VituixCAD

Looks like auto scale setting of .NET form does not freeze font size, and that could lead to some other problems. Thanks for info. Could be worth to restore official release and scaling to 100%. I'll try to avoid bad behavior in the future, though app. scaling works quite okay on win 10. Probably without problems if application windows fit on screen.
 
I'll try to avoid bad behavior in the future...

Looks that custom code for Windows 7 only is too complex and not productive enough because 7 is obsolete anyway. SPL Trace alone would be piece of cake compared to custom graphics such as crossover schema.
Scaling of 100% is recommended until operating system is upgraded to Windows 10 Creators Update (1703) or later.

More information: High DPI Desktop Application Development on Windows
 
Revision 1.1.25.0 (2017-12-27)
Diffraction tool: maximum amount of drivers increased from 12 to 50, and added distance between drivers (Step mm). This helps creating of long line sources with multiple domes or rectangular planars. Long planar line should be simulated as multiple small planars instead of single long.

An externally hosted image should be here but it was not working when we last tested it.
 
Length of wish list is quite short at the moment, maybe 4 features. I've not promised anything so far. I had to take a break because this has been continuous work for last two months.

Hello!

Just discovered this awesome software a few days ago. Really great work!

You mentioned new features. What about adding autoformers as passive cross-over element? For compression drivers these are a common method to use in cross-overs and there are some really good devices available again. I am not aware how complex this would be incorporate.
 
What about adding autoformers as passive cross-over element?

It's quite probable that I won't add autoformers to existing circuit simulator (which plays with impedance blocks in ladder circuit), though it might be technically possible. Bigger plan is to replace whole circuit simulator with some spice library. That would change the game also regarding new components and custom blocks. But, this plan has few major obstacles such as time and experience. Personally I almost hate free form circuit CAD, but it could be acceptable with a template library.
 
Rev 1.1.25.2 (2018-01-09)

Diffraction tool
Authors of DED model used in VituixCAD didn't provide separate model for thin baffles and boxes due to good results with common model. I've not been entirely happy with directivity index of closed box simulation. DI has been about 1...2 dB too much at midrange and off-axis too attenuated. This is now tweaked with multiplier (from my sleeve) which turns observation angle of dipole edge sources smaller in the simulation.

Some tests:

Measured Kef Q100 to hor 0-90 deg off-axis, time window 5.1 ms rectangular:
An externally hosted image should be here but it was not working when we last tested it.


Diffraction simulation to hor 0-90 deg, Dd=100mm.
Measured DI = orange.
An externally hosted image should be here but it was not working when we last tested it.



Measured Kef Q100 to hor 0-180 deg off-axis, time window 5.1 ms rectangular. Do not mind response curves to shadow region. Q100 is cheap crap with lightweight and resonating thin box.
An externally hosted image should be here but it was not working when we last tested it.


Diffraction simulation to hor 0-180 deg, Dd=100mm.
Measured DI = orange.
An externally hosted image should be here but it was not working when we last tested it.



OT: I've tested 0-90 deg simulation also with BDS and Edge (click links). They are not subject in this thread and Dd was slightly smaller. Shortly put, both have high error in off-axis diffraction. I suppose due to diffraction model of Bews and Hawksford which is designed for axial and nearby. In addition, Edge is really horrible, difficult and slow for this kind of simulation because angle setting does not exist and level is normalized by mic distance. Not designed for off-axis simulation.
 
Thanks! A bit lower angle correction factor e.g. 0.75 might be closer to truth, but 0.8 looks adequate as initial value. More prototypes and measurements are needed to define better value and method, but that won't fix all approximations in simulation such as ray model without higher order diffractions, missing box depth setting etc. Not much more needed for the purpose; converting half space to full space at low frequencies and tuning mid/tweeter location. Also ideal piston radiator is adequate for both, though it's not equal to reality.
 
Thanks! A bit lower angle correction factor e.g. 0.75 might be closer to truth, but 0.8 looks adequate as initial value. More prototypes and measurements are needed to define better value and method, but that won't fix all approximations in simulation such as ray model without higher order diffractions, missing box depth setting etc. Not much more needed for the purpose; converting half space to full space at low frequencies and tuning mid/tweeter location. Also ideal piston radiator is adequate for both, though it's not equal to reality.

Very good work Kimmo. Little question. Resistance can be added as the next coil value?
In serial connection I will do it with other resistors, but in parallel it seems to be a problem.
Thank you