Help with 2.45GHz high power application. 3

HP has sold HFSS to Ansoft. There is a free student version available (don't know how crippled it is).


Then click "contact."

Hmmm, interesting - have to think up a good commercial justification for "trying out" this code...

***I understand that Ansoft and HP HFSS are the industry standard for EM simulation

This may sound like a statement from a commercial issued by a respective software vendor, but unlikely represents a real engineering practice. This is no similarity to MS Windows here. HFSS is based on FEM, but modern computational electromagnetics has real alternatives to this method, for instance, FDTD. Many researchers and engineergs find the latter one to be more efficient/convenient for designing systems and components in many microwave applicaitons. In particular, for microwave power engineering, the benefits of FDTD modeling (and respective simulators) have been explicitly formulated - see, e.g.,

I do not think HFSS will be of real help in this case as is never really seemed to model loss in systems all that well, or at least when I tried, It took to long and I was skeptical of the results. You are not very clear of you application, when I was A pipe weld grinder in wyoming, they would always have the welds X-rayed for defects. Is that what you are after? I do not think that microwaves have the resolution, other, than to tell that a welded section is maybe present, Now good weld from bad weld? that is a stretch, I know that PHd's need to find a new nitche to hang your hat, but this is going to be very tough. I would suggest a wide band linear FM or psuedorandom signal such that pulse compression can be employed to maybe, maybe get your resolution sub-inch. This would seem your only hope, I think, However really, I know nothing of nuclear magnetic resonant imagining, if which this is. Forget every thing I just said, and laugh that I just made an ass of myself again, and have fun with you superconduting magnets.

Thanks for the info IMMG. Are there any codes you recommend particularly? Again cash is tight so just something for an initial looksee will do me.

Thanks for the ideas GOTWW. Sorry I can't say more about my application. May be subject to intellectual property rights, and is likely to generate a patent. By disclosing too much information I shoot myself in the foot.

I guess the question at this point is what do you want to do?, simulate the fields in a cavity, coaxial resonator, transmission line or what. As I said, HFSS is somewhat limited in evaluating the skin effects, or discontinuities of conductance. Basically S-parameters, and modal field distributions. The conductivity, skin effects, & associated loss is troublesome for me with the simulators, especially if I have to wait. If your goal is to basically establish a resonant scenerio with alot of surface current, just dig into the old rad-labs, etc. Pick a mode that produces J, that suits your need.

Simulation of fields in a cavity is a good summation. The ability to, say, predict interaction of an aerial with the cavity while simulating skin effect would also be useful. At this point I'm assuming (perhaps falsely) that I won't have to simulate the generator and feedlines. Smith chart manipulation should be good enough for this.

Most of it, as you say, could probably done by picking the appropriate mode and assuming that this will be dominant. Without saying too much, the test piece will be put at one node, and aerial/reciever at another.

I guess my training is different, and taylored be applied to different problems. Using network analysis, scattering parameters w/ clearly defined ports, that I estimate to what you refer to as aerial. And true, if a calibration proceess is employed, the error network that is between the "idealized" measurement instrument and test ports can be "de-embeded" or removed from the raw data collected.

A few notes on high power stuff.
Dielectric is lossy, if you can build something without it, please do.
If you make a stripline power divider, use either a very thin printed circuit board (0.005", 0.010") with 2 Oz. copper and space it with plastic spacers, or use solid metal parts with foam spacers under and above to avoid dielectric. Hexcel Honeycomb is a good spacer material. Use Hexcels Texas representative if you order any, Hexcel is too expensive with their minimum order requirements.

If you need to fill something with liquid, buy Castor Oil, it's totally lossless at this Microwave Oven frequency range. If you put Castor Oil in a normal microwave oven, you can't even feel any heat generated, super low loss. It's good for cooling too.

If you have a cable with dielectric on it, and transition to something without dielectric, there can be breakdown arcing at the surface of the dielectric at much lower powers than you'd ever imagine, so tapering dielectric to air is a good way to help that situation.

I used a Microwave Oven as a source in a Cancer study to heat tumors through the skin, it was a cheap source of power. We purchased a waveguide to coax transition and spaced it away from the antenna in the microwave oven to vary our CW power. Added a new door with type N transition, and voila, had a nice variable CW power. We of course got the extra warranty on the oven, but didn't use it.
Loss in water is about 10 dB per inch at that frequency (measured).

kch

Some good pointers Higgler. Thanks