DIYMobileAudio.com Car Stereo Forum banner

Measuring cabin gain? Need a little advice from the experts.

17K views 14 replies 10 participants last post by  mikey7182 
#1 · (Edited)
Ok guys... I'm trying to interpret what I just observed.

I decided to create a little experiment to (hopefully...) measure the effects of cabin gain so I might take it into account in building my next enclosure.

So here's the premise of the experiment.

#1. Using a sound level meter, measure sine test tones ranging from 20Hz to 95Hz in 5Hz increments with the subwoofer in the car, all windows/doors closed, with the meter exactly two feet from the cone of the subwoofer. Volume set at head unit at 30.

#2. Using the same meter, measure the same test tones with the subwoofer enclosure sitting on my driveway facing away from the car. Meter is again two feet away from the center of the woofer's cone. Volume set at head unit at 30.

The same enclosure was used for both tests. Unfortunatley the specifications of the box elude me other than it is aprox. 1.3-1.5 ft^3 with a slot port. I would have liked to have used a sealed enclosure for this experiment but this box was sitting on the garage floor and since the box is the same for both tests I would assume the size and type of enclosure is a moot point since it is the same for both tests. Crossover frequency on the amplifier was the same for both tests (120Hz low-pass 12dB/oct)
Here's what I observed.




I expected to see an increase in sound level at some point. (Apparently 35Hz-55Hz in my vehicle, +5.8dB at 45Hz) from What I did not expect to see was a decrease. Below 35Hz and above 55Hz this experiment shows that I actually have a decrease in sound level. Most notably around 80Hz.

What this seems to suggest is a tuning frequency of 80Hz would help level out that graph... Doesn't seem right to me.

Can one of the wizards in here explain why I'm seeing that? I believed that cabin gain would provide a boost to all frequencies? If I am correct in that statement, what other factors am I missing here?
 
See less See more
2
#2 ·
The problem is ported designs do not effectively use cabin gain nearly as well as a sealed sub. Just because you have a dip at 80hz doesn't mean you would tune your box to that. You would loose a lot of output below 80Hz if you did that. That would be in the roll off of the sub response when crossed over anyway. It means you wouldn't have to use as steep of a slope. Try to seal that port and redo your measurement, see what happens then. If it were me, I would crossover my sub around 40Hz to integrate with my other speakers for a flatter response. By the way a hand held meter will not give an accurate response because of the filter weighting inside. If you have a microphone and a laptop give Room EQ wizard software a try.
 
#9 ·
The problem is ported designs do not effectively use cabin gain nearly as well as a sealed sub. .
This is not correct. The car doesn't change depending on the box or the style of the box, unless the box is so large that it diminishes the volume of the car substantially. I doubt the box is 50% of the volume of the car.

Here's how you can make a fairly accurate measurement, so long as the woofer and the port are on the same baffle:

1. Measure the response in the car with the box in place and the mic at the listening position. Distance doesn't matter.

2. Place the box outside on the ground. Place the mic 2 meters away also on the ground, facing the woofer. Use the same volume setting and the same sensitivity setting. The ground plane measurement will give you an anechoic response equal to a measurement 1 meter away. The box must me far from other obstructions. 20 feet from big walls or a car should do it.

If the port and the box are NOT on the same plane, then separate measurements of the port and the woofer must be made and spliced together. This is beyond the scope of what you'll be able to do accurately.
 
#5 ·
Wow. This response is pratically identical to my car.

I tried to resolve it by high passing my sub @ 50Hz and turning down the 45Hz EQ on the amp to flatten out that hump. The trying to increase the midbass @ 60Hz on the HU eq to try and get that dip @ 80Hz up a little but it would be improve more to have an EQ with more precision to iron out the bass response.
 
#6 ·
It doesn't look right to me. You should have some gain in your car compared to outdoors not the reverse. You should measure at the position that your head would be in car. What SPL meter did you use and what were the settings on it? Weighting, fast, slow, range, etc. Was the car sealed up? Did you use the exact same set-up for both tests? (seems like you did, just checking)

If you have a laptop or a computer that you could drag outside long enough to do the tests, you can get an external soundcard and download REW (Room EQ Wizard) or an RTA program for free and run much more accurate tests in a fraction of the time it takes to manually do them with a meter.
 
#7 ·
I'm a real zealot about measuring speakers... I've literally done *hundreds* of measurements for my current project.*

But in this situation, you're better off using JBLs data. Just find your car on the graph below:



The reason why I would recommend going this route is that cabin gain can vary depending on where the mic is placed. In addition, measuring the low frequency response of speakers is not trivial; there are entire books on this subject.**

To make a long story short, calculating cabin gain is fairly complex, as you've found out. Unless you're interested in spending a few days learning about gated measurements and splicing together data points from the port and the cone, just use JBLs data.

* my current project: Creating a Soundstage with Waveguides and Psychoacoustics - diyAudio.com Forums

** the book I'm referring to: Powell's Books - Testing Loudspeakers by Joseph D'Appolito

Here's an article on testing loudspeakers from Stereophile. I personally learned from the D'Appolito book, but hey, free is free right? :D

Stereophile: Measuring Loudspeakers, Part Three
 
#8 · (Edited)
Jeez. Sounds like i opened a giant can of analysis. I don't really have the equipment required to accurately preform this test.

JBL reccomends a tuning frequency of 34Hz and an enclosure of 1.75 ft^3.

That gives F3 of 29.5 Hz (aprox) and +4.5 at about 50Hz. Here's the graph.


Now if anything I measured is valid (which it might not be and probably isn't).... around 55Hz I should see the largest boost from cabin gain as well... Is that going to affect my goal of having a linear graph? Should I tune it lower? The graph definatley looks it's prettiest for 1.75ft^3 at 34Hz... I also tried 38Hz for comparison as someone in another thread mentioned a few Hz above Fs might be advisable. F3 of 31Hz and +5.7 at 50Hz... Seems like 38Hz would be the louder of the two...



So from an SQ perspective... I'm assuming the lower playing box would be preferred?
 
#10 ·
Oh, for JBL woofers, use the box volume and the tuning specs in the manual. I design the boxes in the manual to give the widest bandwidth possible with a port that will fit in the box pretty easily and to use the port to minimize excursion over as much of the useable bandwidth as possible. This minimizes distortion, maximizes output over the widest bandwidth, maximizes power handling and eliminates the need for a subsonic filter. Also, the in-car graph shown in the manual is the modeled response of the woofer modeled with the average of the graph that Patrick posted. If you cross the woofer over below 80Hz, the larger variations in the upper part of the bass won't have much influence and you'll discover that the model in the manual is pretty close to what you'll hear in terms of response shape, but overall level will depend a bit on the size of the car.
 
#13 ·
I have been having the same issue, although because I viewed the specs a few days ago, if I go directly to the page in my history, it comes up fine. Here is the link to the W15GTi:

JBL Car & Marine Audio

I do have a question, and this could point more to an inaccuracy in WinISD or my measurements than the spec sheet for JBL. The specs recommend that for the W15GTi, a ported enclosure of 4.0ft^3 is used, tuned to 28hz, with a port diameter of 6" and 15.13" long. Is Vbox in this example gross volume, or net after sub/port displacement? If Vbox in the diagram is gross volume, I should end up with a net volume of roughly 3.75ft^3. If Vbox is net, I should start out with a gross volume of 4.25ft^3. I have modeled both options in WinISD and come back with tunes of 31 and 32hz, respectively. In order to achieve a 28hz tune with those port dimensions, the enclosure would have to be 4.96ft^3 net. Am I doing something wrong, or is WinISD off a bit? Not that it matters terribly to be a few cycles off, but if the goal is to build the enclosure to spec, I'd like to know. :)
 
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top