While we did not have a major release since then, this does not mean that our development team has been idle. Robin Hood Solver has received ten smaller updates since the release of v2.0, released every two to four weeks. As of today, the current version of Robin Hood Solver is v2.0.11 – you can see that easily for yourself if you download a demo version.

What have we changed? A lot of small things (and several large ones). Most changes were prompted by the comments and requests from our users – so, if you test drive Robin Hood Solver and have a suggestion to make, don’t hesitate to send us a message or post a request to our support site.

Here is the list of most important changes in releases up to and including v2.0.11:

• Welcome screen enables you to quickly return to existing projects and to access tutorial and manual
• Robin Hood Solver now uses SI units instead of proprietary units
• all new Functions dock makes scripting with Lua easier
• color palettes are now saved with the project (in initial release, this feature was not implemented – users needed to manually reapply palettes after reopening the project)
• color contours differentiate color areas on objects and make potential/field visualizations even more impressive and even more easier to interpret at a single glance
• improvements to palette handling: users now control the palette using single property (‘color correction’)
• many improvements to dataset handling: better export options, possibility to hide visual datasets from the scene etc.
• it is now possible to automatically create objects out of imported rough (corrugated) surfaces
• all new text-box for editing properties in scientific notation now accepts any kind of input and interprets it correctly (enter 4,445776E-02 or 0,4445776; Robin Hood Solver will get it)
• snapping of objects is now turned off by default: to turn it on, hold the Ctrl key while dragging the object

In addition to that, our developers have spent a lot of effort in finding and correcting  small bugs and implementing subtle appearance changes which greatly improve experience of using Robin Hood Solver. Check it out for yourself – get a demo now!

The company will showcase Robin Hood Solver 2.0 – a top of the line CAE solution for electrostatic analysis, modeling and visualization.  Users can get a free demo version at www.artcalc.com/demo. Company representatives on Hannover Messe will give live demos of the software and help interested users to learn how to use the software to solve their research challenges more efficiently.

Look for us in Hall 12, booth A31. We’ll be a part of joint Croatian booth organized by Croatian Chamber of Economy.

We are looking forward to your visit!

Our new address is Derenčinova 1, HR-10000 Zagreb, Croatia. The new office is just several minutes away by foot from the old one and close to the city center.

See the map:

View Larger Map

Please note that we have recently also changed the phone number. You can reach our switchboard at +385 1 7980-236 and you can fax us at +385 1 7980-566.

Robin Hood Solver v2.0 builds upon praised and well-known Robin Hood method and offers many new features, including:

• new, advanced GUI:
  • create models quickly, with just a few clicks of a mouse
  • choose objects from a predefined list, then change their properties to fit the desired model
  • move and resize objects using your mouse, or change their properties using the Properties dock
  • perform Boolean operations (intersections, unions or subtractions) by clicking the icons on the toolbar
  • display objects as solid, transparent or in wireframe rendering
  • add palettes to better visualize calculation results
  • use clipping planes to see inside your objects
  • view your model from four different perspectives (top view, front view, side view or full 3D view) – even simultaneously
  • use graphs and tables to better understand results and inspect the objects in the model
  • observe the solver while it works: as RHS is solving your model, it graphs the accuracy of the calculation and colors the objects in real time
  • examine the results of calculations using Output dock
• extensive project management capabilities, which make it easy to store multiple models and calculations
• every project is saved as a list of operations – this means that the project file itself takes up a minimum amount of storage space and can be easily copied or sent via e-mail; also – the operations executed in the project can be modified later
• auto-save: all operations are automatically saved to the project file as they are performed; there is no need to use Save command
• it is easy to import objects from DXF files (and other file types) created in other applications
• integrated Lua scripting helps you automate your calculations and exchange of data and results with other tools
• Robin Hood Solver is now true multi-platform solution, available for Windows, OS X and Linux (all features of Robin Hood Solver work on all three platforms

Robin Hood Solver v2.0 is available for testing – fill-out a short form to request a demo version now.

Demo version is free of charge and licensed for evaluation purposes only. There are some limits to functionality. See the version chart for more details.

Full version is also available for our academic and commercial customers.

 The very functioning of the MEMS devices rests upon movable or deformable parts which are manipulated via electric (essentially electrostatic) forces between them. To obtain a faithful and precise simulation of the MEMS dynamics, it is necessary to solve the electrostatic problem as soon as some of the parts moves or deforms a bit. This translates to the requirement of solving many electrostatic problems with very close or similar geometries consecutively. In the optimization of the configuration and geometry of some electrostatic system to achieve a predefined requirement (e.g on the electric field in some region), many geometries and configurations need to be tried out. Irrespectively if the geometries are tested sequentially, or more advanced adaptive algorithms that select geometries are applied, calculations for close geometries are performed consecutively.

In both of these cases, Lua scripting capability of Robin Hood Solver 2.0 makes possible a systematic automated study of a large number of geometries and boundary conditions on conductors. Yet, the Robin Hood method, being iterative in nature, has room for further improvements at the algorithmic level for problems which require solving electrostatic problems on a large number of (similar) geometries.

For a given geometry the solution of the electrostatic problem is found in a Robin Hood iterative procedure. It is expected that for similar geometries the solutions (i.e. surface charge distributions) will also be similar. An example of such geometries obtained by increasingly large deformations of a sphere into ellipsoids is presented in Fig. 1.

Figure 1: Ellipsoids obtained by consecutive deformations starting from a sphere (image generated by Robin Hood Solver v2.0)

Another important question is that if for an initial charge distribution in the Robin Hood method we choose a charge distribution that is close to the actual numerical solution, will the duration of the iterative solving procedure be significantly shortened? In the case of affirmative answer, it would be reasonable to build a procedure of finding solutions for a series of close geometries so that the solution for one geometry is used to produce the initial distribution for the next close geometry. Such a procedure might then reduce the time of simulation of MEMS dynamics or time of finding optimal geometries that produce the required properties of the static electric field. We refer to this procedure as the recycling of the solution. Nontrivial elements of the viability of the procedure are mappings of triangulations between the close geometries and the balance in time of execution between the calculation with the recycled solution and the calculation from scratch for each of geometries. Future R&D will have to decide how big advantage the recycling of solutions will bring to high-throughput calculations in electrostatics.

For a general spatial dependence of the dielectric constant, the space filled with the inhomogeneous medium can be divided into volumes with constant permittivity. Then take these volumes as pieces of homogeneous dielectric for which the description in terms of induced surface charges is suitable. Finally, just run the Robin Hood Solver with this configuration of numerous homogeneous dielectric bodies. Other BEMs would quickly run into trouble following this approach. As the number of volume segments into which the medium is divided grows, so does the number of triangles (surface elements) needed to describe their surfaces. For standard BEMs this number very quickly becomes prohibitively large. The Robin Hood method, on the other hand, can handle very large numbers of triangles and, consequently, large number of volume segments with homogeneous dielectrics. That should allow a sufficiently detailed description of the medium with the spatially dependent dielectric constant.

In practice, the dielectric constant frequently varies just in one dimension. An example of the division of the medium into slices of material with constant permittivity is presented in Figure 1 for the case of variation in z direction.

Figure 1: The division of the inhomogeneous medium into thin slabs of material with constant permittivity (image generated by the development version of Robin Hood Solver v2.0)

For a situation with spherical symmetry, depicted in Figure 2, in which the dielectric constant has a radial dependence, the medium needs to be divided into thin spherical shells of material which are treated as homogeneous dielectrics.

Figure 2: The division of the inhomogeneous medium with spherical symmetry into spherical shells with the same dielectric constant throughout the shell (image generated by the development version of Robin Hood Solver v2.0)

In the systems where the dielectric constant has very large spatial gradients and large variability, the approach presented above might become unpractical, or, in extreme cases, unfeasible. However, useful approximations can still be achieved by the division into larger (compared to the typical scale of variation of the dielectric constant) volume segments which are considered homogeneous with the dielectric constant corresponding to the average dielectric constant in that volume segment.

The broadly based philosophy behind the Robin Hood method in principle also allows its generalization to a method for inhomogeneous media. This important topic is a challenge that will be addressed by the Artes Calculi R&D team in the future. In the meantime, the Robin Hood Solver, by the method of slicing the medium into segments of approximately the same dielectric constant, has a way of dealing with inhomogeneous media as well.