“Yeah Science, Bitch”  -- Jesse Pinkman

I did some research. The concept of tuning a vented box very low is called EBS (Extended Bass Shelf). This site explains it pretty good, the basic idea is getting more bass on low frequencies. You basically take an enclosure 160-180% the size of a “maximum flat” setup and then tune it close to Fs (Free Air Resonance of the speaker). As I already found out the previous owner did tune the enclosure to about that frequency. But how about enclosure size? The manual suggests “maximum flat” enclosures from 64 to 92 litres, suggesting an EBS setup between 102 and 165 litres. The enclosure has around 80 litres not accounting for the speaker and the vents. Well, I guess I found a problem here.

I tried some more online calculators and plenty of them tune close to Fs without even mentioning it. So it’s more than understandable to end up with this setup. But as always, you gotta trust your senses: If it sounds shitty there has to be a reason… Ah, and trust the engineers. If you deviate far from their suggested path you better know what you are doing…

So to wrap this up and continue other important work I’ll go the humble approach. I’ll measure the actual resonance frequency of the enclosure and tune it to something between 36-39Hz for the suggested “maximum flat” response.

Measuring shouldn’t be to difficult, the impedance of the speaker drops at the resonance frequency, so the current spikes ( I = U/R ). So just rig an ampere meter to the speaker, a frequency generator to the amplifier and play with the frequency knob until you reach maximum current. Well, my ampere meter doesn’t like low frequencies at all making it impossible to get a reading. Using a shunt (resistor in series) and measuring voltage I pinpointed the resonance frequency at 35,7Hz. That’s way lower than the expected 43Hz. Good news: That’s the lower end of my target frequency. Bad news: I’ve got to redo the math…

Some site suggests if the distance of the ports to any wall is less than their diameter the tuning gets lower. That’s definitely the case here…  The increase of Volume by damping depends strongly on the material, ranging from 0-20%. If I run the numbers with an estimated volume of 115 litres everything checks out and the estimated resonance frequency with the long vents would be around 24Hz. I guess I have to measure that as well 🙁 Hmmm, 27Hz, not too bad for an estimate…

Hmm, maybe I should give the 31Hz a shot. 18,5 cm for the vents… Let’s say 20 for good measure… After some try and error 21cm turn out to lead to 30.something Hz. Good enough. Yeah, there is more ooompfff on low frequencies, you don’t hear some but feel it in your guts. I had to adjust the woofer level back to +6db on the amp. But it seems to be a trade-off with quality. Most of the time there’s not enough bass and then suddenly it’s killing you when the frequency drops low enough. For example with “Clint Eastwood / Gorilaz”, some low rumbling for the whole song and then the break with a couple of low beats is just plain awful. Somebody suggested that EBS setups often have cancellation issues on the two octaves above the resonance frequency (so 30-120Hz). Yeah, that might be the issue here.

To cut a long story short: EBS doesn’t work out with this enclosure, even if you do it right. I guess that’s 2:0 for the engineers… Disclaimer: I’m working low volume settings (neighbours…) and the placement of the sub is bad. Maybe it’s getting better with optimal placement and if you crank up the amplifier. Maybe not. At least now the sound setup at my desk is ok. Time to actually do some work…

The new subwoofer has bass, I’ll give it that. But sounds like a trumpeting elephant somehow. Well, it’s probably not the speaker itself. The broken crossover / connection panel is already removed. Hmm, let’s have a look at the mods of the previous owner.

He reduced the volume by adding an enclosure for the connection panel, added some foam isolation and adjusted the length of the vents. Alright, the enclosure lowers the volume (obviously), the insulation increases the (virtual) volume of the enclosure (not so obvious). Measuring inside the box is difficult, I’d estimate the volume about 70 litres so with the insulation it’s probably somewhere between 70 and 80 litres. The original vents have 75mm diameter and are 120mm long. The previous owner extended them to 350mm / 360mm using pipes from the hardware store and claims to have calculated everything with some software.

The vents should be the same length. They are extremely close to the rear wall of the box but I couldn’t find proof that this is an issue. Let’s start with the obvious: Some designer made this enclosure and even if it’s a cheapo they probably put some thought in it. The vents adjust the resonance frequency of the box. Changing them by a factor of three means either the engineer who build it was way off or the one doing the modification. I used an online calculator to run the numbers. With the extensions the resonance frequency (fb) of the enclosure is around 29Hz, without approx. 43Hz. The manual of the speaker comes with 12 suggested enclosure dimensions ranging from 36 to 50Hz fb. The interwebs suggest it makes some sense to go for a “bass shelf” by tuning fb to Fs (30,83Hz) or even lower. One formula I found suggests 31Hz being ideal for this setup.

The manufacturer of the speaker and the manufacturer of the enclosure obviously don’t think so. I removed the extensions and the setup is quite close to the “Standard Tight Bass” box design of the speaker manufacturer. Afterwards I had to reduce the amp setting from +7 to +3 (probably db) and it sounds pretty tight.

I guess I’ll rig up a setup to actually measure the resonance frequency to check my estimates and tune the box to “optimal sound quality / flat response”.

"He is dead, Jim" -- Leonard McCoy

(posted a week or so after writing…)

I thought this might have been the first winter (ever?) without getting sick. Didn’t work out 🙁 Whatever… After even netflix became boring I had to something else.

My stereo hasn’t been the same since I moved home. Today I finally looked into the issue. Some peak performance testing before moving to Hannover fried one tweeter completely and hurt the other one badly. The automatic eq / analyzer of the amp covered this failure up somehow (being a four way system the mid speaker range is overlapping with the tweeters). Since I use the amp mostly for surround sound, the other speakers helped covering it up as well. But after I bought some JBL Control One for my Desk it became quite obvious that some frequencies are missing. I still had some tweeters from the stereo of the 190 series ( a car…) catching dust so I went for a quick fix. It turned out they are too noisy but I got them more or less under control thanks to the equalizer.

Linux provides some great tools for audio testing and they are for free. So I cobbled together an audio analyzer using the mic which came with the amplifier for setting up surround sound. The following are all frequency diagrams, x-axis is frequency, y-axis is output volume. The actual numbers don’t matter, what we are looking for is a more or less straight line indicating that the pink noise is reproduced by the speakers “as is”.

From left to right: original tweeter, audio system, gradient

The Speaker with the completely fried tweeter. Note the drop!

Quick fix: The audio system tweeters are a little bit small for the Infinity Reference 61 Mk II Chassis

Audio System tweeters removed from the old Merc. Not too bad at all but a little bit too much. Probably because they are build for a two way system there’s too much overlap on a four way.

Fancy Gradient Tweeters from ebay, used. I don’t like the peak around 8k.

Up to now the audio system + duct tape did a superior performance than fancy (and well fitting) Gradient Tweeters. Since both speakers have different volume levels on high frequencies (which you can’t see on the diagrams) there’s still something wrong. More investigation has to be done I guess. To be continued…