VituixCAD Not Summing Same as XSim or PCD7

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello all, I recently installed VituixCAD and it seems really great! Powerful, fast, etc. Except I started by copying a design I was working on in XSim and I noticed the SPL/filter transfer function was very different between the two programs and I can't seem to figure out why. So I typed it into PCD7 and it agrees very closely with XSim. Anyone else experience this? I would love to make Vituix my go-to program, but I gotta figure this out first. See files and screenshots below.

Before you ask, yes, I messed with the offsets, minimum phase, generator settings, none of it comes close to making it match. I've got to be missing a setting in Vituix somewhere. THANKS!
 

Attachments

  • TC1 docs.zip
    354.2 KB · Views: 41
Before you ask, yes, I messed with the offsets, minimum phase, generator settings, none of it comes close to making it match. I've got to be missing a setting in Vituix somewhere. THANKS!
:rolleyes:

The virtual mic is at "0" and I don't have any other drivers in there that they would interact with. Also - it's in the filter transfer function which is not effected by offsets.
 
diyAudio Moderator
Joined 2008
Paid Member
You can make it look like this.
 

Attachments

  • ph2.png
    ph2.png
    125.5 KB · Views: 237
Just another Moderator
Joined 2003
Paid Member
I was seeing this recently, I was tearing my hair out and couldn't work out why, until it was pointed out to me that the driver was inverted in vituix cad, compared to xsim and speakerworkshop.

It was a third order butterworth so the difference between inverted and non-inverted was not pronounced, and I thought it was the software. :rolleyes:

I never got exact matches between the three packages, but once that was sorted they were much closer.

Tony.
 
Generator's output impedance should be 0.01 ohm (in all programs) and drivers' X,Y,Z=0,0,0 mm and R,T=0,0 deg to be compatible with XSim and PCD. Drivers' response data is located at the origin (0,0,0 mm) of the speaker with programs without mechanical geometry calculation (such as XSim and PCD).
 
Last edited:
Typical error is 0.001 dB with matrix solver so typical error source is user.

this is my experience. I was able to get a close enough match between XSim and VituixCad2 that the two final FRD curves laid on top of each other. However, to get that close, I had to make certain that both programs were using the same insertion loss values for all capacitors and inductors. In other words, a 2 mH inductor in VituixCad might have a DC resistance of 0.28 Ohm, while in XSim, it might be 0.31 Ohm.

There are a lot of little details that all make a small difference. If you want an exact match, all those little details need to be worked out.
 
^Correct. Also MP extraction is one source for differences because it's just subjective lottery. Reliable comparison requires that both programs use exactly the same phase information i.e. possible MP extraction must be done externally before responses are loaded to programs.
Anyway, comparing different programs is not proper method to verify is calculation correct or not. This is just commented in VCAD thread.
 
Generator's output impedance should be 0.01 ohm (in all programs) and drivers' X,Y,Z=0,0,0 mm and R,T=0,0 deg to be compatible with XSim and PCD. Drivers' response data is located at the origin (0,0,0 mm) of the speaker with programs without mechanical geometry calculation (such as XSim and PCD).

Thanks Kimmo, I know to trust you with Vituix, unfortunately like I said in the original post, that stuff doesn't matter here. You can test it yourself. The filter transfer function is 7dB different (inconsistently), not just the driver summation or acoustic calculations.

EDIT - I had tried a wide range of generator impedance values, but not .01Ohm, which seems to fix the difference! Kimmo strikes again.

Typical error is 0.001 dB with matrix solver so typical error source is user.

Ya, that's why I'm here asking for help!

^Correct. Also MP extraction is one source for differences because it's just subjective lottery. Reliable comparison requires that both programs use exactly the same phase information i.e. possible MP extraction must be done externally before responses are loaded to programs.
Anyway, comparing different programs is not proper method to verify is calculation correct or not. This is just commented in VCAD thread.

I have been very methodical about this. I tried several runs of MP being extracted externally and internally or even a combination. There is very little difference as the issue here is the transfer function being drastically different, not phase interaction.
 
Last edited:
Generator's output impedance should be 0.01 ohm (in all programs) and drivers' X,Y,Z=0,0,0 mm and R,T=0,0 deg to be compatible with XSim and PCD. Drivers' response data is located at the origin (0,0,0 mm) of the speaker with programs without mechanical geometry calculation (such as XSim and PCD).

OK Kimmo, you have some explaining to do! I think the issue is the generator impedance. I tried every value from 1-20, but putting in .01 gets me 99% to match the others! I'm not an electronics guy so that doesn't make any sense to me. Got any ideas? Sorry for writing it off at first!
 
^^Rg seem to work okay in VituixCAD. Problem is that you have entered 9 Ohms to Rg which is output resistance of the generator.
Rg in XSim is default 0.001 Ohms as far as I know so you need to enter 0.001 Ohms also to VituixCAD to get equal result. 6 W and 9 Ohms in XSim are just auxiliary/helping parameters for voltage calculation.
 
^^Rg seem to work okay in VituixCAD. Problem is that you have entered 9 Ohms to Rg which is output resistance of the generator.
Rg in XSim is default 0.001 Ohms as far as I know so you need to enter 0.001 Ohms also to VituixCAD to get equal result. 6 W and 9 Ohms in XSim are just auxiliary/helping parameters for voltage calculation.


That's very helpful. My version of XSim only lets me go down to 1 ohm min. however, it just changes the overall level and nothing else. Whereas changing it in Vituix actually changes the transfer function, which makes sense. With Vituix at .01 Ohm it matches the others consistently. I think this is likely an advantage to Vituix, now that I know. As long as I have all 3 programs agreeing, I feel confident to continue. Though I will continue to pursue understanding electronics better so I can wrap my mind around it.
 
Default voltage is 2.82843 Vrms so you should have 1 W to 8 Ohms in helping parameters of XSim.
It's compatible with Eg=2.83 V and Rg=0.001 Ohms in VituixCAD.

Ok, that checks out, thank you. Follow up question - I am of the belief that MP should be extracted and offsets always typed in before starting, but I see SO many people just leave offsets at 0 and be happy with results. This makes for enormous differences in my simulations and I can't understand how people, even veterans, are getting real-world results anything like their simulation, in ANY of these programs. Even IF I only want the sim from a fixed single position you'd still have to go through the same process, right? Thoughts on this?

I have been following the process Bagby layed out to find Z-offset, but I get some weird results with that method on this project.

P.S. I found differences between having MP extracted through Bagby spreadsheets and Vituix, but found that Vituix was always consistent regardless if MP was extracted previously or not, leading me to trust Vituix more and more.
 
Last edited:
kimmosto is right, the ? watts at ? ohms is only a way for you to specify the voltage output of the power amp by stating it in watts, since some speaker manufacturers would specify sensitivity db when driven in effective watts. Kind of confusing, I know.

To change the amp's output resistance or impedance in XSim you have to add a resistor or complex network inline with its output. The amp's source resistance is at 1mOhm or something like that to keep the math from blowing up. To watch the effective source impedance of the amp+network, put a single dummy driver at its output to ground and have Xsim graph the "lookback" impedance of the driver (see the curves menu of an impedance graph when there are drivers in the design). Lookback is the impedance that a driver sees driving it, which can be an indication of how it might be shoved around from external pressures (like from other nearby drivers).
 
Last edited:
After getting the generator in line as per Kimmo, here is the crossover (starting point) in Vituix on the left, and XSim on the right in the first picture. The curves match, but every feature seems to be more exaggerated in Vituix. Maybe some smoothing or something, I haven't really looked into. Regardless, the phase matches and the summation is working in a similar manner.

Second Pic
When I have Vituix extract MP, and I type in XY offsets only (haven't figured out the Z-offset reliably yet) I get this. I left the XSim graph on the right the same.
 

Attachments

  • Screen Shot 2021-03-05 at 1.04.57 PM.png
    Screen Shot 2021-03-05 at 1.04.57 PM.png
    91.9 KB · Views: 107
  • Screen Shot 2021-03-05 at 1.10.52 PM.png
    Screen Shot 2021-03-05 at 1.10.52 PM.png
    93.1 KB · Views: 95
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.