Testing low voltage pentodes without a curve tracer?

[tss]

I have bunch of different low voltage tubes I got from eBay over the years and never used them for anything. What I would like to do is to setup a simple jig to test them.

My initial idea was to test the tubes wired as triodes. See the circuit idea below, this is for some mini russian tubes I have - not exactly sure of the model. The idea is to get an IV graph in order to calculate the tube parameters and use it in a pre-amp circuit.

A DMM will be inserted in series from the power supply to the resistor to measure the current, the anode voltage will be read using a 2nd DMM. Vbias is set to a constant voltage and B+ is stepped from 0 to 30V. This process is repeated after changing vbias. The readings are entered into a spreadsheet and the dots are connected to create the VI curve.

Question - what should be the resistor value? Thoughts about the testing method?

[merlinb]

I would try to keep the anode dissipation under 0.5W. If you use a 470R resistor or larger then the dissipation can't possibly exceed this figure, so you're less likely to burn anything out by accident.
If you use 1k then the voltage across it will be numerically equal to the current through it in milliamps, which you might find useful when working.

[tss]

That's a good idea, easier to measure voltage than current and probably more accurate using the test gear I have.

And I need to watch that Vr * (Vcc - Vr) < 500 you say?

[tss]

OK this is a bit weird... no matter what bias voltage I apply, the VI curve seems the same. How can this be? Something is probably wired up the wrong way?
R=1k

[merlinb]

OK this is a bit weird... no matter what bias voltage I apply, the VI curve seems the same. How can this be? Something is probably wired up the wrong way?
R=1k

If you're not sure what model tube it is, how do you know you've got the connections right?

[tss]

I've wrote it down on a paper and stored with the tubes  
It's 1zh24B - pin out is OK, just checked according to this:

[tss]

Here is a look on the connections:

[merlinb]

I've not used pencil tubes before, but it looks to me like you're supposed to ground pin 6 and only connect pin 3/4 to anode if you want triode mode.

[tss]

3 is a shield "thingy" it should be left floating or connected to ground, however I will try to ground pin 6

[greaser_au]

I'm probably going to look like an idiot (and I do not mind that, it is frequently educational), but it looks to me like this is a directly heated cathode valve... I also hope I'm not missing the point of your thread (if so tell me and I'll get back in my box) ...

First thing:  obligatory caution...  warning - high voltage. Even 90VDC can be as deadly as the mains.

Bear in mind that the circuit shown is a very digital application and (if I'm right about the type) in this case the NOR gate inputs (A,B) are on the screen grids. The digital input voltages are 50V (the valve should saturate) and 0V (which would drive the valve into sharp cutoff). with the plates tied together, if either valve has a logic high (50V) on the screen, the plates (Q)   are both low (<3V). the clue is in the text:
output swing +3V to +90V No load          
+4V to 50V with 3 screen loads and a
single grid drive to 50V.

Gut feeling for more sensible tracing  based on the circuit (but I take no responsibility for any destruction that ensues, please do your own research):
1- 0V (heater)
2- 1.25Vdc (heater/cathode)
3,4-  0V (suppressor grid)
5-  50V  (screen grid)    (tie to B+ via a large resistor  don't undershoot this you WILL burn the screen - typically the Vscreen approaches (but not exceeds) the Vanode.)
6- grid leak (1M???) to ground - input (control grid)
T - 100k to 90V - output (anode)

this likely will not get you where you want to go, but - based on the circuit shown you likely will not destroy it. but your 'curve trace' should get you  in the ballpark.

EDIT:  Found these and these support my assertion - though there is a circuit seemingly using the valve configured similarly to the NOR gate as a preamp:
a diyaudio page: http://www.diyaudio.com/forums/tubes-valves/256429-russian-1zh24b-pentods-design-assistance.html

Datasheet in Cyrillic russian  http://www.radiomuseum.org/tubes/tube_1j24b.html)   use the pin names (Anod, Katod, Hakap, setka) to figure out the characteristics bottom left of page 1  -  note 'B' = V (volts) - maybe one of our russian friends here would care to translate?.

Application discussion of a bunch of different types including this one: http://www.radiomuseum.org/forum/russian_subminiature_tubes.html?thread_id=200277#13

david

[tss]

Thank you for the detailed answer, David.

I'm not planning to run the tube on high voltage, actually I don't have a power supply that goes past 30V. From what I've heard, these tubes can be operational at lower voltages.
This is indeed a directly heated cathode.

So 1 + 2 are the heater supplies.
3 + 4 goes to GND
5 - I did not understand exactly. Why is a large resistor needed? Is it connected to B+ through the said resistor?
6 - Signal input via 1M to GND?
T - 100k to anode, why so large?

Thanks!

[greaser_au]

So 1 + 2 are the heater supplies.
3 + 4 goes to GND
5 - I did not understand exactly. Why is a large resistor needed? Is it connected to B+ through the said resistor?
6 - Signal input via 1M to GND?
T - 100k to anode, why so large?

What I posted was just guesstimates from the NOR gate circuit values and my exposure to 'standard' valve designs - intending to take this towards linear operation - at that point I didn't have a datasheet or linear circuit.  
When I found the oscillator/mixer here:  http://www.radiomuseum.org/tubes/tube_1j24b.html it kinda validated it, except the screen resistor is 2.2M  beware!!!!!  and, of course,  being a mixer, it is intended to be somewhat less than linear.

5: sorry I missed the words to "tie to B+" for the screen grid via the resistor (I have updated - and suggesting a large value resistor).

6: Basically the cathode is biased to 1.2V by the DC cathode supply. For negative bias,  the control grid needs to be tied to ground, but this would ground your signal also- so a large resistor (in that circuit: 1.5M) ' a grid leak resistor' ties it to ground and your AC signal drives the grid directly - testing with a curve tracing setup you should not need it, but you would in an audio arrangement.

A: just what was there. Merlinb's comments here are likely to lead you along the right path.

have a look at Joe Sousa's posts:
"1zh24b IF gain with reduced filament voltage"http://www.radiomuseum.org/forum/russian_subminiature_tubes.html?thread_id=200277#12 for a linear circuit
and
"1zh24b DC operation at low voltage" http://www.radiomuseum.org/forum/russian_subminiature_tubes.html?thread_id=200277#13 for curve traces using that circuit with varying values of grid voltages in pentode mode.

david  

[PRR]

Apologies if you know this and the image just confused me....

This drawing shows G2 driven.

We usually set G1 as input. The grid *closest* to the Cathode. The one with the most control. The other grids then connect to Plate to make "triode mode".

[tss]

Paul, I've actually seen applications using g1 and g2 as the input (and even saw one with the output taken from g2!). So you are suggesting to to connect it like in the drawings below? I'll try that today, as last time I couldn't get see anything changing in the IV curve depending on the "bias" voltage of g2.

[tubegeek]

Triode mode would have the screen grid (g2) connected to the plate and the suppressor (g3) connected to the cathode. g1 is input. Watch out for bias on the DH filament: you will have a DC voltage of an average of +0.6V on the filament the way you show it, and your bias will come from the g1 being negative with respect to that voltage, you also show that in your diagram so I think you've got that covered.

TL;DR: connect g3 to the cathode.

Also: in many pentodes, g3 is already connected top the cathode and you have no choice.

The purpose of the suppressor grid is to knock the electrons that bounce off the plate (secondary emission) back towards the plate by repelling them (the plate is much more attractive to the electrons, being positively charged.)

[PRR]

RCA had a tube where you connected different grids different ways to get very different characteristics.

And yes, you can drive G2. But in a standard-design Pentode you have to drive it very hard with a positive DC voltage.

I don't see why you are questioning driving G1 with your small signal-- that's the most usual way, because it gives the best leverage against the electron-flow off the Cathode.

BTW the usual plan is to drive G1 *negative*. If it goes positive it sucks signal current, which means much less power amplification. In filimentary tubes it is a little complicated because the "cathode" (filament) has a significant DC voltage drop along it. They usually scaled these so with a half-used battery, 1.2V, the +0.6V "average" along the filament, against a ground-biased G1, gave a happy negative bias. But that's for 45V-90V operation. With a puny 24V you may have to dink around to find a useful range.

If you are connecting the other bits "Triode", then where you take G3 is probably moot. When you run "normal" with G2 high and steady, and large output swing at Plate, G3 near cathode/ground bias soothes secondary electrons. But G2 and Plate tied together, this is non-issue. There might be a dissipation consequence, but at 24V you really can't burn it up.

[tss]

Paul, I am not questioning the connections. I am only sharing that I've seen this tube used differently, I don't understand why is that as I am sure one way is better than others. The data sheet is in Russian but does not offer any additional info on circuit application. Thanks!

[merlinb]

Paul, I am not questioning the connections. I am only sharing that I've seen this tube used differently, I don't understand why is that as I am sure one way is better than others. The data sheet is in Russian but does not offer any additional info on circuit application. Thanks!

I assumed it was a space-charge tube which do indeed use the second grid for control. However, googling around I too discover that this particular tube seems to be used both ways.

[PRR]

It is actually a fairly unique tube. Joe Sousa has tested it every which way-- his mentor's motto was "Throw it on the curve tracer!" Joe has a research-class curve tracer, and is a fairly friendly guy. He isn't too-too hard to track down (hint: www.philbrickarchive.org ).

[tss]

I've actually seen a long discussion about the tube (among others similar ones) on the radio museum or tube museum (can't remember which) website but could not understand how the test jig is hooked up.