Sonotrode design software

Hi Ronald,

Welcome to the forum, and thanks for putting up the question!

I've not used Atila but judging by their web site docs
[url]http://www.cedrat.com/software/atila/pdf/atila.pdf[/url]
what they're offering is coupled analysis - electric / mechanical dynamic and electromagnetic / mechanical dynamic. This means that you can analyse piezoelectric ultrasonic transducers whose dynamic behaviour is determined by both the piezoelectric and mechanical properties. (Broadly these would be shape, density and modulus of the front and rear blocks, shape and piezoelectric properties of the ceramic disks).

Other producers of finite element analysis (FEA) programs offer similar coupled analysis capabilities, eg. Ansys in their Multiphysics package have a specific piezoelectric analysis option - more info here:
[url]http://www.ansys.com/industries/mems-feature-pzt.asp[/url]
You can probably find similar options in most of the major FEA programs - Abaqus, Marc etc. (Apologies to those I've left out - if readers know of others that include this option please post the information.)

In terms of proven reliability the FE code from reputable suppliers is rigorously checked, but of course in a specific application the quality of results depends on the accuracy of the model - not always easy to assess. And unfortunately as you've discovered these high-powered FEA packages tend to come with a price tag to match.

The good news (unless you're a transducer manufacturer) is that sonotrode analysis is a much simpler operation involving only dynamic (mechanical) analysis. In fact for many purposes you can get away with the even simpler Eigenvalue (mode-frequency) analysis, which gives good accuracy on most sonotrodes since they're generally very high-Q systems. Using a simpler analysis type makes it easier to get good results (there's less to go wrong!) and permits the use of cheaper, more limited FEA software.

I've been looking a Lisa
[url]http://www.lisa-fet.com[/url],
which appears to offer this kind of basic functionality at a much more reasonable price - ranging from 800 euro for eigenvalue and dynamic analysis to 2000 euro for all their analysis options. Of course even the full version would be much more limited in application than any of the major packages but if all you need is basic static / dynamic / thermal analysis then it may be enough. They even have a free trial version, tho with a 1300 node limit it's not much use for real-world problems (but I guess that's the point!).

Another one to watch is CalculiX
[url]http://www.calculix.de[/url] - which is free open-source software released under the GNU GPL. Sadly it's not yet in a state to be easily used - there's a Windows executable for the solver only (without the pre- and post-processor this has limited usefulness!), the Linux binaries don't run on my (recent) Linux system and the source doesn't seem to want to compile! But the prospect of open-source FEA software is great, and many thanks to the people at MTU Aero Engines and Ing. Buero Schreck for making it happen. If I get the time I'll try to get it compiled under Cygwin (windows) or Linux and put up a new binary for download...

Finally you might look at Don Culp's Computer Aided Resonator Design (CARD)
[url]http://www.krell-engineering.com/software/card/card.htm[/url]
which is a simple finite-difference program specifically to analyse (axisymmetric) sonotrodes. At US$950 the price seems pretty steep for a rather limited program and the installation is dire (Dos-based!) but at least there's a free trial!

Anyone have any other suggestions for reasonably priced sonotrode analysis software? Please feel free to post them here on the board. At some stage I'll set up a new section of the site to cover ultrasonics analysis so all additional information is welcome.

Ronald - hope that helps you. Any other questions pls don't hesitate to ask.

Regards
Chris

Hi Roland,

[color=Blue]you can check online the stresses inside the material, activating a demo frequency[/color]
Not sure what you mean by that - once you've run an analysis the results should be on file until you delete them. In theory only dynamic analysis can give you stresses (since there's no amplitude information in an eigenvalue analysis) but sometimes you can trick the programs anyway - I used to use seismic excitation with Ansys to get a nominal amplitude into the eigenvectors...

[color=Blue]some time a little mistake can cause to start everything from the beginning[/color]
Now that I do understand!

Re. Solid Works, I think a few years back they swallowed up Cosmos (one of the smaller FE packages at the time). The current(?) offering looks good especially for multi-component vibration analysis - drive trains etc.
[url]http://www.wildefea.co.uk/download/software/131/SD-MI_2003.pdf[/url]

Thanks for the feedback.

Regards
Chris