Russian Military Subminature "pencil" valves - Page 3 - UK Vintage Radio Repair and Restoration Discussion Forum

neon indicator · 6th Jul 2011, 9:11 pm

Which valve is it?
What about higher anode resistor?
and regard f- as 0V rail

neon indicator · 7th Jul 2011, 9:11 am

Leaving the High Voltage endurance testing up to murphyv310

Going back to the 1j18b as it's simple and cheap. Only f+, f-, g1 and g2 wires on base plus anode wire on top.

At 45V HT, there is marginally more gain with 12 Ohms in series of f- to 0V than 10 Ohms, due to change in grid bias probably.
With 1.35V LT (NiMH)
12 Ohms is Vf = 1.08 If = 23mA approx
10 Ohms is Vf = 1.13 If = 24mA approx

It's likely this valve will be used for lower level signals so connecting g2 to HT+ rather than via a resistor. If a resistor is used, more than 10K started to sharpen transient response too much!
Anode load for maximum swing is about 33K to 47K with Vg1 = 0V (i.e. -0.25V compared to f-) and HT = 45V. Voltage Gain about 14 to 15 depending on sample tested with R in series with filament = 10 Ohms. With 12 Ohms gain is 16. With 15 Ohms down again to 15.5

G1 "grid leak" was 100k Ohms. No current was detected.

Testing at lower voltage Is a 9V radio or amplifier possible?

Below about 20V HT I found that more gain was obtained in Triode mode, using g2 as anode and grounding Anode. (connecting Anode & G2 is slightly lower gain!)

Below about 22V HT the bias and gain is better connecting g1 to f+ via 100K (no significant current flow).

True Triode mode
g1 -> f+ via 100K, Rload = 47K, g2 as Anode, Actual anode earthed, 200mV squarewave 1kHz:
7V G = 2
9V G= 3.5
12V G = 5
15V G = 6
18V G = 7
22.5V G= 8.2
(10 Ohms, 12 Ohms or direct connection to 0V/LT- for f- made virtually no difference to gain, demonstrating the change with Vg1 = 0V is due to variation of f- volts more than current)

Pentode Mode
g1 -> 0V via 100K (f- less 0.25v), Rload = 47K, Anode as Anode, G2 to HT+ squarewave 1kHz:
Below 21V the gain is better configured as previously
22.5V G = 10.5 (10 with G1 -> f+ via 100k instead)
32V G = 15
45V G = 16

Conclusion
With HT from 24V to 60V (suggested Maximum for 1j18b, though 90V works!) the 1j18b should be operated as a Pentode, as designed. 12 Ohms series on f- reduces filament voltage by about 10% which is likely to double life and actually improves operating point. With 47K load gain is between 14 and 16 across samples and voltage range 32V to 45V. It's likely a gain of 24 is possible with small signals and 68K load, perhaps 42V to 60V.

Operation at 9V is feasible, though 18V is a better choice (3.5 to 7 gain). At 9V to 22.5V batteries the g1 should be biased from LT+ the f+ pin. g2 is used as anode and top wire Anode connected to 0V as additional shield.

For Pentode operation with small signals a 24V HT is feasible, but 44V nominal is likely much better and allows over 16V squarewave output. Though a 67.5V pack is slightly over the recommended voltages, it's likely a good HT choice for performance if you don't want to run off 90V, which does appear to be beyond normal spec if G2 is connected to HT+. With suitable G2 resistor, it's likely 90V HT could be used within operational limits, though a -ve g1 bias likely required to get full signal signal with symmetrical clipping.

dominicbeesley · 7th Jul 2011, 9:55 am

Do keep all this info coming...I'm deciding which valves to get but will be joining in the pencil revolution soon!

neon indicator · 7th Jul 2011, 10:42 am

Thanks for words of encouragement dominicbeesley.

Depends what you want to do. Recommend series resistor on f- to add some negative bias with g1 = 0V (HT- and LT- are common 0V) and increase life, with 1.1V approx filament instead of 1.2V (recommended limits are 0.9V to 1.4V)

1j17b : Not worth getting

1j18b : General purpose, cheap, 23mA filament, only 4 base connections, Minimum load about 33K. 24V to 60V HT as Pentode, 9V to 23V HT as triode.
1j24b : A bit more expensive. 11mA filament, All pentode connections separate as well as a "Shield Wire". Minimum load about 68K. 24V to 60V HT as Pentode.
1j29b : Roughly similar to DL96 in performance. 52mA filament (though much smaller). Can run at 24V to 100V as Pentode. RF/IF/AF "power" output down to 8.2K Ohm load. The 1j29br (marked 1Ж29Бр ) is long life possibly 5,000+ preferred version. Use as PA for audio/RF in it's own right, (also used as IF on Russian Military Radios). Needs about -1V grid bias, can be achieved by 2 x 1j18b filaments in parallel connected between f- and 0V (100uF 3V decoupling needed) and using 2 x AA cells as LT+
1j37b : Dual control Grid Pentode. Filament 52ma?. More like a dual gate jFet than a DK96 Frequency Changer Heptode, Pulling both g1a and g1b negative is cut off. Mixer, frequency doubler, Mixer/Osc or Fullwave detector!
1p24b : Serious Pentode, 10mm instead of 8mm package with ribbed glass. 200ma filament sounds a lot. But this will do 1W in push-pull @ 45V and 8W in push-pull @ 140V. A 50hrs rating for 800W per valve 2us pulse operation!

All are suitable for AF to 60MHz (spec). Some are specified to 100MHz. Likely all apart from 1j18b can do near 150MHz. The 1j18b can operate to over 100MHz.

The 1j18b seems suited to operate as a triode at 9V to 22V.

A google finds many sellers. Russian and Ukrainian sources are likely "primary" sources. Prices 50p to £1.20 each inc postage depending on quantity, so likely cheaper than D-96 series Philips/Mullard NOS. Also manufactured up till 1980 at least. Some I have are 1978 date codes! The "traditional" battery valve production likely ceased in early 1960s. Performance is not great compared with mains valves with 250V HT and 2W heater consumption, but seems as good or better than "western" battery valves. I haven't compared nuvistors though.

What's Next?
I need to test 1j37b..

Keeping to +48V and lower HT while poorer performance than 67V or 90V, makes it safe for "newcomers" and kids. You can use 5 x Alkaline PP3 or 6 x NiMH PP3 for HT and 1.5 Alkaline with series resistor per filament or NiMH with smaller resistor per filament. A single PP3 or 3x coin cells (CR2032) will last a LONG time as grid bias for 1j29b and 1p24b.

No valve sockets, woodwork or metal work. A plugin breadboard works as does veroboard!

Apart from my own projects the next stage is some simple radios:
1) A version of the BBC 1952 one valve set. (maybe with options for 67V and 90V HT as well as +44V to 48V)
2) I criticised the Elektor Valve kit. A four valve radio that drives a loudspeaker. LW, MW and SW options. (or single band).
3) maybe a two valve headphone set for 32 Ohms headset.

dominicbeesley · 7th Jul 2011, 11:48 am

I'm not sure yet, after my recent success with a OTL headphone am I might make a tiny headphone amp but with trannys. then possibly a small (1") 405/32 line telly in a tin.

I'm already playing with some 5874's (dual control pentodes probably similar idea to 1j37b but with a seperate cathode?) to try and make a 405 NTSC decoder but not got round to doing anything yet but they look like they should be pretty simple so 1p24b's might make good cathode/grid drivers for my little 4" colour crt...I like the idea of the decoder/CRT driver being fairly small as otherwise there will be a lot more telly than screen!

Not sure how I'll construct things but might solder them to PCBs rather than try and get sockets...what do you think?

Dom

neon indicator · 7th Jul 2011, 1:05 pm

You can use cut up IC sockets. I'm just soldering. That's what the Russians did and all hobby radios I have seen on web pages (soldering).

1j37b Tests
Imagine a 2N3819 FET with two gates. That's what the 1j37b is like!

With +45HT via 22K to Anode, G3 to 0V, G2 = HT
LT is 1.35V to f+ and f- is via 4.7 Ohm to 0V / LT- / HT- (about Vf 1.08, If = 54mA)

Code:

G1a   G1b   Va
 0    0     13.5
-0.5 -0.5   13.5
-1.6 -1.6   27.2
 0   -2.5   21.0
-2.5  0     24.0 
-2.5 -2.5   38.0
 0   -5.0   34.96
-5.0  0     36 
-5.0 -5.0   45.0

neon indicator · 7th Jul 2011, 1:16 pm

Quote:

Originally Posted by dominicbeesley

so 1p24b's might make good cathode/grid drivers for my little 4" colour crt...I like the idea of the decoder/CRT driver being fairly small as otherwise there will be a lot more telly than screen!

Dom

ha! Cathode Drive:
You could drive 4" CRT with any of them. The 1p24b could drive up to 400V and sink 25mA RMS! You could run it at 150V HT and 8mA for mid grey (1.2W, Max limit is 2.5W)! with cut off @ 400V and white at 30V, 20mA (600mW) approx. I think that would drive a full size TV!

At up to 2048 x 1024 pixels @ 60Hz progressive

dominicbeesley · 7th Jul 2011, 3:01 pm

Maybe that's overkill then :-o I think they need to be sat at about 75V +/- 15 IIRC...

D

neon indicator · 7th Jul 2011, 3:53 pm

neon indicator · 7th Jul 2011, 4:09 pm

http://www.radiomuseum.org/forum/rus...e_tubes.html#4

kalee20 · 7th Jul 2011, 6:29 pm

I'm watching this thread, too. These little valves look cute, and they are cheap. I like the idea of a push-pull amplifier using a pair of 1p24b's.

neon indicator · 7th Jul 2011, 6:29 pm

This is interesting related to 1j37b, about 4/5ths down page
http://www.radiomuseum.org/forum/dua...read_id=211038
Russian Subminiature Tube 1j37b

Quote:

Unfortunately, I don’t have a 6AX5 GT in my possession, so I used for the experiment the 1j37b. The internal construction is completely different in respect to normal radio tubes. You won’t find grids, instead, tiny metal bars and rods. The abbreviation “g” for grid was maintained!

See another 1j37b http://www.elforum.ro/viewtopic.php?f=183&p=646312

neon indicator · 7th Jul 2011, 7:31 pm

AM signal generator to feed your AM radio Via attenuated aerial connection (do not connect an aerial!) using the 1j37b
http://www.radiomuseum.org/forum/15v...ansmitter.html

I have now tested the 1j37b in "Gammatron" mode, with g1' (ga) as "grid" and g'' (gb) as "anode" (4k7, 10K, 33k loads). Also vice versa. I only get a gain of 1/2 but it does invert. This is with 0V, -0.5 and -5V "grid" bias and 6V to 20V HT. Actual Anode, G2 and G3 all at 0V.

Indeed strange.

neon indicator · 7th Jul 2011, 11:40 pm

1j37b osc/mixer

Adding a MW transistor radio coil with main part as anode load tuned with a 100pF capacitor and separate short winding from g1b to ground and g1a 0V, -0.25 (100k Ohm g1a to OV) gives about 40V peak to peak 790kHz (frequency counter held near) and 0.8mA HT current with 32V HT and 75V ppk with 45V HT, 1.65mA Total HT current (High Q coil I guess!)

Adding 200mV 1KHz gives visable AM (but not much modulation depth) on Scope.

Confirmed by Yaesu FT817ND radio (whip held next to wiring,) 1kHz tone on 790KHz AM

G2 and HT+ is +45V
G3 = 0V
f+ =LT +1.35
f- via 4.7 Ohm to OV (LT-, HT- common)

The coil I had has an isolated winding. If yours has no isolated winding about 10t should be more than enough between G1b and OV. If it doesn't oscillate, then swap the wires on ONE coil only.

Don't connect an aerial. It's a frequency changer we are making.

I'll look at a better circuit and test with an RF source to get 455kHz, 10.7 and 32.9MHz IFs

I'll look tomorrow at Reflex and Regenerative single valve radios using 1j18b as triode, 1j37b as gammatron and as Amp/mixer/osc and 1j24b as regenerative pentode with 1j29b or 1j18b or both audio amp.

I've partially soldered up a push pull amp using 2 x 1p24b to give 500mW to 1W on 45V Audio AND 6W to 8W RF simply by switching HT and changing between audio and RF drive.

I've the phase splitter to solve

A transformer is simplest and lowest power consumption, but I'm not sure I have a suitable one. A small 110/220 mains transformer is Audio O/P. 10:1 ratio is fine. I can put the secondaries in parallel or series and try different loads. On RF it's simply part of the power feed.

neon indicator · 8th Jul 2011, 7:06 pm

1j18b used in triode AND "gammatron" mode, 12V to 18V HT as reflex regenerative radio...

g2 is Triode "plate" for RF and "gamatron grid" for AF
g1 is "Grid"
0v is f+ and f- goes to -LT (NiMH battery) via 12 Ohms.
Anode is a "gammatron grid" for RF (1M Ohm to Ov!) feedback

g1 has main tuned circuit and optional whip via loading coil.

about 5 to 10 turns over same coil, or the separate transistor "base" winding on ferrite rod is "tickler" coil in series with 27K Ohm load on g2 with 10nF across it. 1nF from g2 to anode

About 50mV AF from RTE on 252 KHz

neon indicator · 8th Jul 2011, 7:38 pm

What does Gammatron mean? Influencing a current on an electrode by voltage on an electrode that is NOT between it and the Cathode. A suitably constructed dual anode "fullwave rectifier" valve can actually be used as Gammatron type amplifier. Use one Anode as a "grid".

neon indicator · 8th Jul 2011, 9:01 pm

1j18b 2 x PP3 + AA battery radio:

Although the bias is much poorer operating current, the gain is much better with g1 referenced to LT- than to LT+, even at 18V HT.
I'll do up the schematic of valve amp. 2nd stage is 1j18b as audio amp, maybe. I've no high impedance phones or crystal earpiece, so may add 1j29b as speaker amp.

neon indicator · 9th Jul 2011, 9:13 am

Some Russian valves have additional "r" or "v" (Russian "р or P" and "В or в", printed/etched on valve glass, the english "b" is Russian Б or б )

So 1Ж29Б or 1ж29б is listed or transliterated variously
1j29b
1sh29b
1zh29b
The final "b" can look very like a six.
(see http://en.wikipedia.org/wiki/Romaniz...teration_table)

The 1Ж29Бр is thus sold as 1j29b-r and 1Ж29Бв sold as 1j29b-v

What does this -r and -v mean? I suggested earlier that while datasheets might show 500 hrs life, this is for 5% failures. CFL lamps at 2000 hrs is actually 50% failures! So they use a conservative rating. Also as filament current rises due to fresh battery AND/OR Increased Screen and Anode currents you reach a maximum Ik (Cathode current or "ток катод" on Datasheet). Remember it's like a filament lamp, a small increase beyond rating hugely reduces life and a small decrease dramatically increases life.

Quote:

Letter V means that it is a tube of mechanical extra-strong and reliability.
The letter of R means that it is a tube of mechanical extra-strong and reliability (better, than V).
These additional letters were used in the USSR for marking of tubes, created for military industry.

So the 1j29b-r is extra reliable and strong.

neon indicator · 9th Jul 2011, 10:27 am

There is no doubt these filaments designed for NiCd rather than Zinc Carbon with thus a min of 0.9V, max of 1.4V and characterised for 1.2V. The Lifetimes quoted (other than insane pulse power 50hours at 400x peak power!) are likely for 1.4V and Maximum Operational Ik as these are Military data sheets.

NiMH batteries are up to 1.48V straight from charger. Don't charge NiMH connected to filaments! Then for quite a while are 1.35V dropping slowly to 0.95V sensible end point.

So my theory is that a small resistor in series with f- will reduce cold filament inrush and greatly lengthen life. Dropping about 0.2V to 0.25V appears to make little difference to performance, performance drops quickly below 0.8V though. With Alkaline battery the fresh cell is nearly 1.6V and drops much more slowly than Zinc Carbon. Perhaps the series resistor per filament should be larger by x1.5 to x2 with Alkaline. Resistors should be per filament rather than a single resistor in case a filament fails or a valve taken out of circuit.

Normally the "bias" should be for Anode current very much less than "Ik max".

Code:

1) Ia max is Va Greater than 30V, Vg2 = +45V, Vg3= 0V approx
2) Last column is max recommended "HOT" Class A bias, not ideal bias.
3) Vg= Gnd means "g1" to LT-, resistor Lt- to "f-" and LT+ to "f+"
4) For sensible gain the Anode load should be about x2 to x3 or more the listed value
5) g2 (screen grid) either direct to HT or not more than Anode Load x1.5 and 
not decoupled (negative feedback)
6) Filament currents approximate 

Valve   Ik   Ia abs  Ia max   Ia      Load If Series R          Comment
       Max    Max      bias   Vg=Gnd          filament
1j17b   5      1.7     1.2?   1.2?   >24k  54 4R7               Not tested much
1j18b   5      1       0.65   0.65   >27k  23 12R
1j24b   1.6    1.5     1.0    0.27   >56k  11 22R
1j29b   8      5.5     4      4      >3k9  53 10R per filament   Needs g1 < "f-" by -1V for 2mA
1j37b   4.5    2.4     1.7    1.43   >22k  54 4R7               g1a & g1b 0V to -3V, both -5V cut off
1p24b  25     24      18     15      >820 200 2R2 per filament  g1 -5V below f- is about 1mA

At 45V HT class A 1p24B about 600mW o/p. At 140V HT a pair in Class AB might do > 6W
(not much Voltage gain though)

neon indicator · 9th Jul 2011, 10:49 am

1) Put two (or more) filaments in series/parallel.

A pair of 1j18b are 46mA. The 1j17, 1j29 or 1j37 is about 54mA. These can benefit from a bias about -1V below f- for the g1 (control grid(s) ).

Thus you can use 2 x NiMH cells (2V to 2.8V supply) as LT+

the g1 of 1j18b valves and their filaments f- resistors (12 Ohms) go to LT-
You need 53-48 = 5mA more, so put 1.35/5mA = 270 Ohms in Parallel.

The 1j17, 1j29 or 1j37 f- resistor(s) then go to the f+ of the 1j18b valves and are decoupled with 100uF (3V tantalum min) + 100nF + 1nF to LT- to avoid crosstalk (cathode modulation and negative feedback on the upper valve). The 1j17, 1j29 or 1j37 control grid(s) are taken to LT-
The Shield (if fitted) can go to LT- (0V or Ground). G3 can go to 0V(LT-), 1/2 LT+ or LT+ (f+ of the 1j18b valves). Optionally for better Low frequency isolation you could connect the centre tap of the two NiMH cells (or Alkaline) to the f+ of the 1j18b.

Similarly you could have two sets of such circuits and use a 1p24b in series filaments mode (centre tap must be decoupled, and f+ of the 1j29b etc) to use 4 x NiMH cells with 100mA consumption for 7 valves
4 x 1j18b
2 x 1j29b (or 1j17b or 1j37b)
1 x 1p24b (filaments in series). The 1p24b g1 to LT- would give then approximately 2mA Ia, instead of over 12mA if connected to "f-" .

You could drop the 2 x 270 Ohms on the 4 x 1j18b and add a 1j24b in parallel (11mA + 23 x 4 is close enough to 54mA x 2).

2) Use a PP3 as grid bias. Bias can be about 0.2uA to 2uA per valve. With 4 valves and suitable grid to 0V resistors and 4M7 from Battery to Grid, any voltage from 0V (f-) to -8V can be set. With 6 valves many thousands of hours is possible. In high voltage (145V) class B push pull 2x pp3 might be needed. Have a method to check grid bias supply. On the 1p24b particularly failure is bad.

3) Pick off part of the Local Oscillator (if there is one) and rectify, filter and smooth as a -12V rail.

6th Jul 2011, 9:11 pm	#41
neon indicator Retired Dormant Member Join Date: Jun 2010 Location: Co. Limerick, Ireland. Posts: 1,183	Re: Russian Military Subminature "pencil" valves Which valve is it? What about higher anode resistor? and regard f- as 0V rail

7th Jul 2011, 9:11 am	#42
neon indicator Retired Dormant Member Join Date: Jun 2010 Location: Co. Limerick, Ireland. Posts: 1,183	Re: Russian Military Subminature "pencil" valves Leaving the High Voltage endurance testing up to murphyv310 Going back to the 1j18b as it's simple and cheap. Only f+, f-, g1 and g2 wires on base plus anode wire on top. At 45V HT, there is marginally more gain with 12 Ohms in series of f- to 0V than 10 Ohms, due to change in grid bias probably. With 1.35V LT (NiMH) 12 Ohms is Vf = 1.08 If = 23mA approx 10 Ohms is Vf = 1.13 If = 24mA approx It's likely this valve will be used for lower level signals so connecting g2 to HT+ rather than via a resistor. If a resistor is used, more than 10K started to sharpen transient response too much! Anode load for maximum swing is about 33K to 47K with Vg1 = 0V (i.e. -0.25V compared to f-) and HT = 45V. Voltage Gain about 14 to 15 depending on sample tested with R in series with filament = 10 Ohms. With 12 Ohms gain is 16. With 15 Ohms down again to 15.5 G1 "grid leak" was 100k Ohms. No current was detected. Testing at lower voltage Is a 9V radio or amplifier possible? Below about 20V HT I found that more gain was obtained in Triode mode, using g2 as anode and grounding Anode. (connecting Anode & G2 is slightly lower gain!) Below about 22V HT the bias and gain is better connecting g1 to f+ via 100K (no significant current flow). True Triode mode g1 -> f+ via 100K, Rload = 47K, g2 as Anode, Actual anode earthed, 200mV squarewave 1kHz: 7V G = 2 9V G= 3.5 12V G = 5 15V G = 6 18V G = 7 22.5V G= 8.2 (10 Ohms, 12 Ohms or direct connection to 0V/LT- for f- made virtually no difference to gain, demonstrating the change with Vg1 = 0V is due to variation of f- volts more than current) Pentode Mode g1 -> 0V via 100K (f- less 0.25v), Rload = 47K, Anode as Anode, G2 to HT+ squarewave 1kHz: Below 21V the gain is better configured as previously 22.5V G = 10.5 (10 with G1 -> f+ via 100k instead) 32V G = 15 45V G = 16 Conclusion With HT from 24V to 60V (suggested Maximum for 1j18b, though 90V works!) the 1j18b should be operated as a Pentode, as designed. 12 Ohms series on f- reduces filament voltage by about 10% which is likely to double life and actually improves operating point. With 47K load gain is between 14 and 16 across samples and voltage range 32V to 45V. It's likely a gain of 24 is possible with small signals and 68K load, perhaps 42V to 60V. Operation at 9V is feasible, though 18V is a better choice (3.5 to 7 gain). At 9V to 22.5V batteries the g1 should be biased from LT+ the f+ pin. g2 is used as anode and top wire Anode connected to 0V as additional shield. For Pentode operation with small signals a 24V HT is feasible, but 44V nominal is likely much better and allows over 16V squarewave output. Though a 67.5V pack is slightly over the recommended voltages, it's likely a good HT choice for performance if you don't want to run off 90V, which does appear to be beyond normal spec if G2 is connected to HT+. With suitable G2 resistor, it's likely 90V HT could be used within operational limits, though a -ve g1 bias likely required to get full signal signal with symmetrical clipping.

7th Jul 2011, 9:55 am	#43
dominicbeesley Octode Join Date: Nov 2004 Location: Hebden Bridge, West Yorkshire, UK. Posts: 1,885	Re: Russian Military Subminature "pencil" valves Do keep all this info coming...I'm deciding which valves to get but will be joining in the pencil revolution soon!

7th Jul 2011, 10:42 am	#44
neon indicator Retired Dormant Member Join Date: Jun 2010 Location: Co. Limerick, Ireland. Posts: 1,183	Re: Russian Military Subminature "pencil" valves Thanks for words of encouragement dominicbeesley. Depends what you want to do. Recommend series resistor on f- to add some negative bias with g1 = 0V (HT- and LT- are common 0V) and increase life, with 1.1V approx filament instead of 1.2V (recommended limits are 0.9V to 1.4V) 1j17b : Not worth getting 1j18b : General purpose, cheap, 23mA filament, only 4 base connections, Minimum load about 33K. 24V to 60V HT as Pentode, 9V to 23V HT as triode. 1j24b : A bit more expensive. 11mA filament, All pentode connections separate as well as a "Shield Wire". Minimum load about 68K. 24V to 60V HT as Pentode. 1j29b : Roughly similar to DL96 in performance. 52mA filament (though much smaller). Can run at 24V to 100V as Pentode. RF/IF/AF "power" output down to 8.2K Ohm load. The 1j29br (marked 1Ж29Бр ) is long life possibly 5,000+ preferred version. Use as PA for audio/RF in it's own right, (also used as IF on Russian Military Radios). Needs about -1V grid bias, can be achieved by 2 x 1j18b filaments in parallel connected between f- and 0V (100uF 3V decoupling needed) and using 2 x AA cells as LT+ 1j37b : Dual control Grid Pentode. Filament 52ma?. More like a dual gate jFet than a DK96 Frequency Changer Heptode, Pulling both g1a and g1b negative is cut off. Mixer, frequency doubler, Mixer/Osc or Fullwave detector! 1p24b : Serious Pentode, 10mm instead of 8mm package with ribbed glass. 200ma filament sounds a lot. But this will do 1W in push-pull @ 45V and 8W in push-pull @ 140V. A 50hrs rating for 800W per valve 2us pulse operation! All are suitable for AF to 60MHz (spec). Some are specified to 100MHz. Likely all apart from 1j18b can do near 150MHz. The 1j18b can operate to over 100MHz. The 1j18b seems suited to operate as a triode at 9V to 22V. A google finds many sellers. Russian and Ukrainian sources are likely "primary" sources. Prices 50p to £1.20 each inc postage depending on quantity, so likely cheaper than D-96 series Philips/Mullard NOS. Also manufactured up till 1980 at least. Some I have are 1978 date codes! The "traditional" battery valve production likely ceased in early 1960s. Performance is not great compared with mains valves with 250V HT and 2W heater consumption, but seems as good or better than "western" battery valves. I haven't compared nuvistors though. What's Next? I need to test 1j37b.. Keeping to +48V and lower HT while poorer performance than 67V or 90V, makes it safe for "newcomers" and kids. You can use 5 x Alkaline PP3 or 6 x NiMH PP3 for HT and 1.5 Alkaline with series resistor per filament or NiMH with smaller resistor per filament. A single PP3 or 3x coin cells (CR2032) will last a LONG time as grid bias for 1j29b and 1p24b. No valve sockets, woodwork or metal work. A plugin breadboard works as does veroboard! Apart from my own projects the next stage is some simple radios: 1) A version of the BBC 1952 one valve set. (maybe with options for 67V and 90V HT as well as +44V to 48V) 2) I criticised the Elektor Valve kit. A four valve radio that drives a loudspeaker. LW, MW and SW options. (or single band). 3) maybe a two valve headphone set for 32 Ohms headset.

7th Jul 2011, 11:48 am	#45
dominicbeesley Octode Join Date: Nov 2004 Location: Hebden Bridge, West Yorkshire, UK. Posts: 1,885	Re: Russian Military Subminature "pencil" valves I'm not sure yet, after my recent success with a OTL headphone am I might make a tiny headphone amp but with trannys. then possibly a small (1") 405/32 line telly in a tin. I'm already playing with some 5874's (dual control pentodes probably similar idea to 1j37b but with a seperate cathode?) to try and make a 405 NTSC decoder but not got round to doing anything yet but they look like they should be pretty simple so 1p24b's might make good cathode/grid drivers for my little 4" colour crt...I like the idea of the decoder/CRT driver being fairly small as otherwise there will be a lot more telly than screen! Not sure how I'll construct things but might solder them to PCBs rather than try and get sockets...what do you think? Dom

7th Jul 2011, 1:05 pm	#46
neon indicator Retired Dormant Member Join Date: Jun 2010 Location: Co. Limerick, Ireland. Posts: 1,183	Re: Russian Military Subminature "pencil" valves You can use cut up IC sockets. I'm just soldering. That's what the Russians did and all hobby radios I have seen on web pages (soldering). 1j37b Tests Imagine a 2N3819 FET with two gates. That's what the 1j37b is like! With +45HT via 22K to Anode, G3 to 0V, G2 = HT LT is 1.35V to f+ and f- is via 4.7 Ohm to 0V / LT- / HT- (about Vf 1.08, If = 54mA) Code: G1a G1b Va 0 0 13.5 -0.5 -0.5 13.5 -1.6 -1.6 27.2 0 -2.5 21.0 -2.5 0 24.0 -2.5 -2.5 38.0 0 -5.0 34.96 -5.0 0 36 -5.0 -5.0 45.0

7th Jul 2011, 3:01 pm	#48
dominicbeesley Octode Join Date: Nov 2004 Location: Hebden Bridge, West Yorkshire, UK. Posts: 1,885	Re: Russian Military Subminature "pencil" valves Maybe that's overkill then :-o I think they need to be sat at about 75V +/- 15 IIRC... D

7th Jul 2011, 3:53 pm	#49
neon indicator Retired Dormant Member Join Date: Jun 2010 Location: Co. Limerick, Ireland. Posts: 1,183	Re: Russian Military Subminature "pencil" valves More on the 1j37b "mixer" http://www.radiomuseum.org/forum/spu...ube_radio.html Images in article http://home.comcast.net/~jlrmsousa/1..._audio_vol.gif http://home.comcast.net/~jlrmsousa/s...er_charger.jpg http://home.comcast.net/~jlrmsousa/s...o_labelled.jpg http://home.comcast.net/~jlrmsousa/s...adio_front.jpg http://home.comcast.net/~jlrmsousa/sputnik_blocks1.gif http://home.comcast.net/~jlrmsousa/sputnik_lo.gif http://home.comcast.net/~jlrmsousa/a...rf_mix_if1.gif http://home.comcast.net/~jlrmsousa/s..._ppdet_vol.gif

7th Jul 2011, 4:09 pm	#50
neon indicator Retired Dormant Member Join Date: Jun 2010 Location: Co. Limerick, Ireland. Posts: 1,183	Re: Russian Military Subminature "pencil" valves http://www.radiomuseum.org/forum/rus...e_tubes.html#4

7th Jul 2011, 6:29 pm	#51
kalee20 Dekatron Join Date: Feb 2007 Location: Lynton, N. Devon, UK. Posts: 7,081	Re: Russian Military Subminature "pencil" valves I'm watching this thread, too. These little valves look cute, and they are cheap. I like the idea of a push-pull amplifier using a pair of 1p24b's.

7th Jul 2011, 7:31 pm	#53
neon indicator Retired Dormant Member Join Date: Jun 2010 Location: Co. Limerick, Ireland. Posts: 1,183	Re: Russian Military Subminature "pencil" valves AM signal generator to feed your AM radio Via attenuated aerial connection (do not connect an aerial!) using the 1j37b http://www.radiomuseum.org/forum/15v...ansmitter.html I have now tested the 1j37b in "Gammatron" mode, with g1' (ga) as "grid" and g'' (gb) as "anode" (4k7, 10K, 33k loads). Also vice versa. I only get a gain of 1/2 but it does invert. This is with 0V, -0.5 and -5V "grid" bias and 6V to 20V HT. Actual Anode, G2 and G3 all at 0V. Indeed strange.

7th Jul 2011, 11:40 pm	#54
neon indicator Retired Dormant Member Join Date: Jun 2010 Location: Co. Limerick, Ireland. Posts: 1,183	Re: Russian Military Subminature "pencil" valves 1j37b osc/mixer Adding a MW transistor radio coil with main part as anode load tuned with a 100pF capacitor and separate short winding from g1b to ground and g1a 0V, -0.25 (100k Ohm g1a to OV) gives about 40V peak to peak 790kHz (frequency counter held near) and 0.8mA HT current with 32V HT and 75V ppk with 45V HT, 1.65mA Total HT current (High Q coil I guess!) Adding 200mV 1KHz gives visable AM (but not much modulation depth) on Scope. Confirmed by Yaesu FT817ND radio (whip held next to wiring,) 1kHz tone on 790KHz AM G2 and HT+ is +45V G3 = 0V f+ =LT +1.35 f- via 4.7 Ohm to OV (LT-, HT- common) The coil I had has an isolated winding. If yours has no isolated winding about 10t should be more than enough between G1b and OV. If it doesn't oscillate, then swap the wires on ONE coil only. Don't connect an aerial. It's a frequency changer we are making. I'll look at a better circuit and test with an RF source to get 455kHz, 10.7 and 32.9MHz IFs I'll look tomorrow at Reflex and Regenerative single valve radios using 1j18b as triode, 1j37b as gammatron and as Amp/mixer/osc and 1j24b as regenerative pentode with 1j29b or 1j18b or both audio amp. I've partially soldered up a push pull amp using 2 x 1p24b to give 500mW to 1W on 45V Audio AND 6W to 8W RF simply by switching HT and changing between audio and RF drive. I've the phase splitter to solve A transformer is simplest and lowest power consumption, but I'm not sure I have a suitable one. A small 110/220 mains transformer is Audio O/P. 10:1 ratio is fine. I can put the secondaries in parallel or series and try different loads. On RF it's simply part of the power feed. Last edited by neon indicator; 7th Jul 2011 at 11:56 pm.

8th Jul 2011, 7:06 pm	#55
neon indicator Retired Dormant Member Join Date: Jun 2010 Location: Co. Limerick, Ireland. Posts: 1,183	Re: Russian Military Subminature "pencil" valves 1j18b used in triode AND "gammatron" mode, 12V to 18V HT as reflex regenerative radio... g2 is Triode "plate" for RF and "gamatron grid" for AF g1 is "Grid" 0v is f+ and f- goes to -LT (NiMH battery) via 12 Ohms. Anode is a "gammatron grid" for RF (1M Ohm to Ov!) feedback g1 has main tuned circuit and optional whip via loading coil. about 5 to 10 turns over same coil, or the separate transistor "base" winding on ferrite rod is "tickler" coil in series with 27K Ohm load on g2 with 10nF across it. 1nF from g2 to anode About 50mV AF from RTE on 252 KHz

8th Jul 2011, 7:38 pm	#56
neon indicator Retired Dormant Member Join Date: Jun 2010 Location: Co. Limerick, Ireland. Posts: 1,183	Re: Russian Military Subminature "pencil" valves What does Gammatron mean? Influencing a current on an electrode by voltage on an electrode that is NOT between it and the Cathode. A suitably constructed dual anode "fullwave rectifier" valve can actually be used as Gammatron type amplifier. Use one Anode as a "grid".

8th Jul 2011, 9:01 pm	#57
neon indicator Retired Dormant Member Join Date: Jun 2010 Location: Co. Limerick, Ireland. Posts: 1,183	Re: Russian Military Subminature "pencil" valves 1j18b 2 x PP3 + AA battery radio: Although the bias is much poorer operating current, the gain is much better with g1 referenced to LT- than to LT+, even at 18V HT. I'll do up the schematic of valve amp. 2nd stage is 1j18b as audio amp, maybe. I've no high impedance phones or crystal earpiece, so may add 1j29b as speaker amp.

9th Jul 2011, 10:27 am	#59
neon indicator Retired Dormant Member Join Date: Jun 2010 Location: Co. Limerick, Ireland. Posts: 1,183	Pencil valve lifetime There is no doubt these filaments designed for NiCd rather than Zinc Carbon with thus a min of 0.9V, max of 1.4V and characterised for 1.2V. The Lifetimes quoted (other than insane pulse power 50hours at 400x peak power!) are likely for 1.4V and Maximum Operational Ik as these are Military data sheets. NiMH batteries are up to 1.48V straight from charger. Don't charge NiMH connected to filaments! Then for quite a while are 1.35V dropping slowly to 0.95V sensible end point. So my theory is that a small resistor in series with f- will reduce cold filament inrush and greatly lengthen life. Dropping about 0.2V to 0.25V appears to make little difference to performance, performance drops quickly below 0.8V though. With Alkaline battery the fresh cell is nearly 1.6V and drops much more slowly than Zinc Carbon. Perhaps the series resistor per filament should be larger by x1.5 to x2 with Alkaline. Resistors should be per filament rather than a single resistor in case a filament fails or a valve taken out of circuit. Normally the "bias" should be for Anode current very much less than "Ik max". Code: 1) Ia max is Va Greater than 30V, Vg2 = +45V, Vg3= 0V approx 2) Last column is max recommended "HOT" Class A bias, not ideal bias. 3) Vg= Gnd means "g1" to LT-, resistor Lt- to "f-" and LT+ to "f+" 4) For sensible gain the Anode load should be about x2 to x3 or more the listed value 5) g2 (screen grid) either direct to HT or not more than Anode Load x1.5 and not decoupled (negative feedback) 6) Filament currents approximate Valve Ik Ia abs Ia max Ia Load If Series R Comment Max Max bias Vg=Gnd filament 1j17b 5 1.7 1.2? 1.2? >24k 54 4R7 Not tested much 1j18b 5 1 0.65 0.65 >27k 23 12R 1j24b 1.6 1.5 1.0 0.27 >56k 11 22R 1j29b 8 5.5 4 4 >3k9 53 10R per filament Needs g1 < "f-" by -1V for 2mA 1j37b 4.5 2.4 1.7 1.43 >22k 54 4R7 g1a & g1b 0V to -3V, both -5V cut off 1p24b 25 24 18 15 >820 200 2R2 per filament g1 -5V below f- is about 1mA At 45V HT class A 1p24B about 600mW o/p. At 140V HT a pair in Class AB might do > 6W (not much Voltage gain though)

9th Jul 2011, 10:49 am	#60
neon indicator Retired Dormant Member Join Date: Jun 2010 Location: Co. Limerick, Ireland. Posts: 1,183	Methods of getting g1 more negative. 1) Put two (or more) filaments in series/parallel. A pair of 1j18b are 46mA. The 1j17, 1j29 or 1j37 is about 54mA. These can benefit from a bias about -1V below f- for the g1 (control grid(s) ). Thus you can use 2 x NiMH cells (2V to 2.8V supply) as LT+ the g1 of 1j18b valves and their filaments f- resistors (12 Ohms) go to LT- You need 53-48 = 5mA more, so put 1.35/5mA = 270 Ohms in Parallel. The 1j17, 1j29 or 1j37 f- resistor(s) then go to the f+ of the 1j18b valves and are decoupled with 100uF (3V tantalum min) + 100nF + 1nF to LT- to avoid crosstalk (cathode modulation and negative feedback on the upper valve). The 1j17, 1j29 or 1j37 control grid(s) are taken to LT- The Shield (if fitted) can go to LT- (0V or Ground). G3 can go to 0V(LT-), 1/2 LT+ or LT+ (f+ of the 1j18b valves). Optionally for better Low frequency isolation you could connect the centre tap of the two NiMH cells (or Alkaline) to the f+ of the 1j18b. Similarly you could have two sets of such circuits and use a 1p24b in series filaments mode (centre tap must be decoupled, and f+ of the 1j29b etc) to use 4 x NiMH cells with 100mA consumption for 7 valves 4 x 1j18b 2 x 1j29b (or 1j17b or 1j37b) 1 x 1p24b (filaments in series). The 1p24b g1 to LT- would give then approximately 2mA Ia, instead of over 12mA if connected to "f-" . You could drop the 2 x 270 Ohms on the 4 x 1j18b and add a 1j24b in parallel (11mA + 23 x 4 is close enough to 54mA x 2). 2) Use a PP3 as grid bias. Bias can be about 0.2uA to 2uA per valve. With 4 valves and suitable grid to 0V resistors and 4M7 from Battery to Grid, any voltage from 0V (f-) to -8V can be set. With 6 valves many thousands of hours is possible. In high voltage (145V) class B push pull 2x pp3 might be needed. Have a method to check grid bias supply. On the 1p24b particularly failure is bad. 3) Pick off part of the Local Oscillator (if there is one) and rectify, filter and smooth as a -12V rail. Last edited by neon indicator; 9th Jul 2011 at 10:55 am.