How so? Bandpass is an efficiency boost and in fact I would say that I could get as much as you are getting but in a much smaller package.
I don't see where the efficiency comes from. Because the "port" is a horn? But at these frequencies the horn is just a single lumped mass, just like a port, there is no gain.
I don't see where the efficiency comes from. Because the "port" is a horn? But at these frequencies the horn is just a single lumped mass, just like a port, there is no gain.
gedlee said:
The efficiency bump is because we're getting useful output from both sides of the cone.
This was something that I found interesting about dipole subwoofers, which I'd never considered. If the baffle is large enough, it can be more efficient than a sealed box, because we're getting output from both sides.
I modeled this in Mathcad using Martin King's worksheets and it confirmed this.
(www.quarter-wave.com)
Upstairs in my new home I have a baffle that's the size of my king bed that's a dipole sub. It's literally under the bed - talk about stealth! The efficiency on this thing is HUGE. It literally shakes the house. Of course part of this is because it's radiating right into the floor.
This set of subs will go in the living room downstairs, which is where I have the Summas.
Now you could argue that a DUAL reflex bandpass would be nearly as efficient, which is true, but I've always thought they sounded "weird." I built a few back in the early 90s, when Peter Mitchell turned me on to your work. That was actually what inspired me to learn your work, was reverse-engineering bandpass subs fifteen years ago.
John
This just doesn't sit right with me. Dipoles have a very low radiation efficiency so it seems to me that they could never equal a monpole. I've always thought that the "both sides of the diaphragm" argument was flawed because it ignores that fact that these two sides are out of phase and that getting them in-phase to radiate efficiently is no small feat. The dual bandpass system cannot achieve any greater output that a single, it can just extend this output a little lower - just like a ported enclosure, but it does so at the cost of a dipole roll-off rather than a monopole.
I have found since I switched to all monopole outputs that I prefer the results.
The Acoustic Lever - now that does add a lot of efficiency. Real honest to goodness monopole efficiency. Some day those will dominate the landscape.
This just doesn't sit right with me. Dipoles have a very low radiation efficiency so it seems to me that they could never equal a monpole. I've always thought that the "both sides of the diaphragm" argument was flawed because it ignores that fact that these two sides are out of phase and that getting them in-phase to radiate efficiently is no small feat. The dual bandpass system cannot achieve any greater output that a single, it can just extend this output a little lower - just like a ported enclosure, but it does so at the cost of a dipole roll-off rather than a monopole.
I have found since I switched to all monopole outputs that I prefer the results.
The Acoustic Lever - now that does add a lot of efficiency. Real honest to goodness monopole efficiency. Some day those will dominate the landscape.
gedlee said:
The tapped horn is quite a clever invention I'd say. It's true that it's challenging to get both sides to radiate in phase, but I'd have to say that Danley has come up with a clever way of doing it.
99% of the programs out there ignore where the port is located. One thing that's intrigued me about hornresp and Martin King's worksheets is that they factor it in.
For instance, let's say you build two vented boxes. Each has the same driver and the same volume. One has the port located near the cone; the other is an eight foot tall sonotube and the port is located on the other side of the enclosure. The in-room response of the sonotube sub will be different because there's a 7msec delay introduced by the distance from the port to cone. This delay will do some interesting things to the in-room response. In addition, the *direction* of the port has some interesting ramifications, due to high frequency peaks radiated via the port.
King's software in particular is quite clever, and easier to use than hornresp IMHO. Unfortunately it can't model tapped horns. It CAN model transmission lines, and I've long thought that the tapped horn is quite close to a transmission line (but not quite.)
If I can find a few minutes I should throw together some comparisons with Martin's software to see how it models the hypothetical vented box described above.
I fired up my favorite program for modeling bandpass subs quickly - bandpass boxmodel. I actually bought this program when I was a teenager, as mentioned above, was trying to reverse-engineer bandpass subs. Nowadays it's free.
Using a "textbook" bandpass we can get fairly flat response, as the pic shows. The efficiency is very VERY low though; about 82db. The box is impressively small though; just a hair over four liters! These MCM subs love stupid small enclosures.
The F3 is too high though; just 49hz. Not much of a "sub" woofer.
To optimize this sub I'd probably do what I did with the tapped horn - add a resistor to increase QTS, which will extend the low end response. The resistor will also make the enclosure size more manageable, as a five liter enclosure contributes to a ridiculously long port.
Using a "textbook" bandpass we can get fairly flat response, as the pic shows. The efficiency is very VERY low though; about 82db. The box is impressively small though; just a hair over four liters! These MCM subs love stupid small enclosures.
The F3 is too high though; just 49hz. Not much of a "sub" woofer.
To optimize this sub I'd probably do what I did with the tapped horn - add a resistor to increase QTS, which will extend the low end response. The resistor will also make the enclosure size more manageable, as a five liter enclosure contributes to a ridiculously long port.
Attachments
I've spent more than a little time with transmission lines and from what I can tell they just don't pan out. Sure a port some feet away will be "different", but thats totally unpredictable and completely room dependent. There will be modes that this distance will excite better and those where it will excite less. I just don't see that a general trend can be established.
To me there is simply no way to uncouple the source problem from the room problem at very low frequencies and I am now firmly of the opinion that multiple monpoles cannot be beat for efficiency and smoothness. It would take a lot of data to get me to believe otherwise since I have a lot of data to support this position.
SPEAK has always been able to model transmission lines and in playing around with them I was never able to get anything useful. They can even do damping at different locations in the line (although this feature is not obviuos as to how to imliment).
I looked at the tapped horn briefly and I agree that to me it was a transmission line. Horns at low frequencies, like in your example, just don't work as horn. They are too small and only act as very big ports. Of course big ports are a good thing - lower loss - but a straight port is a lot easier to build if it works the same as a horn.
To me there is simply no way to uncouple the source problem from the room problem at very low frequencies and I am now firmly of the opinion that multiple monpoles cannot be beat for efficiency and smoothness. It would take a lot of data to get me to believe otherwise since I have a lot of data to support this position.
SPEAK has always been able to model transmission lines and in playing around with them I was never able to get anything useful. They can even do damping at different locations in the line (although this feature is not obviuos as to how to imliment).
I looked at the tapped horn briefly and I agree that to me it was a transmission line. Horns at low frequencies, like in your example, just don't work as horn. They are too small and only act as very big ports. Of course big ports are a good thing - lower loss - but a straight port is a lot easier to build if it works the same as a horn.
Patrick Bateman said:
John
Send me the driver data and I'll see what SPEAK says. I don't trust other programs as I have found them incorrect on too many occasions.
What you are showing looks very bad to me. Thats way too much bandwidth for a bandpass. They don't work well beyond one, maybe two octave. But in a narrow bandwidth they really do work well.
I'm really not trying to scuttle my own project here, but a one dollar resistor makes that MCM woofer VERY attractive in a bandpass.
I've attached a pic which shows the response of the SAME woofer once you bump up it's QTS to 0.6 via an increase in QES.
The difference is night and day - it extends the response a full octave!
Of course you need an enclosure that's five times as big, but that's expected due to the increase in QTS.
The bottom line is that you're looking at a $25 driver that can reach down to 25hz in an enclosure that's barely over a cubic foot.
In 2004 we had a discussion about buying a bunch of cheap subs to implement your multiple sub idea, and this sure looks like a good candidate. I have a hunch that three or four of these $25 subs in small bandpass boxes would clobber the sound quality of a "cost-no-object" subwoofer.
Of course the rich guys should go out and buy three or four "cost-no-object" subs, but that's a whole 'nother ballgame.
I've attached a pic which shows the response of the SAME woofer once you bump up it's QTS to 0.6 via an increase in QES.
The difference is night and day - it extends the response a full octave!
Of course you need an enclosure that's five times as big, but that's expected due to the increase in QTS.
The bottom line is that you're looking at a $25 driver that can reach down to 25hz in an enclosure that's barely over a cubic foot.
In 2004 we had a discussion about buying a bunch of cheap subs to implement your multiple sub idea, and this sure looks like a good candidate. I have a hunch that three or four of these $25 subs in small bandpass boxes would clobber the sound quality of a "cost-no-object" subwoofer.
Of course the rich guys should go out and buy three or four "cost-no-object" subs, but that's a whole 'nother ballgame.
Attachments
dammit, i'm really starting to wonder if I should build a bunch of these. Here's a pic of the excursion at 25 watts. Keep in mind we're not going to feed these with much power, since I'll be using a minimum of three.
In my humble opinion the biggest advantage of the bandpass sub is something you haven't discussed. The biggest advantage of the bandpass sub is that it's STUPID easy to build. I've made bandpass subs out of sonotube in an evening.
The trickiest part about this tapped horn that I'm building is that it's sooooo looooooong. Because it's eight feet tall, sealing the panel inside of the tube is going to be a b1tch. And if you skimp on the seal, the whole design is RUINED. Any kind of leak is going to be a BIG problem.
My plan was to cut access holes every three feet, seal the inside panel, then replace the access panels. Definitely a p.i.t.a.
In my humble opinion the biggest advantage of the bandpass sub is something you haven't discussed. The biggest advantage of the bandpass sub is that it's STUPID easy to build. I've made bandpass subs out of sonotube in an evening.
The trickiest part about this tapped horn that I'm building is that it's sooooo looooooong. Because it's eight feet tall, sealing the panel inside of the tube is going to be a b1tch. And if you skimp on the seal, the whole design is RUINED. Any kind of leak is going to be a BIG problem.
My plan was to cut access holes every three feet, seal the inside panel, then replace the access panels. Definitely a p.i.t.a.
Attachments
gedlee said:
The extremely wide bandwidth in that first plot is because it's a "textbook" design which would never work in the real world. The volume of the chambers is so small, it would never behave this way in real life. It's because the MCM woofer has a tremendously low VAS combined with a very low QTS. An author for Audio Xpress found this out the hard way a few years ago when he tried to put it in a vented box. The vent wound up being larger than the enclosure!
The resistor seems to help a lot.
Here's the specs I'm using: http://www.diyaudio.com/forums/showthread.php?postid=1565560#post1565560
I personally measured three or four, they're all very close to what Krutke measured.
cm^2
I found the driver on the web. Very attractive.
I'll look into making a sub with two facing a central ported enclosure. This will cancel the box vibrations. I'll do a broad band sub as I think that these are the most important and would work best with this inexpensive driver. Why two? Because I doubt that one could take the output of the plate amp - I know what they claim the power handling is, but I also know what to expect from a 2" voice coil and a cheap driver.
I found the driver on the web. Very attractive.
I'll look into making a sub with two facing a central ported enclosure. This will cancel the box vibrations. I'll do a broad band sub as I think that these are the most important and would work best with this inexpensive driver. Why two? Because I doubt that one could take the output of the plate amp - I know what they claim the power handling is, but I also know what to expect from a 2" voice coil and a cheap driver.
Patrick Bateman said:
John,
Why not just slice the tube longitudinally, then glue the two pieces to your flat divider? Easy to control the whole seam this way. You could even taper your cut if you want a tapered tube. There shouldn't be much pressure in the tube, but if you are worried about integrity, it would be simple to wrap the entire reassembled tube with reinforced tape, or for a decorative touch, a nice braided rope.
Sheldon
gedlee said:
Hi Earl,
The tapped horn is most like a 6th order bandpass, and most analogous to a vented box vs. a transmission line comparison. I've been designing plenty of 4th and 6th order bandpasses lately, and there are many similarities and a few differences which can of course work for or against you. For the rest reading along, the important difference lies in the significant distance/lengths involved related to the design vs. the operating bandwidth, so you have to account for 1/4 wave resonances and other effects which again can be problematic or beneficial.
The comparison of a 4th order bandpass to a 6th order makes for a reasonably parallel comparison of a horn and a tapped horn. As with most higher order systems including more variables, they are also easier to screw up, requiring a little more fine tuning or accuracy and completeness in the modeled parameters.
gedlee said:
Earl,
Is your definition of "work well" implying there is useful gain in sensitivity? I've always seen bandpasses as offering the advantage of allowing the flexibility to trade bandwidth for sensitivity with suitable drivers. While sensitivity is traded off, bandwidth of one decade is certainly possible.
Interesting. They got rid of the price break at ten.
http://www.mcmelectronics.com/product/55-2421
I've had these around for years - used them in two seperate projects already.
I cloned the autotuba by figuring out the dimensions of it using this site:
http://home.columbus.rr.com/weswhitmore/autotuba/
Also made a couple for this project here:
http://www.audiogroupforum.com/csforum/showthread.php?t=62292&page=2
http://www.mcmelectronics.com/product/55-2421
I've had these around for years - used them in two seperate projects already.
I cloned the autotuba by figuring out the dimensions of it using this site:
http://home.columbus.rr.com/weswhitmore/autotuba/
Also made a couple for this project here:
http://www.audiogroupforum.com/csforum/showthread.php?t=62292&page=2
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.