[AMRadio] RE: Testing Transmitting Tubes

Donald Chester k4kyv at hotmail.com
Thu Sep 23 19:50:46 EDT 2004

>Yes, but there are practical issues. This is certainly not a speedy method. 
>Changing tubes out of a working rig puts a good tube in jeopardy of a 
>mis-handling accident. How are results interpreted for a single tube 
>changed into a push-pull final? You have to keep a different transmitter 
>around for each tube type that you want to test (hmmmm, maybe this is an 
>advantage, not a disadvantage....). Etc. Etc.

I once built up a panel with meters, and variable bias supply and plate 
supplies, to simulate actual operation.  The problem is measuring the peak 
emission capability of a tube.  I have seen tubes that seemed to have plenty 
of emission at normal voltages go belly-up on modulation peaks.

When I test tubes that have been sitting for a long time, I go through the 
standard burn-in process with periods of filament-only, filament with 
reduced grid drive, filament with full grid drive plus reduced plate 
voltage, and finally full filament, grid and plate voltage.  Once I have 
determined that the tube appears to be reasonably good, I run the plate 
voltage and grid drive so that the tube operates at about 95% of its rated 
plate dissipation for about 4 hours, then observe the tube again.  I usually 
do this with the test panel, not in a transmitter.  I have seen apparently 
soft tubes test normal after the 4-hour burn.

I'd say it is best to check each tube using several different methods to see 
every possible aspect of the tube, and then you have a fairly good idea of 
its condition.  There is no magic test to plug a tube in and instantly get a 
"good", "fair" or "poor" evaluation for sure.  I have seen tubes that tested 
extremely high, but went completely soft after only a few hours of 

I also have one 833A with a bright shiney spot burnt on the plate, and the 
glass envelope is actually sucked in over that spot.  Obviously someone let 
it run at extreme overload for a substantial length of time.  The glass has 
a silvery-smokey appearance on the inside of the envelope.  Yet that tube 
operates just as well as any of the best, well cared for ones in my 

And yes, I have lost tubes due mishandling.  However, it is still a good 
idea to "rotate spares" occasionally to keep t h e tubes from gassing up, so 
you always run that risk.  I just try to be extra careful when handling 
tubes, including removing them from the sockets and removing grid and plate 
caps.  It is very easy to break a tube at the grid or plate cap if you 
stress it too much.

Ideally, you would need a pulsed power supply and oscilloscope or some kind 
of accurate peak reading instrument, so you could operate a tube at its 
maximum rated voltage times current without exceeding plate dissipation and 
burning it up.  It would then be simple a matter of comparing the observed 
plate voltage/plate current/grid voltage curves with the charts in the 
manufacturer's published specs.

I run pushpull triodes in my rig.  To test one tube, I use a "neutralising 
dummy" in the other socket.  It consists of a defunct tube on which I have 
cut off one of the filament pins.  I insert it in the socket opposite the 
one I want to test.  I have found that you have to back off the loading of 
the transmitter.  Otherwise one good tube tries to run the full plate 
current of a pair of good tubes.  I back off the loading until the plate 
current registers one half the rated plate current.  That way, I can run the 
tube in the final (final running at half power), and check the envelope and 
trapezoid patterns on the scope, and be sure that I am really checking the 
tube in question, without the influence of its mate.

Of course, the same problem exists with a parallel tube amplifier.  The idea 
is to make the amplifier work with just the one tube you want to test at its 
normal parameters.

If you don't have a defunct tube to use as a neutralising dummy, just 
disconnect one of the filament leads to the socket and use a good tube as 
the dummy.  Just make sure the filament by-pass caps stay in place so the rf 
circuitry is unaffected.

I try to keep a log of each tube, and match them as closely as possible in 
the final.  I use a Staedtler Lumocolor black permanent ink overhead 
projector marker to write a number on the tube for positive identification 
later.  I usually mark the number on a plate or grid cap or base of the tube 
in addition to the glass envelope just to make sure it doesn't rub off, but 
I have tubes with still legible markings that I marked over 15 years ago.  
The heat doesn't seem to affect the permanent ink of that brand of marker.

Most of the tubes I keep around and most concerned with testing are the same 
ones I use in my various transmitters, so the availability of a test unit is 
not a problem.


FREE pop-up blocking with the new MSN Toolbar – get it now! 

More information about the AMRadio mailing list

This page last updated 17 Jan 2018.