UnKnown Triode Transmitting Tube.

[W2PFY]

Suppose you came across a triode transmitting tube that is so obscure that information cannot be found on it. You think the filament voltage is somewhere between 5-10 volts. You can guess that the plate voltage could range from 1500 to 3000 volts. Where would you start if you wanted this tube or tubes to be run in class B or C?

There must be an old buzzard way to set up or find initial characteristics for something unknown?

[Tom WA3KLR]

If you got some numbers off of the envelope or base already, share them.  Someone may find info. 

Also if you could post some close-up photos with a ruler or state dimensions, that may help someone to identify the tube.

Electrical measurements could be made if nothing else once the proper filament voltage determined - how to determine type of filament?  Create grid voltage, grid current  & plate current plots, pulse measurements of max plate current, plate dissipation empirically determined (incipient plate glow), etc.  Not worthwhile though unless you have a few of them, in my opinion.  Interesting exercise - for geek. 

[k4kyv]

If you have worked with a variety of transmitting tubes, you can get a rough estimate of the plate dissipation and rated plate voltage by looking at the size and shape of the plate structure.  Short the grid to the plate to form a diode.  Connect the plate to an a.c. supply, such as a plate transformer at somewhat less than the estimated plate voltage (or a variable HV DC supply if you have one), through a current limiting resistor and a DC milliammeter, returned to the filament.  Use a filament transformer fed through a variac.  With the voltage applied, the tube will act as a rectifier, and the DC plate current will be indicated by the current through the meter.  Starting at zero, gradually run up the filament voltage until you reach a point where the plate current no longer increases with increased filament voltage.  This should be roughly the rated filament voltage.  Keep these readings momentary, to avoid overheating the plate, also watching for plate glow.

You could then experiment with grid bias voltage to determine the cut-off voltage, and amplification factor, and static plate current at a given voltage.  Then maybe throw together a triode oscillator circuit or class-C amplifier into a dummy load to refine the measurements for the optimum filament voltage, grid bias and grid bias.  If this is done carefully, these tests should be possible without damaging the tube.

With more sophisticated test equipment, you could measure the characteristic curves of the tube and compare those to known tubes, or design a circuit around the results of your experimental curves.

If it is really a good tube in good operating condx, it should be possible and worthwhile to experimentally figure out how to use it.

[WA1GFZ]

Sounds like the day I bought 7- 4CX3000A for $100 and didn't have any data for a month.

[W3SLK]

Don said:

If you have worked with a variety of transmitting tubes, you can get a rough estimate of the plate dissipation and rated plate voltage by looking at the size and shape of the plate structure.  Short the grid to the plate to form a diode.  Connect the plate to an a.c. supply, such as a plate transformer at somewhat less than the estimated plate voltage (or a variable HV DC supply if you have one), through a current limiting resistor and a DC milliammeter, returned to the filament.  Use a filament transformer fed through a variac.  With the voltage applied, the tube will act as a rectifier, and the DC plate current will be indicated by the current through the meter.  Starting at zero, gradually run up the filament voltage until you reach a point where the plate current no longer increases with increased filament voltage.  This should be roughly the rated filament voltage.  Keep these readings momentary, to avoid overheating the plate, also watching for plate glow.

I think it was John WA4DWW, who bought an Amperex AX9903(?) and did that. To this day, he and I couldn't find any data on it. He used it in his HB rig until just recently when he said the power was begining to roll off.

[KE6DF]

Attached is some data on the Amperex 9900 series of tubes.

Note that I found another document that said the 5894 was another number for the 9903 -- but this may not be the tube you were talking about as it's a fairly small tube.

Perhaps you meant one of the other 9900's.

Dave

[W2PFY]

Thanks Don & guys, This is for the GL-434A tube. I'll need to set up a blower for it as it has an external anode. I thought  about the way Don suggested would be the way to go. Since I never tried anything like this, it's great to know there was vitrifaction to a plan I didn't know would work. If I started this myself, I believe the note still would have gone out to be sure it was the right way to go.

Thanks   

[Tom WA3KLR]

Now that we know it is a GL-434 – I see that this tube was covered before in the AM Forum:

http://amfone.net/Amforum/index.php?topic=19143.0

http://amfone.net/Amforum/index.php?topic=18648.0

A photo of the GL-434 that was posted in that thread:

http://home.comcast.net/~n6jv/gl434.html

“Tube Lore” by Ludwell Sibley says it is also known as a ZP434.

Nobody came up with specs.  You know this, now I know "the rest of the story".

[Tom WA3KLR]

Use a filament transformer fed through a variac.  With the voltage applied, the tube will act as a rectifier, and the DC plate current will be indicated by the current through the meter.  Starting at zero, gradually run up the filament voltage until you reach a point where the plate current no longer increases with increased filament voltage.  This should be roughly the rated filament voltage. 

Don,

This is incorrect.  I see no basis in physics for what you state that the tube current will level with increasing filament voltage.  The physics of a hot cathode has increasing current output with increasing temperature.  This is a high order monotonic function.  The cathode temperature keeps rising with increased filament power.  In fact as the temperature gets higher, the milliamperes of emission per watt of filament power increases also. 

Following your procedure the filament will probably get damaged. 

Perhaps you are thinking of a constant filament voltage and loading the plate circuit/increasing plate voltage until the saturation current is seen?

[Tom WA3KLR]

I’ve been skimming through Chapter 4 of “The Thermionic Vacuum tube and It’s Applications” by H. J. Van der Bijl, M.A., PhD., a classic in the field.  I read through it many years ago.  The effect I am referring to with no current limiting mechanism is associated with an equation called Richardson's equation.

Yes Don, the phenomenon of cathode temperature saturation can occur, you are right there, I didn't remember this part.  A space charge can build up causing a current limiting mechanism.  However, it all depends on a given tube’s designed cathode temperature and plate voltage applied as to whether the plate current will seen to level off with further increase of cathode temperature.  The number of electrons emitted from the cathode does increase with temperature but an equilibrium is formed by the space charge; the additional electron flow returns to the cathode in the cathode temperature saturation mode.   

We are both talking about different portions of the same family of curves.  Never having explored this, I don’t know if this cathode temperature saturation is experienced in most tube applications.

Out of curiosity, I will find a good triode tube here (but not a big transmitting triode!) and do the filament voltage versus plate current test on it to the point of filament burn-out and post an Excel graph.  It will be interesting to see what I come up with.

[Rob K2CU]

Tom,

I agree with you that it is very risky to attempt to derive design paramters for filament voltage by that method. I suppose that there is a difference between directly and indirectly heated cathodes, and I suspect that this big tube is a directly heated type. These things are also big  on current. The 3CX3000's that I have on the shelf are something like 7.5V @51A. I believe that they use low voltage to minimize the violtage gradient along the filament. But, consider this: If you put some small tube in a tube tester and then adjust load current so that it is just at the Bad/Good threshold on the meter, or even below, just step the filament voltage by one click for a short period and watch the current rise up. Tube testers usually measure the emission and you can play with the filament voltage and see how it affects the current as displayed on the meter. you can even see how cooking a "bad" tube can actually restore it some what when you drop back to the correct filament voltage.

probably the bigger issue is what type of service the tube was designed for?

but, it is an intersting mystery.

can always use it as a lamp stand.

[KE6DF]

The page below is for the 7c29 which is said to be a successor to the GL-434a

http://www.radiomuseum.org/tubes/tube_7c29.html

Was used by Transmittertube (Radio)
Filament -: Direct / If: 28 Ampere / Vf 10.5 Volts / 
Description VHF, integral radiator transmitting triode. Thoriated tungsten filament. Mu 29, 500W anode dissipation, 110MHz full ratings. Forced air cooling required.
Probably derived from GL-434A.   

[Tom WA3KLR]

Thanks for the data Dave.
- - - - - - - - - - - - - - - - - - - - - -

I did a filament voltage versus plate current test on an indirectly-heated triode tube here today. 
I tied the grid to the cathode. 
Plate voltage was 300 Vdc from a regulated lab supply.
The filament voltage was from a dc lab supply. 
Starting voltage for the filament was zero Volts. 
The new filament voltage setting was run for one minute before the plate current meter was momentarily connected to the tube plate connection.

Below is a graph of the filament voltage versus plate current.  For now, the actual voltages and currents are hidden, scaled into Tommyvolts and Tommyamps.

What is your guess as to the rated filament voltage in Tommyvolts?  Correct answer will be given later.

[KE6DF]

OK, my guess is that the rated filament voltage was 60 tommyvolts.

[Rob K2CU]

Tommy,

I'm guessing 25 to 50. It would interesting to see a curve for a thoriated tungsten (directly heated) filament too!

[AB2EZ]

I agree with Don et al., that determining the correct filament voltage is critical.

This article:

http://www.w8ji.com/vacuum_tubes_and_vaccum_tube_failures.htm

contains a chart of the proper operating temperature for popular cathode types (which, as Tom pointed out, determines the rate at which electrons are produced that have sufficient kinetic energy to leave the surface of the cathode)

I think that, given the effect of too high or too low a temperature on the lifetime of the cathode, the only reliable way to "guess" the proper filament voltage would be to observe the temperature of the cathode v. the applied filament voltage.

If the cathode is visible, one could use the blackbody radiation formula (color spectrum) to observe the temperature of the cathode. Comparing the spectrum of the radiation from the unknown tube to that of a known tube with the same type of cathode might be a good way to calibrate the measurement.

This, of course, is a lot of effort to expend to get to a position where one could use the tube.

Stu

 

[Tom WA3KLR]

My filament voltage versus plate current experiment results:

The rated filament voltage is 31.5 Tommyvolts or 6.3 Volts; the Volts was scaled up by a factor of 5.

The tube tested was a metal octal 6SQ7 triode/duo-diode.  This is an indirectly-heated oxide-coated cathode.  Test set-up photo below.  The filament has not been blown.  With 17 Volts applied to the filament, the fuse popped in the plate current meter when touched to the plate lead.  The cathode and grid structure probably warped to the point where the grid was shorted to the plate for the next application of the plate voltage.  The filament current was about 0.55 Amps, so the dissipation was almost 5 times the proper value (9.35 W/1.89 W).

In the 1940 RCA Vacuum Tube Design Lectures book available from the pmillett site, Lecture 3 page 25 lists the 6K7 ( a similar tube) heater temperature:

3.0 volts  817 C.  (1090 K.)
4.5 Volts  987 C.  (1260 K.)
5.5 Volts  1092 C. (1365 K.)
6.3 Volts  1167 C. (1440 K.)
7.5 Volts  1287 C. (1560 K.)
14.0 Volts  1725 C. (1998 K.)

Below is the test results graph in actual Volts and milliamperes.

[Tom WA3KLR]

Yep Stu, if you can see the filament, this is real good. 

Bright red is 1200k
yellowish red is 1400 K
incipient white is 1600 K
white hot 1800K and up.

The old directly-heated cathode designs had the possibility of being run as high as 2500K but its a choice between the increased emission efficiency and plunging lifetime.

[Tom WA3KLR]

A second 6SQ7 has volunteered to take part in the cathode temperature saturation test.  The purpose of these tests has been to see if the rated filament voltage can be determined by this kind of experiment.  The 6SQ7 has an indirectly-heated oxide cathode rather than the thoriated Tungsten directly-heated cathode tube that started this thread.  The test has been re-run at a much lower plate voltage, 100 Volts, in an effort to show the saturation effect.  The first run was at 300 Volts.

The transconductance of the second tube is a little lower than the first; tube 1 = 1000 uS., tube 2 = 850 uS.; minimum specification for the 6SQ7 is 750 uS.

The current saturation was not seen.  See updated Excel graph below:

[W2VW]

Secondary grid emission is a good indicator of poisoning from cathode over emission in indirectly heater power toobs. This process takes time. Normal tubes with high hours may act similarly.

Some large toobs take up to a week in continuous operation to show grid current changes from filament wattage reduction. The balance is between peak power and cathode material finding it's way to the grid. This may be one reason MRI tubes get pulled with so much life left. 

[KM1H]

The AX-9903 is indeed the daddy of the 5894. It was an updated 829B but the plate pin seals fractured under any strain in the intended mobile service so the 5894 quickly replaced it.

Carl
KM1H