DiyMobileAudio.com Car Stereo Forum banner

1 - 12 of 12 Posts

·
DIYMA Founder
Joined
·
4,757 Posts
Discussion Starter #1 (Edited)
This is a short, simple tutorial on how to interpret the results of your typical subwoofer box modelling software such as Winisd.

Here's our good old trusty Ascendant Audio Atlas 12" modelled up using the manufacturer supplied t/s parameters and Winisd. Notice the shape of the rolloff, and an F3 point of about 52hz. Not bad.



Now take a look at the actual, measured frequency response (semi-anechoic) at 1 watt, 20 watts, and 75 watts of power.



First off, you can see that while Winisd was fairly accurate, there was still some differences between the actual response and the calculated response. Note that the actual -3db point is around 56hz vs. predicted Winisd 52hz, and the actual -6db point is ~48hz vs. Winisd's prediction of 40hz.

This is due to variance in t/s parameters between the actual sub, and the manufacturer's specifications. We can also see that the frequency response changes with different power outputs. At 400 watts of power I'm sure the response looks quite different, but unfortunately my amp would only do 75 watts before clipping :)

Also, it's pretty evident that Winisd does not take into account the upper end response of the driver.

Now let's look at the same sub in the trunk of a Honda Accord (Gold line).



Now that looks nothing at all like what we modelled :) So you can see that the car's interior itself and placement has the greatest effect on the frequency response.

So to sum up, what do box modelling programs like Winisd tell you?

They can tell you:

1. The low end frequency response of the box+sub "ONLY", based on the manufacturer's supplied t/s parameters. The actual t/s parameters can vary by quite a bit, and will change the low end frequency response. For a ported enclosure, I wouldn't even think of using the manufacturer supplied t/s parameters, but rather I would measure the actual t/s specs myself. Reliable manufacturers are generally around 10-15% deviation from what I've noticed, while I've seen some t/s parameters that were off by at least 50%.

2. Frequency response changes with power output. The more linear your subwoofer's motor, the less shift you will see as power increases.

3. Frequency response changes DRAMATICALLY with the listening room/car, and placement.

In conclusion, Winisd can be a powerful tool for comparing different driver's low end response in a variety of enclosures against each other, but the actual frequency response in-car is going to depend on quite a few more factors that Winisd(and similar programs) cannot account for.

One last thing that I forgot to mention. If you subtract the in car response from the sub+box response, you can find the transfer function of your vehicle. Applying this transfer function to Winisd's calculated response can give you a much better estimate of how a sub wil sound in your car.
 

·
DIYMA Founder
Joined
·
4,757 Posts
Discussion Starter #3
Hobbes26 said:
Cool test. Can you elaborate more on the measurement conditions and methods?
Sure. For the box+sub I just used a nearfield technique. Basically, you just measure the speaker up close to the cone to swamp out all the room and baffle effects. It gives you a very clean, artifact free frequency response. It's the closest I can get to anechoic without going outside to some field.

For the car measurement, I placed the sub in the trunk facing the rear. Mic was a little above the driver side headrest. Standard "rta" time window of 300ms. Interesting how the combination of seats and rear firing just kills the top end response, limiting it to about 80hz before it drops off at about -24db/octave.

I probably should have used a crappier sub to illustrate some points such as t/s parameter shift and bl compression, but I didn't have any on hand :)
 

·
Registered
Joined
·
104 Posts
Very good :)

This should be a must-read for all the beginners who uses WinISD or similar to give them a "final" optimal box size. "what do you mean it will sound crap in my car? WinISD suggests a 30l closed box and the graph says it has a -3db point at 26hz!" and so on..

Can the transfer function be extracted from the measurements in some way, so that it later can be included in the rest of the simulations i.e. in WinMLS?

I really must get some software and measuring equipment before taking on a car stereo project...
 

·
DIYMA Founder
Joined
·
4,757 Posts
Discussion Starter #5
epifant said:
Very good :)

This should be a must-read for all the beginners who uses WinISD or similar to give them a "final" optimal box size. "what do you mean it will sound crap in my car? WinISD suggests a 30l closed box and the graph says it has a -3db point at 26hz!" and so on..

Can the transfer function be extracted from the measurements in some way, so that it later can be included in the rest of the simulations i.e. in WinMLS?

I really must get some software and measuring equipment before taking on a car stereo project...
Subtract the in car response from the anechoic response and you have the transfer function.

Also using a more advanced box modelling software such as Klippel or Leap will allow you to see how the response changes with different power levels as well as getting a realistic picture of distortion vs. spl.
 

·
Registered
Joined
·
104 Posts
Re: What do Winisd/box modelling programs really tell you?

BTW: is it just me who`s noticed that WinISD says "doing nothing" on startup? :D
 

·
DIYMA Founder
Joined
·
4,757 Posts
Discussion Starter #8
The most brainless, easiest way is to go to Partsexpress.com and buy the Dayton woofer tester. It's a worthy investment if you're going to be doing it on a consistent basis.
 

·
Registered
Joined
·
31 Posts
It should be noted that WinISD can not show inductance drop offs

however that FR graph you measured is identical to what Bassbox6 shows when you input inductance.

Also you have to remember that inductance is frequency dependent and shorting paths and other additions will not show up on FR graphs with much magnitude till after 300hz.
 

·
Registered
Joined
·
15 Posts
Great, I thought I was making some headway. I'm building my first sub box and want to do it right the first time.

I just started using Winisd attempting to design a sealed enclosure for a 12" sub. I was using boxes from .8 to 1.0 cubic foot sealed. After all the setup and learning all the assumptions are worthless?

Initially I was making sure the cone excursion did not exceed the Xmax limits of the subs I was modeling. I used a LP xover to make sure the subs stayed below their Xmax limits. I this a what I should be doing or am I going about this the wrong way?
 
1 - 12 of 12 Posts
Top