Vituixcad2

[Doufuss]

Calling @Foz @Passinwind

I used Vituixcad2's enclosure, diffraction, and calculator apps to simulate the box impedance and general performance of the network. I decided to go with a 1st order filter on the 3015 and 2nd order filter on the mid. The 1st order filter also has a zobel network. In order are the spl chart, directivity, impedance, power dissipation, and crossover schematic.


Using the simulator to apply 45 volts, power dissipation for the alpha is 105 watts with a crest factor of 3db.

5.4 ohms minimum impedance.

Fellas do you see any areas of concern?

 

[agedhorse]

Either there's an inaccuracy in your model or the Zobel network is incorrectly modeled and/or way too aggressive for a nominal 8 ohm cabinet.

Look at how much region there is of sub-nominal impedance. That's way too much and can damage some amps unless you call it s 4 ohm nominal box.

I would eliminate the Zobel, or scale it WAY back. That's more of a hi-fi technique that IME often results in worse real world performance.

 

[Doufuss]

Either there's an inaccuracy in your model or the Zobel network is incorrectly modeled and/or way too aggressive for a nominal 8 ohm cabinet.

Look at how much region there is of sub-nominal impedance. That's way too much and can damage some amps unless you call it s 4 ohm nominal box.

I would eliminate the Zobel, or scale it WAY back. That's more of a hi-fi technique that IME often results in worse real world performance.

What should I be looking for in terms of minimum impedance for an 8 ohm box?

 

[Doufuss]

Minimum impedance is 5.4 ohms at 172 Hz. Then the flat line is right around 7.5 ohms.

 

[agedhorse]

It's a combination of minimum impedance AND the amount of bandwidth that resides below the nominal impedance value.

 

[Foz]

"I would eliminate the Zobel..."

+1

 

[Doufuss]

Thanks for the reply Foz. I went back over it and turns out I missed inputting the impedance plot for the alpha. Here's the new impedance graph

 

[Foz]

Did you take the zobel out?

 

[Doufuss]

I'm working on that. It's tough getting everything to work together. Ideally, a two way, 2nd order would be best, but I'm trying a bunch of different combinations of filters atm.

Back in a bit!

 

[agedhorse]

1. As is, that would be a nominal 4 ohm speaker by a wide margin

2. Because I am familiar with this driver, something is very wrong to end up with an inpedance curve as you have.

3. Quit playing around with Zobel networks and different combinations of filters and revert back to a simple pair of second order passive Butterworth or LR filters (or somewhere in between). I would also recommend underlapping the crossover points by about 5-10% on a Butterwoth type alignment.

4. I think you will find your impedance curves VERY different. You are currently creating a problem that doesn't need to exist.

5. I haven't looked at the lead-lag phase analysis but it's possible that there are some issues with this too, which can make some amps unstable.

 

[Doufuss]

Agreed. My goal is a two way 2nd order butterworth crossover.

Modelling it in a simpler tool like xsim or PCD give me very different impedance plots, and everything looks hunky dory.

With Vcad2 the enclosure tool models the in box impedance, and that's where it all of a sudden gets complicated. Here's the parameters: Woofer - 87l cab, Qa of 30, Ql of 10 (these are estimated) and tuned to 46 Hz. Mid - 4.8l cab, moderately stuffed with open cell foam.

Here's the 6a:



Here's the 3015:

 

[Doufuss]

Here's a model of a two way:



Not bad impedance wise, minimum is 6.1 ohms. But, there's a nasty spike at 1300 Hz that I'm not happy about.

Crossover:

 

[Doufuss]

But even at this the cab is, at best, 7 ohms. Increasing the cap on the woofer reduces the 1300 Hz spike a bit, but drops the impedance a bit from 6.1 ohms to 6 ohms.

Also, there is a big dip in the power response at around 700 Hz.

 

[Foz]

I think a big part of your problem is your 3015 impedance sweep still looks nothing like the one you claim to have cloned from here:

KAPPALITE™ 3015 - Loudspeakers | Eminence Speaker

 

[Doufuss]

I think a big part of your problem is your 3015 impedance sweep still looks nothing like the one you claim to have cloned from here:

KAPPALITE™ 3015 - Loudspeakers | Eminence Speaker

View attachment 3653135

The impedance graph IS different, because I modelled it in cab.

SPLtrace free air of eminence graph:

Winisd impedance graph in cab:

Vituixcad2 impedance graph in cab:

And yes, I did use that graph initially but it won't produce the correct impedance values to build a crossover. Because the impedance in box is very different than in free air.

 

[Doufuss]

Here is the graph in vx2 without the box:

 

[Doufuss]

Regardless, using the eminence supplied impedance plots for these two drivers drops the minimum impedance for this loudspeaker to 5.7 ohms.

Here:
And crossover:

It's a thorny problem.

 

[agedhorse]

Adjust the corner freq. of the mid driver up ~200Hz and add ~2 ohms in series with the mid HPF before the cap.

I'm still not convinced the impedance plots are accurate. With the 0.3 ohms DCR of the choke, the whole plot should be a little higher.

 

[Doufuss]

Adjust the corner freq. of the mid driver up ~200Hz and add ~2 ohms in series with the mid HPF before the cap.

I'm still not convinced the impedance plots are accurate. With the 0.3 ohms DCR of the choke, the whole plot should be a little higher.

Thanks Aged Horse. Adding that resistor really smoothed out the top end. Looks good.

Turns out, the DCR of the LP coil as per Erse audio is 0.178 ohms and HP 0.5 ohms.

Changing these numbers didn't help, the minimum impedance is still 5.7 ohms.

 

[agedhorse]

I would actually run an impedance sweep of the finished cabinet to see how far the model deviates from the real world.