Play with Crossover Design and Simulation - Free

[ErikSquires]

Since someone will ask, everything here is free, and I make no money nor do I promote a service whether you build the LM-1 speaker kit or not. I'm just trying to promote audio education.

For this you must have Windows (probably works with Wine too). :( I'm sorry.

Have you ever been curious about how all the parts and charts work? Want to get behind the scenes of the reviews you read about? Well, here is your chance.

I have posted the design for the LM-1 speaker on line. It is a free, 2-way design specifically intended to be used on a bookshelf or desktop.

http://speakermakersjourney.blogspot...f-version.html

However, I have also made the simulation files available so that novices can learn about crossover design, and experienced speaker makers can play with their own crossovers. One of the very difficult tasks in speaker design is correct electro-acoustic measurement of all the parts but for the LM-1 I've done all that for you.

The simulation files are available from DIYAudio in this thread.

http://www.diyaudio.com/forums/multi...ailable-2.html

To use them, you'll need to download XSim from here.

http://www.diyaudio.com/forums/softw...-designer.html

Any questions please post in the DIY forums.

[Millstonemike]

Erik, I do believe your on to something here. Starting a design from scratch for the first time can be daunting. This is a way to see the guts of a working design. The learning by experimenting with incremental changes ...

[tomzarbo]

Erik,
Good idea. Being able to show how measurements and modeling software was used to design/build a speaker from the ground up can only be helpful. I envisioned the same thing with a more reasonable speaker such as the Overnight Sensations, since many people have those. I don't consider myself quite there yet with measurements and simming to pull it off.
I'm glad you're making this public like this, it's worth a close look. I bet it's real educational. Thanks for doing this.

TomZ

[ErikSquires]

Thanks guys.

Tom, the LM-1 and Overnight Sensation are very close in crossover complexity. Not that you should ever evaluate a design based on part count, my point was just that they are both examples of simple 2-way crossovers. One thing I was able to do, luckily, was keep both drivers in positive polarity in the LM-1. You'll see that in the measurements page.

[ErikSquires]

Millstone,

Thanks, you put into words what I really could not. That is the goal, to let the interested do as little or as much as they'd like to get a real visceral feel for how all the components work and interact electro-acoustically. Maybe it will also encourage more to build the LM-1 speakers too and get hands-on with crossover mods.

Best,


Erik

[PWR RYD]

Why do you feel it's lucky that you have both the drivers wired in positive polarity? Is it then unlucky in your mind if the filter design requires one of the drivers to be wired out of phase?

[Paul K.]

Good question.
Paul

Why do you feel it's lucky that you have both the drivers wired in positive polarity? Is it then unlucky in your mind if the filter design requires one of the drivers to be wired out of phase?

[ErikSquires]

PWR,


The very short answer: It is my impression (but not based on any rigorous study) that small 2-ways often have to invert the tweeter.

To make this alignment possible much had to come together, such as driver spacing, surface mounting the driver, etc. So, it's pretty lucky.

Does having positive polarity on all drivers have any audible benefit? I doubt it. I'm just happy about it. I would not use it to claim it's better or worse than any other 2 way speaker or kit. I just wasn't expecting it.


Best,


Erik

[corradizo]

Thanks so much!! I am very excited to work my way through this!! I'm just starting out, perfect timing.

[ErikSquires]

Over at the DIY forums we also started a thread on analyzing the Seas A26 kit and alternatives with XSim if anyone is interested.

[Brian Steele]

Over at the DIY forums we also started a thread on analyzing the Seas A26 kit and alternatives with XSim if anyone is interested.

How do you model the effect of a z-offset between the drivers in XSim?

[ErikSquires]

How do you model the effect of a z-offset between the drivers in XSim?

Pretty much any change you want to do to a part, if you can make the change, is through the "Tune" menu. Right click on a part and select "Tune."

When you select a driver you'll see the distance entry which is called "mod delay" While you can add delay to any driver, I set the tweeter to 0" and the woofer to 1.1"

[ErikSquires]

In case anyone is interested, I posted a separate tutorial on measuring speaker distance here. You can try the example yourself with the XSim files provided. Delete all the crossover parts and set your speaker distances to 0".

[bohdan]

Hi Erik,

Assuming, that your XSim blue plot shows summed "Woofer alone" + "tweeter alone" response, then this simulation plot should be exactly the same as your red measured "Woofer + tweeter".
These plots should be the same without any additional distance shifts, because the 1.32" woofer distance is already embedded in all woofer measurements that go into your simulation.

Best Regards,
Bohdan

[ErikSquires]

It may help you to undestand that OmniMic which was my measurement tool, captures relative phase but not time (i.e. group delay). The reason is it does not generate a signal, but instead relies on a test CD with signals to be played, so omnimic has no way of telling how long sound took to travel from driver to microphone. OmniMic FRD files include only frequency, phase and amplitude.

When you measure the tweeter and woofer together you now get the interference pattern caused by the woofer's delay, or the red trace as it were. Now this is a little confusing, but Omnimic will see the entire speaker as 1 driver, it can't tell what's woofer and what's tweeter. It has no way of going "Aha! The woofer is 1.8" behind the tweeter!"

When you first input the separate frequency and impedance files in XSim the combined simulated frequency response will look like the blue trace, but of course, since the delays are wrong, it won't match reality. It's up to us to input or infer distance into the "mod delay" setting. As you approach the correct delay the simulation will slowly approach the actual measured response.

Of course, other measurement tools may make direct measurement of distance easier. If you find this helpful, please send a Pocket CLiO to my address at..... (Actually I think it also has this issue).

By the way, the simulation in Xsim correlated perfectly with the measured responses of the finished speaker, within the limits of in-room measurements, so, except for the missing port contributions, I'm pretty confident in the simulation.


Best,


Erik