wideband traveling wave tube amplifiers (understanding specs)

[SArepairman]

So I am interested in a helix TWT that is wide band (I was hoping for a 2 octave one, as the wikipedia says exist). I wanted to design my own TWT amplifier including PSU and protection circuits, and then hopefully a transmitter to connect to it.

Some examples:
http://www.ebay.com/itm/Litton-Electron-Traveling-Helix-Wave-Tube-Amplifier-L-5967-K-Band-to-GND-1-3900-/130435966117?pt=LH_DefaultDomain_0&hash=item1e5e96e4a5
what is kband to ground???

this would be a 20-40GHz 1 octave TWT?

http://www.ebay.com/itm/L3-Comm-Electron-Helix-Wave-Tube-Amplifier-L-6073-00-K-Band-to-GND-1-3900-/130435967862?pt=LH_DefaultDomain_0&hash=item1e5e96eb76
same?

How do you read the frequency specs on this thing? How expenisve/hard to find would it be to get a 1-4GHz (2 octave) helix TWT? how about 2-8 GHz (also 2 octave)?

I'm more interested in a TWT transmitter, but recievers are interesting too, also, how frequency agile are these things? what is the speed limit for scanning the output across their bandwidth?

also, what does price vs power level look like for these guys? im guessing the ones I linked are <100W.

What does a 2KW 1-4GHz TWT cost?

[T3sl4co1l]

I know very little about these, so I can't be of much help;

What do you mean by "speed limit"?  If it were a tuned amplifier, it would have to track the frequency the whole way.  A self tracking tuned amplifier would be amazing, and such a thing either exists or can be made, but is very slow and ponderously expensive (phase tracking circuits + servo drives + vacuum variable capacitors!).  Slow as in, takes seconds to track.

A true wideband amplifier simply amplifies what goes through it, so there is no frequency-time trade-off, it's all frequencies [in the passband], all the time.

Now, I understand a TWT is dispersive (a consequence of the helical transmission line), which means you won't get phase coherence -- a square pulse in will end up as ripples (much the same as ripples on a pond -- gravity waves are dispersive).  The frequency domain components will end up amplified more or less correctly (given the peaks and dips characteristic of the amplifier), but the phase shift or time delay will vary significantly with respect to frequency.

If you're looking to do something like wideband pulse or burst RADAR, you'll probably have a hard time.  If you're doing chirped, you'll need to calibrate against the chirp characteristic of the tube itself, but given that, it should do just fine.

So, it depends what you want to do with it.

Tim

[SArepairman]

is the dipersion effect in any way related to operating power? i assume not (as you say its a t-line effect) but its voodoo to me

[T3sl4co1l]

I wouldn't be surprised if it varies a bit, but probably more dramatically so with bias voltage or focus (electrons being what they are).

I recall reading that identical tubes can be well enough matched (mechanically and electrically) to use power combiners for greater power output.  This requires phase errors under 90 degrees to be efficient, so for there being, what, a few hundred cycles worth of phase shift down each one, that's not too bad (<1% phase error?).

Tim