A klystron has multiple resonant cavities with a coaxial electron beam. The electron beam causes the cavities to ring-- thereby transferring energy from the beam to the electromagnetic wave propagating down the slow wave ( group velocity much less than c) transmission structure comprised of the series strings of resonant cavities.
In short, tune all the cavities to about the same frequency and yo have a very high gain, very efficient narrowband amplifier.
Stagger tune the cavities and you get a broader band amplifier with less gain and lower power conversion efficiency.
In 1977 the Russians were running klystron efficiencies in the 60 to 80% range. Note that both electrical and power efficiencies were measured. Here is an english translation:
http://www.slac.stanford.edu/cgi-wrap/g ... s-0181.pdf
Consider a conventional waveguide with a rectangular cross-section
Constricting the wide dimension with a very thin conductive sheet introduces reflections such as those that would be developed by inserting a shunt inductance at the iris location.
Constricting the narrow waveguide dimension is the same as introducing a shunt capacitance.
Constricting both dimensions is the same as introducing both a shunt capacitor and a shunt inductor. Depending on the geometry of the remaining aperture , the effect is the same as introducing a shunt connected, series resonant circuit, or a shunt connected parallel resonant circuit.
When inserting multiple irises, the distance between the irises and the characteristic impedance of the guide between the irises inserts an equivalent series inductor or series capacitance.
So irises or in the limiting case simple conductive protrusions such as a threaded screw allow one to construct the equivalent of lumped constant impedance matching and or filtering networks.
One old trick is to tune a radar antenna for minimum vswr by taking a ball pean hammer to the waveguide. A couple dents in the right place and you are good to go.
The internet and Google are a marvelous research tool:
Waveguides Impedance Matching
COAX TRANSMISSION LINE (NASA) vs WAVEGUIDE (SHAWYER/NWPU):
I suspect that the waveguide approach even with its attendant tuner, isolator & iris would cause less spurious thrust signals that a piece of coax which could be leaky and cause spurious thrusts due to EM interactions with local fixed pieces of apparatus. I wish you luck in getting all that 340 waveguide gear along with the frustrum thruster onto a balance!
The magnitude of the time rate of change of power (dP/dt) to the test article can be much greater using a magnetron and WR340 waveguide verses a narrow-band solid-state microwave source tied to an RG-142 coaxial cable. (A fire hose compared to a straw.) And since we all know by now that dP/dt is king in this business, it’s worth the time and money to investigate this angle of the problem
Yes waveguide can handle large high power levels. WR340 waveguide power level vastly exceeds the power level rating of RG-142 however, if you do a waveguide to coax power rating comparison, to be fair the coax should have about the same cross-sectional area as the waveguide and should have the same dielectric- ie air or largely air.
In such a comparison coax still has a lower power rating than waveguide but only by a factor of 4 or so. The limitation in both cases is the same, dielectric breakdown - arcing) and conductor overheating
However I think coax is still worth a consideration as i don't think you will be using a multi-megawatt source and the practical advantages are significant. Still I agree with you waveguide is better with respect to insertion loss and power rating but coax is much much cheaper, much much more flexible, much much lighter and much easier to work with.
There are coaxial cables much better than RG142 but still fairly ordinary that work well at 2 GHz. Look for cable with air dielectric.
Another cautionary note is the power rating specification you quote for WG340 and RG-142 both are I assume accurate but do not apply to your application. Both waveguide and coax and as far as i know all transmission line structures have maximum power rating that only apply when operating with a 1;1 VSWR. For example if you have a waveguide or a coaxial line- load combination that has a 2:1 VSWR, the maximum power rating drops by at least a factor of four because of the presence of a standing wave on the line that creates two to one voltage amplitude peaks and two to one current ratio peaks. Power handling limitations come from dielectric breakdown and conductor overheating (I^2*R) A 2:1 vswr requires a factor or four reduction in power to stay within the dielectric break down limits and the conductor heating limits.
Necessity of 3 stub tuner: consider the 3 stub tuner as tuning (minimizing the VSWR) the section of waveguide leading to its junction with the cavity as if this junction looks like a lossy (& phase rotating) absorber to the waveguide. Re: necessity of iris at the junction of waveguide and cavity: consider the difference in geometry, ie impedance between waveguide and cavity. Apart from the simple necessity of impedance matching, you also want to maximize the one-way power flow into the cavity from the 'source' ie the waveguide. The iris acts essentially as an shunt-connected radiating element into the cavity, its efficiency given primarily by its geometry. We've never been satisfied with calculations on iris geometry versus waveguide & cavity geometry so always made variable-sized slit irises which could be manually manipulated. The one in the photos is 1 cm wide on a 1/8" thick brass plate. We've also used a simple nylon threaded rod one end of which is able to be screwed across a simple constant-area iris to vary the VSWR.
Here is a microwave waveguide filter and impedance matching text that might well be useful.
- Matthaei, George L.; Jones, E. L.; Young, Leo (1980). Microwave filters, impedance-matching networks, and coupling structures. Dedham, Mass: Artech House Books. ISBN 0-89006-099-1.
This is an absolute classic, a canonical text. IIRC correctly it is a collection of papers and research reports from the surge of post WW2 radar work.
There is a chapter in this text that I think might well be particular relevant, a chapter discussing how to generate extremely high microwave power levels for testing of waveguide windows. Windows in this context are insulating sheets crossing the waveguide aperture allowing pressurization - or evacuation- of the waveguide structure.
The trick was to make a closed loop of waveguide and feed that loop with a directional coupler. The point is the waveguide loop was a resonant structure having a high Q and therefore having the equivalent of very high circulating currents. This way a 10 kW test source could test waveguide windows at equivalent power levels of several megawatts.
Seems to me there may be some lessons to learn from that. One thing that comes to mind is maybe the resonant cavity EM thrust device could better be treated as a two port thereby easing the impedance matching requirements.
You might want to pick up a copy of Skolnik (MIT) books:
- Introduction to Radar Systems, McGraw-Hill (2002)
- Radar Handbook (3rd ed.)
Mark Richards (Georgia Tech) is the other prolific author I would trust, but Skolnik's Radar Handbook is required for your level of work.