The question “What to do to turn on an old TWT amplifier that has not run in 1-5-10 years” comes up regularly on various mailing lists. I have been designing TWT amplifiers and power supplies for the best part of the last 30 years, so I have a few ideas of my own about it.

Here it goes:

On TWTs, the cathode runs negative with respect to ground. Most TWTs likely to be of interest to hams run anywhere between -4 and -15 kV. Some run much higher.

This will kill you faster than going to school or to the Post Office.

SSSOOO, be VERY careful before trying to adjust, measure or check anything inside a TWT amplifier.

Old amplifiers may have damaged wiring, and that may not be immediately noticeable. The insulation does not need to be completely gone to strike your hand if it comes close, so BE CAREFUL and KEEP YOUR HANDS OUT.

When trying to bring up an old TWT for the first time, the best thing to start with is run the heater (Standby mode) for a while, an hour or two, before applying the beam voltage. Then a prayer may not be a bad idea, depending on your convictions…

Running the heater for a while will clean the cathode and make it more likely that the beam will be focused. A contaminated cathode has non-uniform emission, and that's bad for focusing. The heater by itself when powered does work a little bit like a vacuum pump, but does not pump much. However it can help a slightly gassy tube turn on. That's another reason why it's a good idea to run the heater by itself for a while on an old tube. Then, when you apply the beam voltage, and assuming the tube does not trip the helix current protection immediately, check the helix current. It will probably be fairly high but should drop over the first few minutes or hours of operation. Do not run any significant RF power until the helix current has dropped to something like 1/4 of full scale or less.

In all cases, safe distance from the equipment is recommended (just kidding.)

If the tube trips immediately, try running the heater only again over night (8 hours or so) and try again. If the tube still trips, there will be little hope you can recover it without more elaborate equipment.

Someone indicated he was advised to “limit the cathode current”. Unless the tube has a grid or a focus electrode with adjustable voltage, the cathode current is not adjustable. For a given tube, the cathode current is only a function of the cathode voltage, and the focusing only works at the nominal cathode voltage. Lower the cathode voltage without refocusing the tube (something that is not easily done at home), and you can kiss the tube goodbye. The worst defocusing is typically around 80% of the nominal voltage, and most TWTs will defocus to the point of damage when the cathode voltage is lowered by 10%, some much less.

If the tube has a Grid or a Focus Electrode, adjusting the voltage down (more negative for an FE, less positive for a Grid) will reduce the beam current without too much effect on focusing and make it more likely that the tube will turn on without tripping the helix current protection. The grid normally runs positive with respect to the cathode (anywhere from +20 to +200V) and the FE normally runs negative (0 to -75V or so). In both cases, reducing the grid or FE voltage by 10-30V will make it much more likely that the tube will turn on. A large negative voltage on the grid or the FE will turn the beam completely off (-250V for a grid, -1200V for an FE), but the tube should not be operated with the grid or FE voltage in-between the ON voltage and the OFF voltage, as it will defocus badly. Of course, actual grid and FE voltages are tube dependent and may vary outside the range I indicated.

When a tube is gassy, even if it turns on, it is very likely to arc during the first few hours of operation (or minutes, or seconds). The power supply should be designed to take the arcs without permanent damage (they occur pretty much regularly in most tubes, particularly the old ones). There is not much you can do to limit the frequency of arcs, and the arcs actually clean the tube inside, so the good news is that if you have arcs, they should decrease in frequency and that would be an indication that the tube is cleaning itself.

The power supply protection against arcs is typically by the way of arc current limiting resistors. These are usually low value non-inductive resistors in series with the HV capacitors (typical values are 10 to 100 Ohms). They have to be of a type that can sustain the full cathode voltage for a brief moment without themselves arcing internally. In older supplies, they were power carbon composition resistors. As these became obsolete, my company ran a lot of tests to find a suitable replacement. We have found that certain types of MIL-SPEC power wire wound resistors were perfectly acceptable, so that's what we now use. However, be careful that most wire wound resistors will NOT be a good substitute for carbon composition. Nowadays, there are other types of composition resistors that are available, and I suggest you look at the usual distributors (Digikey, Allied, Mouser) for those.

When trying to recover a tube with excessive helix current, the tube manufacturers modulate the beam current (using a grid, FE or cathode modulator) so that the tube runs only at 0.1% or less duty cycle (like 1uS pulse width at 1 kHz). At that low duty cycle, the tube can survive being very badly focused while the tube specialists refocus the beam. Sometimes, simply running the tube at low duty cycle for a while makes the helix current drop. That's because the electron beam works like a pump and force the atoms floating inside the tube, in what should be vacuum to embed themselves in the cathode. As the helix current drops, the duty cycle can be increased progressively over a period of many hours and sometimes days.

If the tube has a grid or a focus electrode, a modulator could be used relatively easily to modulate the beam, but if not, you need a cathode modulator, and that's another story.

I have heard people say they were told to never run a TWT without RF. This has no basis in reality with tubes made in the last 30 years. All TWTs are designed to run with or without RF.

There are 3 possible reasons why an old tube may run better with some RF, and neither should affect a tube made in the last 20 years that is in reasonably good condition.

1. the tube has more gain (no gain compression) with no drive and tubes that are marginally stable because of age or damage are more likely to oscillate with very low or no RF. This may be a problem on a development tube, but these problems should be resolved before the tube goes into production.
2. with no drive, the beam remains focused all the way to the collector, and on single collector tubes, the collector dissipation is maximum with no drive (class A amplifier), so if the cooling is marginal, indeed no drive is not good.
3. with a little bit of RF drive, the beam will spread a little bit and if the tube is not well focused, it is less likely that a significant portion of the beam current will fall on one single helix turn, which is the worst thing that can happen to a TWT. I have known tube engineers who always brought new tubes up with some RF and some who did not, so the jury is still out on that one as far as I am concerned. If you feel better with some drive, drive the tube with at most -10dBm at the center of the frequency band.

I have been dealing with TWTs professionally for about 28 years and I have not come across a production tube that did not like no drive, as long as it was cooled properly and stable.

Particularly when bringing up a tube that has not run in a while, I recommend not to run RF, as it will cause an increase in helix current, and if the tube is not well focused to begin with, it will simply make matters worse and the tube more likely to trip.

Of course, like any power amplifier, input and output should always be terminated. TWTs being generally very broad band amplifiers, the loads should be matched over a large band as well. Even if you do not intend to run full power to begin with, the output termination should be rated for the full power as the tube may oscillate.

If you do not know the gain and bandwidth specification for the tube you are operating, start with -10 dBm drive and check the frequency response and gain. Very few power TWTs will get near saturation with -10 dBm. From there, you can go to the point of highest gain and look for saturation. Do not exceed the drive level that gives saturation. Not only you won't get more power, but a number of bad things can happen in the tube if you do that for a while.

Typical .141 semi-rigid coax is good for at least 20W at 10 GHz in ham applications (intermittent operation like CW or SSB), as long as the run is not too long and the connectors are good quality. I have run 50W at Ku band (14 GHz) in a short piece for a short time because I had to, but I definitely do not recommend it. I have tested tubes capable of 100W at Ku band that were designed with SMA connectors at the output, and that's a recipe for disaster, regardless of the cable used (in my case, the SMA was going directly into a waveguide transition, but even then, the specially made SMA connector did not survive, what were they thinking?)

Didier KO4BB