Tapped Horn For Car

[mwmkravchenko]

I measured the inner driver alone mounted in the horn. You are right in that it does decouple from the driver in the mouth. You just picked the wrong end of the frequency response.



Of note is the little bit higher resonant point. The other thing to point out is the lack of the secondary peak. But when wired in reverse I get exactly the same peak as the driver in the mouth. The combined impedance graph has been posted.

The drivers are being effected by the air load presented by the coupling chamber. The response is shaped by the narow slot at about 1 cm wide. In simulations chaning this greatly effected the response.

Now I have to ask again where are the extra dbs coming from? The horn length is the distance from the begining of the coupling chamber to the end of the mouth. It is from this point that the first driver that is situated in the mouth of the enclosure is loading the box. That is a standard way of length measurement in a horn. The second woofer tapps the box at 1/4 of the way out or there abouts. This is not a perfect position for the driver. It is a practical one. Remember that this is an exercise in smallest volume biggest clean sound. The first design posted is a true horn with a falling response. I will build that one when I have thoroughly worked out what this little moster is doing. This box is not designed specifically to mirror a falling response in a car. But it does have by default a 12db slope/octave at the low end. It should be noted that driving a horn below it's resonant point on the low end is a good recipe for a quick death just as it is with a vented box. The cones flap and you don't get much sound.

Mark

[mwmkravchenko]

Does anybody who has been peeking at this thread have any experience with a H baffle? If you look at the so-called fold it could work out as a bit of an H-baffle. Just a thought.

Second fishing expedition is this. One gentleman pointed out that the drivers load the box from three sides. As this is true the radiation gained from the surfaces that actually drive the enclosed box volume summ and what do we get with this? Higher SPL???

This design is not a true acoustical impedance transformer or horn. As Eva pointed out there maybe a very peaked response on the output. That I have not verified with accurate measurements. I'm digging up my old program suite and hoppong that I still have some of my assorted toys to do this fairly quickly.

Mark

Thinnin is what happens when you sand cabinet parts all morning and your mind starts a wondering.

[Eva]

At this size, it's probably 6th bandpass with flared ducted ports what is going to help you most. Your current layout is quite close, but the inner driver suffers too much acoustic short circuit. You could make the duct longer...

[mwmkravchenko]

Ok generated the test signals. 30hz to 100hz by 5's Have printed up the graph paper and have the meter and voltmeter at the ready. So I will either find good or bad. But what I find I will report.

Mark

[mwmkravchenko]

I'm still not unhappy!

Here is the raw data:





So to put a marketing perpective on it we have =/- 3db from 40 to 100 hz. There is a nasty resonance in the car at these levels between 70hz and 85 hz. With some judicous use of a loudness contour the low end is decent down to 35 hz. So this little unhorn is doing what it was designed to do. Hell if I know what to call this little thing. It works as a narrow bandwidth subwoofer. And that is exactly what most subwoofers are asked to do. It is very efficient at 106 db/watt into 4 ohms and can stand a bit of eq that will tame the problems. Any sub measured to that level of precision in a car will wobble up and down in a similar fashion. So this is more than respectable for two 6.5" drivers in an enclosure that we don't know what to call. So I present the unhorn.

Mark

Time to get a movin for the plans you think?

[Eva]

You have to measure dB/W outdoors in half space to compare with other subs and to get an idea of how efficient it really is.

[mwmkravchenko]

Hello Eva

I designed this box for the car environment. All of it's design compromises are made so that it functions as designed in a car environment.

If you are informed enough to ask the question of why I did not measure it out in a half space environment then you will know that automatically without any corner reinforcement the efficiency will drop by 6db maybe even 8db. The low end will suffer to the tune of rapidly rolling off below 60hz instead of 30hz. So a loss of an octave of output.

This enclosure is a carefull dance between size and low frequency extension. I have carefully documented it. And to the credit of David McBean whose sofware helped me design this I was able to come up with a wierd little hybrid unhorn enclosure that gets some of the gain a proper horn loading would generate. The no free lunch clause is that it works only in a small acoustical space. It operates in a very narrow bandwidth. But a subwoofer usually operates over narrow bandwidth of 1 1/2 to 2 octaves anyway. 80-40 is one octave 40 to 20 is the second octave. This little box pulls off 1 1/2 octaves. Gives you a tremendous SPL capability and has a rather clean character of reproduction.

But the theme tapped horn for car maybe telling where it was intended to be used. Therefore that is where I measured it.

Mark

[Eva]

I have experience with cabin gain effects, I have done several custom subs for cars like yours during the past 10 years, usually with very limited space available. 6th-order bandpass always seemed to be the best approach because it provides high efficiency with low cone displacement.

But you can't compare efficiencies inside a car unless it's always the same car. 106dB/W may not be good at all for your cabin gain. Who knows if with some improvements you could be getting 110dB/W.

Of course, the frequency response will change in half space, but at least this allows to compare efficiencies to well known figures. If you get well over 90dB/W at 60-80Hz in half space, you can consider it efficient.

My half space target frequency respnse was usually maximum efficiency betwen 80Hz and 150Hz, falling 3dB or so to 60Hz, then another 12dB to 30Hz, then falling at 24dB/oct below that. In some cases I was able to approach 100dB/W half space at 80Hz.

Then cabin gain flattens it nicely...

[mwmkravchenko]

Fair enough Eva.

I apologise because I thought you were only trying to poke holes in the design.

I have seen the work you put into your electrical engineering projects and should take your opinion a little more seriously.

A bandpass enclosure will get you a maximum gain in efficiency of 7db over an octave and a half maybe two at most. asked to play loud out of the passband you have the food processor effect. Shredded woofers. I don't build that kind of box for anyone as kids generally destroy their woofers faster than they can open the box and change them out. Bye bye woofers. You are a brave man Eva.

As for increasing efficiency I played with the simm for a couple of days. You can only trade bandwidth for efficiency. No free lunch on this one. With greater efficiency came lower bandwidth. What I ended up making was what I felt made the best compromise.

For the sake of fairness I will test it outside ground plane with this caveat. I only have so much wire comming out of the car. So it will not be half space as in against a wall floor juction but a true ground plane with no boundary reinforcement. So what ever I come up with will need to be bumped up 3 db to reflect a floor wall boundary effect.

Mark

P.S. A sealed box version with two drivers maybe a simple test to see the cabin gain on an equal footing. Actually that could be easily modeled in a number of programs that I should install. Will be back on this one.

P.P.S. Eva don't you sleep sometimes? What time is it over there?

[Eva]

I think that your box design does not correlate well to the hornresp model, it could probably be improved by extending the first section, that's why I ask you to measure "unbiased" efficiency, and compare to hornresp results (half space too). If it's close or better, then the box layout is ok. I bet you will get a few dB less than predicted.

I didn't build one note boxes, tunings were well spaced, typically around 37Hz and 100Hz for minimum overall cone displacement, with moderate resonant Q. Some of my layouts were not very different from yours. The cabin and the proximity of surfaces inside the car detunes the "ports" to the low side, usually requiring fromt chambers just like yours, without a port or even a wall sometimes. We only had problems with one crazy kid blowing speakers until we found something with a big enough voice coil.

btw: I have strange sleep time (it's 5 AM or so). My current main customer is overseas and is probably having dinner now.