VituixCAD

^Very good.
I just made small change that driver instance in the crossover is linked automatically to driver data in Drivers tab as long as number of driver instances is smaller or equal to number of drivers in Drivers tab. In other cases linking should be done manually. This could be helpful in practice, but not always. Feature is available in the latest built.

Basically Drivers tab is just storage for measurement data. Actual/physical driver instances are added to crossover schematic, and they could be linked to any driver in Drivers tab. For example, you can have 20 full-range drivers in a project (added in the crossover) and all of those are using the same measurement set (Driver #1) in Drivers tab.
 
Thanks kimmosto,
I appreciate your detailed explanation. It makes perfect sense now. I'm starting to wrap my head around your fabulous software. Still need to read the help manual from start to finish. Your YouTube videos are also a great help in learning more about the program. Could you at some point make a video(s) detailing your steps to designing a successful Xover? It would be very much appreciated from newbies like myself.:D Love your program and the hard work you put into it!

Happy Holidays,
Rich
 
Rev. 2.0.9.0 (2018-12-09)

Enclosure
* Added calculation of T/S parameters with added mass or sealed box method. Feature integrated into same window with extended impedance model solver. Window opens with 'Calculate T/S...' button in 'Edit parameters'/'Add new driver' window. Two impedance responses, Re and Dd/Sd are needed. Z1k and Z10k of basic Z model are detected from free air impedance response.
* Extended impedance model solver uses loaded free air response as a reference within 1-20 kHz with 1/6 oct. resolution. Manual target points (4) are used if free air response is not loaded.

This feature is normally in measurement programs, but might be useful in VCAD because works with txt/zma impedance responses files i.e. does not require original mls/sin/lim/mdat files.
Added mass and sealed box methods are for cone drivers only but that could be quite adequate for a while.
 
^That's so true. I tested few different methods to specify fs (before knowing L). Quite common system is to detect frequencies f1,f2 where Z=sqrt(Zmax*Re), and then fs=sqrt(f1*f2). This could give a bit strange result if symmetry of Z peak is not expected. Also for drivers with very low inductance.
We could also boost this basic formula with frequency at measured peak: fs=(f1*fpeak*f2)^(1/3), or include also frequency where polarity changes: fs=(f1*fpolar*fpeak*f2)^(1/4).
Current method is that f1 and f2 are skipped and possible small offset down is accepted. Looks more stable than with f1 and f2, imo.
 
Extended Z model solver in the same window would be fully capable to fit also any number of T/S-parameters to measured impedance curve within 10...20000 Hz. But excessive focusing to fs is probably overshoot in this case. VCAD is using impedance response as interpolated to 1/48 oct. resolution. Original file exported with measurement program has some other resolution which is also interpolated from constant frequency step FFT data. In addition, expected <0.5 Hz error in VCAD is probably less than tolerance of individual drivers.
So, T/S-calculation with measurement program is recommended to make science with much better accuracy.
 
Rev. 2.0.9.1 (2018-12-14)

Enclosure
* Added Mms and BL checkboxes to T/S Calculation. Parameters can be calculated with free air impedance response if Mms or BL is known and trusted.
* Manual target points of extended Z model removed, assuming that free air impedance response is always available for Z-curve fitting.
* Lossy inductance is subtracted from impedance responses before detection of fs and Q-factors.


This feature should be quite final now. I've also tested curve fitting for T/S parameters, but advantages are not so significant. Iteration would be okay for fine tuning with free air response when Mms or BL is known.
 
Hi Vasslist,
You have to select a driver first. It looks a little grayed out but try it and you will see that you can select the driver and then you can use the upper part. The upper part is really to work on individual drivers first prior to working on the crossover area.
 
Sorry I was not clear. If you look at the Optimizer Box, you will see Axial response of Driver grayed out. There is a box next to it with a downward arrrow that also looks to be grayed out but you can click on the arrow and select your driver. The upper part then becomes active. Hope this helps,
Jay
 
The latest incarnation of TSP window (rev 2.0.9.3) shows measured and simulated impedance responses. Simulated curve is either free air/IB or closed box (Q=500) depending on selected method.

An externally hosted image should be here but it was not working when we last tested it.


Some empirical weighting factors are added to both methods to get smaller difference between available measurement data and simplified simulation, and between different methods and programs with T/S calculation.

Manual adjustment of Mms or Bl enables playing with Mms/Cms relation. Graph shows changes in simulated Z.
 
The latest incarnation of TSP window (rev 2.0.9.3) shows measured and simulated impedance responses. Simulated curve is either free air/IB or closed box (Q=500) depending on selected method.

An externally hosted image should be here but it was not working when we last tested it.


Some empirical weighting factors are added to both methods to get smaller difference between available measurement data and simplified simulation, and between different methods and programs with T/S calculation.

Manual adjustment of Mms or Bl enables playing with Mms/Cms relation. Graph shows changes in simulated Z.
Kimmo, thank you very much for this feature. Finally, you also implement measurement to VituixCad. :)
 
Finally, you also implement measurement to VituixCad. :)

I'm closing measurement features step by step. Two steps are still missing: generating output signal and actual measurement of input signal :)

About TSP calculation. It would be nice if all programs give equal result with the same measurement data (impedance responses), and with both methods. Reality is more or less disappointment, and differences/tolerances could be 10...20% in Rms, Mms, Cms and Bl.
Free air part (Re, fs, Qms, Qes, Qts) is not the problem, though some programs could have small systematic bias also in fs.

My decision to use some weighting factors to reduce differences between methods is just a compromise which may help in average but doesn't guarantee that (Rms...Bl) result with individual driver is accurate. Error is expected also with added mass which is preferred due to better accuracy.

Fortunately extreme accuracy is not important. Simulated SPL and impedance curves could be totally okay with errors >15% (Rms...Bl). Few % shift in system resonance (fb) or Zmax does not kill.
 
Rev 2.0.9.4

Mic-in-Box -function added to Calculator tool is kinda experimental. It's just 2nd order high-pass which compensates "room gain" if sub/woofer is measured so that mic is located inside the box. Generally, this method should be slightly better for closed boxes than traditional near field (outside the box), but frequency range could be more limited.
This feature could also be in Merger tool, but I suppose it won't be so popular as NF measurement due to more difficult mic arrangement.
 
I'm closing measurement features step by step. Two steps are still missing: generating output signal and actual measurement of input signal :)

About TSP calculation. It would be nice if all programs give equal result with the same measurement data (impedance responses), and with both methods. Reality is more or less disappointment, and differences/tolerances could be 10...20% in Rms, Mms, Cms and Bl.
Free air part (Re, fs, Qms, Qes, Qts) is not the problem, though some programs could have small systematic bias also in fs.

My decision to use some weighting factors to reduce differences between methods is just a compromise which may help in average but doesn't guarantee that (Rms...Bl) result with individual driver is accurate. Error is expected also with added mass which is preferred due to better accuracy.


Fortunately extreme accuracy is not important. Simulated SPL and impedance curves could be totally okay with errors >15% (Rms...Bl). Few % shift in system resonance (fb) or Zmax does not kill.
In my opinion, the fundamental problem is the value of the measurement signal. The very pleasant procedure is used by SB accoustic. The datasheet is on the page.