What high end RF/MW/MMW software is the nicest?

[coppercone2]

What EM software do you think is the easiest to use and connects well to solidworks?

I am scared that HFSS is the 'autocad' RF equivalent. I don't want to learn how to use a program that is requires complete psychosis to use (like eaglecad is completely maddening in comparison to Altium).

Which one is the solidworks equivalent, meaning easiest to use, most intuitive menu options, nicest interface etc. Even if its less accurate I would trade that for ease of use and having actual ergonomics/user interface considered in its design.

I have lost a great deal of time struggling with eaglecad and autocad (in the way past). When I got altium and solidworks it was like getting a blow job. I don't want to repeat the eaglecad experience. If you compare eaglecad and altium, eaglecad just feels like its broken/retarded. Another analogy to those not familiar might be using the very nice PicC compiler instead of Microchips compiler, which is pure venom. It made me feel like I was trying to fuck a torn foam couch cushion that a bobcat pissed all over and considered his property.

So, which one will save my monitor from physical battery and make me not wanna hang my computer up like a punching bag in the basement? The linux mint instead of OpenBSD.

I don't think I want to consider free software, given my experience in Photoshop vs GIMP (omfg I would rather learn how to read Egyptian hieroglyphs then decipher the user interface for that, I am convinced it was designed by people completely fried from LSD, the menu options and button pictures and everything just felt completely disconnected from eachother, it was like sorting through random drawers in a messy filing cabinet) , or Microsoft Word Excel vs OpenOffice excel (which is the least offensive open source Excel program I used but the features are severely limited, the user interface feels kinda OK though but it still sucks compared to MS interface in terms of being intuitive, I seriously feel like microsoft can read your mind when it comes to menu options and stuff, it just flows and its a pleasure to use). Solidworks kind of feels like a microsoft product in terms of how pretty much you can randomly click stuff and you get to where you need to go, even if its somewhat less efficient then knowing the hardcore production way of doing things like duplicate features etc. That's what I want. Photoshop is kind of like that too, I was able to get stuff 85% the way I wanted just quickly browsing around the well organized menues and stuff, whereas gimp you need to go on google to draw a freaking line.

I know if you train on those crappy programs for a long time you might find things you like and can justify the cost but as a weekand warrior and person with low stress tolerance I can't deal with that kind of crap. I don't want to use them unless there is no alternative or I am being paid well to deal with the deficiencies.

And I am not railing on the open source community, because Autocad is super expensive but it still has the side effect of vaporizing your blood.

[D3f1ant]

I have no suggestions but your post is was funny

[TheUnnamedNewbie]

To be honest, AutoCAD and Solidworks are just quite different tools, and AutoCAD has some history which is why it is a bit different to use. I've used a number of mech cad tools (from free to sell-a-kidney-for-a-years-license), and SolidWorks was my least favorite by a long shot.

If you are a weekend warrior person, none of the tools used in industry is going to be for you, because all of them cost more than a car. You say you consider AutoCAD to be super expensive. I suggest you sit down for this, because HFSS costs more than 100x that (assuming AutoCAD costs about 2.2k)

That aside: What do you want out of your tool? Not all tools are equal, and not all tools can do the same things. Different types of solvers can also give different accuracies depending on what you are doing, and different speed versus what you are sweeping (EG, time-domain analysis can speed up wide-band sweeps, but not all tools can do it, and it can give you less precise results at any given point vs a single frequency f-dom analysis).

HFSS is my go-to tool for 3D simulation. I tend to import STEP files which you can export in any self-respecting bit of mech CAD. The Ansys electronics desktop suite is quite nice, though I don't use it that often. I've heard good things about HFSS 3D (more for PCB layout (verification) and analysis, not hard microwave simulation).

CST is another big one, and our antenna designers seem to prefer it over HFSS. Significantly cheaper than HFSS (still 80k), but comes with less bells and whistles.

Another tool I use quite a bit is Keysight ADS. The momentum solver is a fast 2.5D solver we use a lot for simulating on-chip inductors and transformers. Rumor has it they are looking at adding full Analog IC CAD tools in there to allow the entire chip design to be done in ADS (and get rid of Cadence's monopoly) but I haven't heard any updates about this. Also don't know what it costs.

And then there is Sonnet which I haven't used. Another tool I have heard of is COMSOL Multiphysics, a few colleagues of mine use this to simulate microwave heating. Again, can't say much more about it.


I do have to add that there were a lot of tiny, smaller companies providing 3D EM simulation at IMS. I don't know which ones, but those might be worth looking into.



Oh, and regarding the software being intuitive - I have not really had to much problems with this, but at the same time, there is not a tool like this that I do not have the manual open for when using. These are so complex and specialized bits of software - not just design tools but simulation tools. Your choices will impact accuracy and poor simulation setup can result in poor performance at best, just completely wrong results at worst.. And I know this from experience, since we had to redact a paper at the last moment because it turns out we made a mistake in the setup and got the losses of a simulation just completely wrong (100 dB wrong).

[coppercone2]

Just 'theoretically' speaking with price not being an object.

I am interested in:

-Simulation of electromagnetic behavior of box with circuit board in it. I doubt I would be able to get any kind of RF models for any of the parts I use (say DDS chip from AD, high speed op-amps like 1-4GHz gain band width (the serious opamps) for things like active circulator design). I suspect that the optimal PCB layouts might result in boxes that look weird like having triangle shapes and stuff. I suspect that setting up a usefully accurate electromagnetic simulation of a PCB is outside of my interest because I think you would either need some kind of automated characterization system or do alot of cold calling and begging to get useful models, nor do I find it that interesting on a PCB level to justify simulation.

-Simulation of waveguide things, like those filters that have the little disks that hang off of little rods that you adjust (BPF), I can see myself making something like that if I ever get a lathe and mill, end launchers, different forms of waveguide like circular, flexible, so I can learn to see what happens electromagnetically due to impedance mismatches and to analyze the effect of manufacturing tolerance on waveguide properties.

-Simulation of horn antennas. I see that with tools like the $6000 dollar 5(or was it 6) axis mill that is capable of making cuts in materials like aluminum and even stainless steel or titanium (at a slow rate but how many antennas will I be making), I am interested in those horns like the ones that have dual polarity, bastardized conical shapes, etc. I am not really that interested in other antenna types because the horn seems to be kind of like the best antenna for what I am interested in, I was looking at eventually being able to make my own series of standard gain horn antennas for a wide microwave frequency range  when my fabrication skills reach maturity.

-Simulation of dielectric materials to try to optimize a home made EMI chamber with home made foam that I can categorize with the dielectric measurement methods as suggested by someone in another thread, to see how things like layout of cones, geometry and materials composition all come together in anechoic chamber dynamics.

-simulation of microwave reflectors for making things like dishes that are used to communicate with satellites

-simulation of polarization filters (I guess its just a special type of dish)

-simulation of microwave heating apparatus for small samples (like making your own microwave oven with a waveguide feed for heating small things like 'shelenk ware', which would fall under the subset of high pressure medium temperature chemical reactions that have unique reaction dynamics due to the nature of the microwave heating the sample

All these things are not really super dB sensitive and I am suspecting that my manufacturing methods will result in big errors.


I am not that interested in simulating something that will be 0.005 dB accurate, I noticed that most electrical work I do ends up being kind of ratiometric in nature anyway. I thought that certain programs might be catered to designing things with really narrow pass bands, phase coherence, stability

I am more interested in kind of cool physics apparatus then very tight specifications on communications systems, super high repeatability, metrology grade PCB's for use in microwave equipment etc. I don't really see myself doing board level microwave design, it is the least interesting to me because you need to get the PCB made in a factory anyway, it seems the other stuff you can maybe do in a machine shop with enough iterations. I don't think having SMD parts models would be useful for me at all.

Visual representation of fields is very important to me so I learn to imagine what is going on inside of microwave cavities, resonators, antennas, and so on. More so then bode plots and measurement graphs of insertion loss. I also am interested in seeing how microwaves react with bulk dielectrics for things like paraffin lens in experiments


Also stuff like trying to model a structure (say a shed, small building, signs, fences, other common obstructions) in solidworks with realistic materials to see how radio/microwaves are effected by it so I can better understand what is happening with outdoor antenna uses and behavior of radar systems to better understand things like radar cross section of airplanes, stealth aircraft features, etc. Maybe even trying to make a model of my lab and seeing how it compares to real life when lightly irradiated.

Alot of it is that I want something like a 'gmod' (if you played the half life 2 mod) for messing around with RF stuff. I don't expect to make any money learning any of this for any kind of manufacturing purpose. I just really can't imagine half the stuff people talk about and it would help me so much to get cool RF predator vision.

Like with solidworks, I am modeling heat-sinks and cooling systems, but I am never gonna build it like the model, I just want to learn the relationships between different elements to get a good idea of how I can over engineer something properly. Exact simulations I think are typically useful if you want to save money in large scale production, but it would be too much of a pain in the ass for me to build 5 heatsinks with different fin spacings to get an idea of how the spectrum of design choices effects thermal transfer. I am trying to put boundaries on absurd design choices, not hone anything. I know microwaves are alot more sensitive then thermodynamics however, but I don't really see myself sitting there for days on end trying to shave off a fraction of a dB or something, its just really easy to do something completely ridiculous with electronics.

 I know its possible to kinda do things by learned intuition and feel, especially after reading documents from the 1970's where they tend to use peacewise approximations and stuff for research, rather then simulating some super refined model... I wanna be able to get that without all the aids a well funded researcher back then would have. (like the work of John Kerr with his dual ridged horn designs)

A useful simulator would probably save me thousands of forum posts.