily the Russian valves mount.

[fetteler]

Yes I can see where you are coming from - there's a tremendous satisfaction to be had by using a little ingenuity and staying with an all valve design
All this solid state stuff can make you a bit lazy.....
Cheers,
Steve.

[neon indicator]

Been designing Industrial & Telecoms Electronics from DC to 11GHz off and on since 1979.

Last valve gear I worked on apart from some Amateur Radio TX/RX was repairing valve L.O. Multiplier chain for a TWT based microwave link in BBC in 1976 or 1977. I did Triode Characteristic in Physics at School, very little valve experience since my teenage years.

[neon indicator]

2 x PP3 and 1 x AA or 4 x AA
1 valve regenerative radio 1j18b

Additionally:
3 x 1j18b as Audio Amplifier for 32 Ohm phones or 64 Ohm speaker:

All 4 filaments are connected in series and decoupled between each valve with capacitors to ground. Small series resistor value depending on if NiMH, Alkaline or 5V via USB power. Consumption 22mA to 23mA from 4 x AA cells (4V to 6V LT depending on cell type and state of charge)

"P.A" stage is 1j18b in "Triode mode" with miniature mains transformer 12-0-12 to drive speaker or earphones. Connected between g2 and HT+, the anode is Earthed. You can connect the speaker direct, but performance is poorer. the g1 is biased of f+ via 1M Ohm due to low HT, which is in series with LT+.

The two Pre-amp stages are identical Pentode Mode with 100K Anode load and g2 to HT+, the g1 is via 1M to f- I'll try the 1st stage with 220K, 330K, 560K later...

About 20mV in gives 2.5V to 3V at g1 of final "P.A.". I think the radio part will be 10mV to 50mV. Volume control needed between the 1st two preamp stages.

Total HT current excluding radio valve is 200uA @ 18V, 12V to 20V HT works.

So over 110 hrs for 4 x AA and likely 500hrs+ for 2 x PP3

[neon indicator]

1j18b as very low power Audio Out:

For best results so far for 12V to 20V HT
(1j18b g2 only as anode below 14V, Anode + G2 up to 20V, above 24V, use as Pentode with G2 to HT)



8 Ohm speaker I found a 1:17 Turns ratio transformer (actual turns 45: 400+400), Altai/Eagle part PT631T, Blue tape. Maplin sell it as an LT770 . CT is Primary in this application.
Primary is about 40 + 40 Ohms and secondary about 0.5 Ohms or less. It may be useful for a higher power audio amp using 2 x 1j29b in push pull (45V HT)


32 Ohms speaker/earphone is better with an 1:3.25 (6.5:1+1 per o/p) transformer (actual turns 2500:400+400). Altai/Eagle part PT631M Blue tape Listed as 20K primary 1K secondary at Maplin as HX82D LT44, but an LT44 is I think 1250:400+400 with yellow tape. CT is secondary in this application, 32 Ohms on either secondary.

PT631M
I think this transformer might do well as splitter to feed 2 x 1p24b PA valves. The driver could be 1j18b or 1j24b and grid resistors adjusted for best frequency response. 9:1 impedance ratio, DC primary 1.37K Ohms, DC secondaries 65 Ohm + 65 Ohm

[murphyv310]

Hi.
Well my experiments certainly don't tie in with yours Neon! I am using 1j18b valve, and have made up some AF amps and oscillators and a one valve reaction AM radio.
Firstly the Filament volts are not critical the valves seem happy between 1.0v-1.5v with best performance between 1.2 & 1.45v. The filament must be wired correctly with the dot on the valve connected to filament +. Bias can be obtained from either the - filament connection or a potential divider between the two, Filament + is taken as ground. I have tried all sorts of odd wiring mixtures and the only successful way to use these valves is in the the normal conventional way, any other gives poor performance, only one thing that improves is the frequency the valve will handle, not I say handle as gain in this instance is low possibly between 2 & 5 times.
Connecting the Valve as a triode A-G2 give the best performance at RF up to about 20mhz then the valve rapidly falls off.
HT seems to be not too critical and 30v works well.
I made up a simple reaction set for Long wave and both RTE & Radio 4 came in with reasonable volume before the onset of audible feed back.
Here is a photo of the project After it fell on the floor "mutter mutter"!


It's also interesting that the Russians didn't wire the G2 to the top of the valve and the anode to the base pinouts or even to Filament + if the performance was greatly improved by this method of connection.

Can you put some pictures up Neon of some scope traces and pictures of your work as I would like to see what you have done and check that I am following correctly your work!

[neon indicator]

The filament volts are only critical for long life. Also at higher powers there is more filament current from "cathode" current, rather than LT supply. The valves designed for NiCd batteries hence the 0.9V to 1.4V spec sheets. NiMH are bit more than NiCd on average and Alkaline a good bit more, hence the recommendation of series filament resistor. Also has the advantage of bias -0.25V to -0.5V more negative with grid if which desired can be an advantage.

It's only at very low supply volts with lower loads that there is any advantage to non-Pentode operation. Above 24V HT, Normal Pentode connection is best.

Some of the valves definitely need a grid more negative than f- even for Class A.

Breadboard of 1j18b in Triode Mode (anode earthed as HT only +12v!)

(Wired as oscillator with "tank" at G2 as Anode and the small separate winding on coil at grid to ground)

Injecting 1kHz 200mV at Anode grounded via 100K Ohm!


I did take photo, I thought of 4 x 1j18b on Breadboard with filaments in series. Not on camera, so here is next stage:

Designed for 4 xAA cells in series and 2 x PP3 (HT works 12V to 20V)
4th valve filament is 1j18b as Regenerative Radio.

Breadboarding a different amplifier, for 40V to 50V HT only:

Uses Push Pull 2 x 1j29b in class AB, so a grid bias battery needed (1 x PP3 and a 4M7 + set resistor.) Driver is 1j18b and phase splitter by transformer.

[neon indicator]

Some comparison of size:


Traditional direct and slowmotion MW Radio tuning capacitors, T092 BC547 Transistor, 1j29b Russian valve (only the 1p24b is 2mm fatter) and "traditional" valve with maybe 2W heater?


Four kinds of polycon. Much soldering melts the internal plastic. Top two are common dual gang MW/LW and quad Gang MW & VHF sections.


Various quality of MW LW combo coils. The MW needs to be about 0.5mH and Long Wave over 4mH. Should an old style 1 valve Regenerative radio on 18V use and air core coil?
For MW maybe 300 turns on 1.9" approx 4.7cm air cored coil.
For LW, going to a larger diameter cuts down turns. 330 Turns on 3.7" diameter.



Various MW only coils. I have a large rod MW only one which is of course from a scrapped USA transistor Tuner Amp

Rule of thumb:
Coils don't want to be any shorter than diameter and any longer than diameter x 2

P.S.

I'll do up pretty schematics and post them.

[murphyv310]

So why then can I not get much performance with a grounded anode? They work very well conventionally wired, am I missing something?

[neon indicator]

Originally Posted by murphyv310

Connecting the Valve as a triode A-G2 give the best performance at RF up to about 20mhz then the valve rapidly falls off.
HT seems to be not too critical and 30v works well.
I made up a simple reaction set for Long wave and both RTE & Radio 4 came in with reasonable volume before the onset of audible feed back.
Here is a photo of the project After it fell on the floor "mutter mutter"!

Thanks, very interesting. Above 32V performance improves.

Originally Posted by murphyv310

It's also interesting that the Russians didn't wire the G2 to the top of the valve and the anode to the base pinouts or even to Filament + if the performance was greatly improved by this method of connection.

The Russians not interested in low voltage operation (Less than 45V). Performance from 9V to 25V is terrible with any valve.

Originally Posted by murphyv310

Can you put some pictures up Neon of some scope traces and pictures of your work as I would like to see what you have done and check that I am following correctly your work!

At higher voltages of G2 (> 40V) the Pentode performance improves. High frequency performance can be "tweaked" when HT is 45V by the "right" value of resistor that is NOT decoupled between G2 and HT. Anode load is important. Higher is more gain, but HF starts to suffer. Too low a load and HF also suffers. The 1j18b is about the poorest for RF, as the G3 is connected internally to Filament/Cathode.

At higher frequencies G3 needs to be RF earth and voltage apparently can be other than 0V.

Decoupling both ends of filament and good RF layout...

"Unusual" connections are only for 9V to 20V operation. Normal 36V to 50V (All valves), 36V to 100V (1j29b) and 36V to 180V (1p24b) are best connected "conventionally".

Construction
It seems to me that the rods are in concentric circles around the filament. Hence chopping the G1 into two groups produces the 1j37b with a g1a and g1b. The base wires seem to be on a button and base done a bit like B7A, but wires, rather than pins. (US and Mullard "pencil" valves appear to use inline wires and simply crimp the base of the tube). The rods you see are the Anode and are all connected together at the top, so exiting the anode wire via the evacuation tube at top seems logical. I don't know how the "getter" was activated, but it's very uniform on top. Very shiny when you peer in.

This is the best article I found: http://www.radiomuseum.org/forum/rus...ure_tubes.html

Worth re-reading several times.

Pictures from http://www.hi-ho.ne.jp/ux-45/russian.html and http://www.radiomuseum.org/forum/rus...ure_tubes.html

These show why at very Low voltage where "space charge" predominates that "Triode Mode" works better than Pentode and as voltage gets lower the Anode is less important and can be used as a "gammatron" type grid to modulate by electrostatics the G2 used as an Anode. There are no fine Grid wires or meshes.

[neon indicator]

Originally Posted by murphyv310

So why then can I not get much performance with a grounded anode? They work very well conventionally wired, am I missing something?

No, you are right, for a given value of right. In normal voltages and loads the normal wiring is best. But if with low load and low voltage, then you can slightly improve on absolutely terrible performance (driving 1K and only 12v HT)

Above 15V HT to 22V tying Anode and G2 together is marginally better gain. Above 24V tying G2 to HT+ or via a resistor no larger than x1.5 to x2 of Anode load is best (Pentode).

At 9V to 15V see a drop in gain if Anode is tied to G2! Floating is better gain. But it acts as a Gammatron grid, so has to be grounded or it will pickup!

But this is at low values of Load (< 10K). At 100K to 200K load, even at 16V the 1j18b works better as a Pentode.

Put just mA meter to HT and plot HT current vs G1 from 1V above f+ to 1V below f-

Do it 5V to 20V. Stop raising G1 any time HT current exceeds 1.5mA

1) G2 & Anode tied
2) G2 only and Anode floating
3) G2 only and Anode grounded.

If you have either more than 18V HT or more than 56K Ohm load you may not see this

Also do put 12 Ohm in f- lead (NiMH) or possibly 15 R to 18R if Alkaline and connect g1 to LT- and put HT greater than 35V and see gain with 56K and 330K loads and G2 direct to HT, then G2 via 470K and G2 via 680K to HT (keeping the 56K and 330K loads)
Compare frequency response and shape of a 200mV 1kHz Square wave.

P.S. See this too http://www.radiomuseum.org/tubes/tube_1j24b.html

[neon indicator]

P.S.
If you want gain, use a 1j24b with 470k to 560K ohm load and 44V to 55V HT.

Next stage should be 1M to 4M7 grid bias and not a 1j29b or 1p24b

[murphyv310]

Hi.
I might get a chance to have a further play around today, it depends what else I need to do. But I think I will order some other types today.
Thanks for the info Neon, Good luck.

[neon indicator]

There are different kinds of applications and reason to use these Russian Tubes:

• Recreate old 1920s and 1930s single valve Radios
• Low cost
• Easily Available
• Low heater power consumption
• Create "high" radio using many valves, inexpensively.
• Get a "Valve" sound for Musician etc.
• Make hybrid equipment with no limit of ICs, Transistors etc
• Very low battery voltages (9V to 22.5V). Below specified Operation
• Medium Performance Battery operation (33V to 50V)
• High Performance Battery operation (50V to 90V)
• Higher voltage applications (100V to 400V) Exceeding specified Operation except power types

Traditional Battery valves are specified usually for 67.5V and 90V Nominal for HT and 1.5V LT using Zinc Carbon Dry cells.

These "Pencil Tube & Rod Electrode" Russian valves apart from "power types" 1j29b and 1p24b are designed for 45V Nominal HT and 1.2V Nominal LT using NiCd rechargeable cells or equivalent power sources and purely for Military applications. Hence the datasheets are not the same assumptions as Philips/Mullard DL96, DAF96, DK96 and DF96 family (the closest family in terms of performance)

Rough Application equivalents (not at all identical)

• 1j17b: Maybe like DF91, but not variable mu. Obsolete?
• 1j18b: DAF96's Pentode section. AF, IF, RF up to 60MHz (100MHz possible)
• 1j24b: DF96 but not variable mu
• 1j37b: DK96, but a Pentode, not Heptode. Also can be used as AGC and variable mu (g1a signal and g1b mu / AGC setting). Mixer and Mixer/Osc
• 1j29b: Quite like DL96, but higher heater power. Also used for IF and RF Driver.
• 1p24b: No Battery equivalent. Maybe Audio wise like an EL42, but 1/6th heater power! AF/RF 2W dissipation and up to 150MHz.

[neon indicator]

I'll do details later of a MW Regen receiver with 1j18b, with 18V HT, maybe as low as 8V.

But I'm a bit short any decent MW here since Cork and Athlone MW all closed.


So 1j18b Test Generator is needed. Works as low as 7V HT by putting the HT- to LT+ battery. The "f+" is fed by 12 Ohms, which gives exactly 24mA at 1.5V. "f-" to LT-

Oscillator is MW ultra-miniature ferrite 2.5mm thick, 8mm wide and 40mm long. About same size as the valve!

With "conventional" wiring the lowest volts for reliable operation with G2 to HT and Anode has Tuned load is about 16V to 18V.


Using G2 as Anode allows operation down to 7V. Real "Anode wire" connection can be decoupled to OV via 470pF and used for AF modulation in.

The isolated smaller number of turns go between g and "f+". The larger bunch of turns from G2 to HT+. Tuned with 47pF.

The Audio modulation via jack socket and 0.47 uF ceramic (2x 220nf on my board). All other capacitors for decoupling.

You only need on/off on LT as there is zero HT current with no filament supply.


"Anode" RF level is 1.25V RMS (3.5V pkk) approx at 600KHz when tuned with 47pf (scope probe likely pulling frequency low). Connect to Radio by one turn of wire around the ferrite rod to the Aerial and Earth terminals. Or 2 turns on Radio rod and Generator rod via twisted pair.

350mV RMS Audio is 50% modulation, ( 1v pkk) so about 0dB is about 100% modulation. The 12k between Anode and +HT adjusts modulation point depending on HT volts. 18K seems about right for a Alkaline PP3.

LT consumption on Fresh Alkaline AA battery is 24mA
HT consumption on Fresh Alkaline PP3 is 210uA (0.21mA)

I'll post up schematic and full details at a later date at appropriate places.

[murphyv310]

No progress I am afraid here today, as I was waxoyling the Doblo and putting a couple of things to rights for the MOT test.
The weather was wonderful and not to be missed. Perhaps Wednesday!