In case you're wondering Maxwell is a Finite Element Analysis package from Ansys for low frequency EM fields [1]. It looks quite neat. I never used it, but I worked for a competitor and we had something similar.
I really wish AI could help with learning open source FEM software so far having grok or gemini write elmer/fenics code has been a very bumpy ride my only saving grace is the fact that I can ask my colleagues for guidance and have them provide a working base to expand from. It would do wonders for the hobbyists scientists as well I can't count the times I had a idea for a product but it failed because I couldn't simulate the problem to see if it could work out.
Interestingly, a few weeks ago I fed Claude a curve from a datasheet (actually a screenshot from the web) via its UI with some of its internal tools enabled, and, with minimal prompting, it turned that plot into an interactive tool that analyzed it. (Good results took four attempts, all in the same conversation. Try 1 had JS errors and did not work at all. Tries 2 and 3 had problems with the visualization. Try 4 was usable.) Oddly, Claude corrupted some of the text in the sheet.
I would not trust this for real work — a real tool like SheetScan is likely much more reliable and is unlikely to hallucinate data.
My experience working at large size company that paid those sort of sums for a different FEA software is that we got an "application engineer" assigned to us to answer any questions we had and provide needed documentation if anything was needed. He actually sat and worked from our own offices one day per week.
HN has a 'second chance' mechanism where the mods will essentially bump up an article that didn't get much attention the first time around, but one if the side ffects is that all the timestamps get reset.
In case you're wondering Maxwell is a Finite Element Analysis package from Ansys for low frequency EM fields [1]. It looks quite neat. I never used it, but I worked for a competitor and we had something similar.
[1] https://www.ansys.com/products/electronics/ansys-maxwell
I really wish AI could help with learning open source FEM software so far having grok or gemini write elmer/fenics code has been a very bumpy ride my only saving grace is the fact that I can ask my colleagues for guidance and have them provide a working base to expand from. It would do wonders for the hobbyists scientists as well I can't count the times I had a idea for a product but it failed because I couldn't simulate the problem to see if it could work out.
Interestingly, a few weeks ago I fed Claude a curve from a datasheet (actually a screenshot from the web) via its UI with some of its internal tools enabled, and, with minimal prompting, it turned that plot into an interactive tool that analyzed it. (Good results took four attempts, all in the same conversation. Try 1 had JS errors and did not work at all. Tries 2 and 3 had problems with the visualization. Try 4 was usable.) Oddly, Claude corrupted some of the text in the sheet.
I would not trust this for real work — a real tool like SheetScan is likely much more reliable and is unlikely to hallucinate data.
File under "unexpectedly interesting and cool". I'd never seen such good visualizations of magnetic fields before, or of the magnetization process.
Ansys is such a powerful range of software but the guides on all features are basically non existent. (The tutorials do not count).
Ansys also costs 5 digits per seat per year.
My experience working at large size company that paid those sort of sums for a different FEA software is that we got an "application engineer" assigned to us to answer any questions we had and provide needed documentation if anything was needed. He actually sat and worked from our own offices one day per week.
How do you learn it then?
Was this not posted five days ago with the same comments? Why are the times shifted?
HN has a 'second chance' mechanism where the mods will essentially bump up an article that didn't get much attention the first time around, but one if the side ffects is that all the timestamps get reset.