More useful perhaps, but tests are usually done at a given voltage, then advanced by a specific increase in level, until power compression sets in, obvious mechanical distress noises are heard, or the magic smoke is released from the test cabinet.
Various cabinets then can be compared for the different metrics of SPL, distortion and phase.
SPL normalization for a listening A/B test requires EQ to be fair, otherwise a sub with flat LF response will sound “louder” than a upward trending sub at the same average SPL level.
Adding EQ in a listening test is a “no no” to some people, in general penalizing most more efficient subs which usually will have a rising upper response.
Art Welter
As I said earlier, the fact that it reduces distortion so much below cutoff is prima facie evidence that it is working. That means it's your turn to present evidence to the contrary. If you have some data that shows something else is what caused the reduction of distortion, have at it. But you don't, you're just being argumentative.
I didn't bother to look at the data and refresh myself with it. Please forgive me that I forgot what the specs were on this design I did six years ago. I only remembered that the distortion was very low single digits in the passband, and rose into the low double digits below cutoff. I didn't remember that it was 15%, not the 28% you quoted. At the time you made that statement, I took you at your word and hadn't really started going over this stuff with a fine tooth comb.
We've argued jots and tittles because that's the argument you wanted to make. You have done this on purpose, focusing solely on the 30Hz value because that's just about where your box crosses the xmax line and distortion is just in the double digits. Go down a few Hertz and it's high double digits; A few more Hertz and it's triple digits. That's something I don't need charts to know - Everyone with any measurement experience reading this thread knows that.
I understand being zealous about the point that bass-reflex boxes have a place, that a huge basshorn isn't right for everyone. In my experience, I find that a few smaller direct radiating subs spread around the room is far better than one large hornsub. That's especially true if the owner would use EQ to boost the bass below horn cutoff - a really bad idea, in my opinion. I can also understand your application, a "club sub" that's easy to carry.
But honestly, Art, your continued persistence is just plain wrong. This is the kind of badgering that prompted me to suggest the $1000 wager on the cooling plugs. Anyone that would have taken that bet would have been stupid, but if you're going to persist, I think you should be forced to put your money where your mouth is. That's what I thought then, and it's what I think now.
You have assumed, like so many others do, that just putting the magnets out in free air will help cool them. But you have done nothing to test that. I have. When I actually did the work, and performed the tests, I was surprised to find that the temperature of ambient air didn't do squat. I mean, nothing, nada. In the range they're used in, a slight heating or cooling of the air doesn't help or hurt. Whether the air surrounding the magnet is 50° or 90°, the difference in the local ambient is too small to make much difference. The venting works pretty well in this range, and +/- 20° is basically the same thing.
The biggest improvement can be made by getting the heat out of the motor core. That's what I found, and you can take advantage of that or you can ignore it.
Same thing with the push-pull drive. It reduces distortion by a measurable amount, a remarkable amount, actually, and yet you keep trying to dismiss it. You're like a flat-earther that stubbornly argues the earth is flat.
The data is all there. Push-pull drive works extremely well. Take it, use it to your advantage. Or ignore it.
Last edited:
Wayne,
You seem to again be confusing me with someone else.
I have spent enough time in the air as a passenger and a pilot to know the earth is round.
I have not ignored your data, despite you disparaging mine, nor did I dismiss the effect of cooling plugs, horn loading, or push pull drive.
In fact, I have re-posted links to your data in this thread, and others regarding your paper on passive crossover design.
I also have never “assumed” putting Lab 12 magnets “out in free air” is superior to your cooling plug, though it is obvious that sealing them in a small chamber is worse than free air- space heaters work better in a room than outdoors :^).
I also am not zealous over my ported design, or I would not have replaced it for my use, about 80% outdoor shows.
I am not doubting that push pull may offer an advantage, but you still offer no comparative tests between
normal loading and push pull in any of your designs.
Wish you would offer some data backing up your claim that it halves distortion instead of your repeated ranting .
I’ll repeat what I wrote in post #79:
If the distortion figures below Fc on the LABhorn you previously posted are correct (and the cabinet was properly built and had no air leaks, etc.) then it is obvious that the 12Pi distortion is much lower below Fc.
The compression chamber to front chamber relationship is different in the 12Pi compared to the LABhorn, that may actually be the primary contributor to the lowering of distortion.
Although I have no doubt that push pull may reduce distortion, without trying your design “normal” compared to push pull, there is no data to tell what percentage the push pull contributes to the reduction.
Although I would not put much money on a bet, probably not enough to even cover your time for the experiment if I lost, my bet would be the difference in distortion below 30 Hz with the 12Pi “normal” (as opposed to push pull), but with speaker magnets both towards the horn throat chamber, would result in less than a 10% difference in distortion below Fc, and 2% above.
If you want to stop my “badgering”, simply provide the data I have asked for repeatedly.
If you didn’t actually do push pull to normal loading comparative tests, that’s OK, but just admit it already, stop blowing smoke.
Art Welter
Yes, I have. The LABhorn measurements at 40 watts show what that horn does in terms of distortion. My distortion meaurements are an order of magnitude better. This is strong evidence that push-pull drive is extremely effective.
Exactly, I know. That's the point.
Front chamber size plays a small part, but it's mostly at high frequencies because it is the acoustic low-pass filter it creates that makes the difference. It reduces high-frequency harmonics, mostly third and higher.
We do see distortion fall as frequency rises. The natural reduction of excursion as frequency rises and the acoustic low-pass filter of the front chamber both help reduce distortion up high.
But really, the falling distortion towards the top of the passband is probably more due to the reduced excursion than the front chamber. The fact that all woofers have reduced distortion as frequency goes up (and excursion goes down) tends to support this as being more of an influence than the front chamber.
Regardless of what happens at the high end of the passband, the front chamber isn't doing squat to reduce distortion at the very low end and below cutoff.
In fact, I built three separate versions of the 12Pi hornsub, each with different front and rear chamber sizes. The changes affected response ripple, but not distortion. In all three versions, distortion was an order of magnitude lower than the LABhorn. This is strong evidence that the front chamber size plays only a minor role in the distortion components, at least at lower frequencies.
One really shouldn't need that evidence as it is obvious that the front chamber size can't affect harmonics falling in the passband, where they would be when the horn is driven under cutoff. But the data is there, nonetheless.
This rules out the cause of the order-of-magnitude lower distortion as being caused by front and rear chamber differences. That leaves push-pull drive as the only possible explanation. Very telling evidence that push-pull drive is effective.
Last edited:
If what you wrote is correct, we should also clearly see an order of magnitude reduction of distortion with a ported cabinet operating push pull rather than normal.
I’ll test a dual lab 12 ported cabinet both ways and post the results later today.
Art Welter
Art, don't forget to use DJK's plenum style of symmetric line-up of the drivers to get the biggest effect of the PP concept.
That is exactly what I won't do.
A plenum reduces distortion in normal facing speakers too.
Most of the proponents of PP also use a plenum, which as an acoustical LP filter attenuates HF, reducing distortion, as distortion is upper harmonics.
Makes it hard to determine which does more, the plenum or PP.
Adding a plenum would require me to do four tests, (and build something I have no use for, with wood I don't have) and since I only have equipment to do that with discrete tones, testing takes a fair amount of time.
I'll keep my test simple, first testing my stock cabinet, then reversing one cone, (reversing the polarity on that cone) and test again.
It is 9:25 AM now, I'm heading to the shop now to get started with the tests.
Art
Good point there Art, but if you want to copy Waynes style you need two cabs with the drivers arranged symetric. That way you can make sure the plenum is 7.6" and opened from 4 sides. That way the filter effect of the plenum is negligible.
You could also have 3 side plenum opening in case you want the drivers to have the same back loading.
You could also have 3 side plenum opening in case you want the drivers to have the same back loading.
Last edited:
A little more elaboration is warranted, I think. I realize that some people understand these principles, but many DIYers are still learning, and "sound bite" arguments can easily sway them in the wrong direction.
Two features that are sometimes argued about are the push-pull drive and the cooling plugs. Unbiased people don't generally argue, they just read and make up their own minds. But people with agendas sometimes do. The argumentative person is usually someone that makes a subwoofer, mostly commercial manufacturers and some DIYers, all wanting to have the most successful, most popular subwoofer system. So to them, it is really important to win the "popular opinion" vote. If they aren't running push-pull drive and cooling plugs, they will probably attack those features or in some other way try to minimize the system in the reader's minds. So I wanted to write a few more words, to put some of the most common arguments in their proper light.
The first one, the push-pull drive:
Before the 12Pi prototype was built and measurements posted, naysayers simply said it wouldn't be particularly effective. But once they saw the measurements, showing the 12Pi to have order of magnitude lower distortion, the arguments changed to "Sure, I see the lower distortion, but that doesn't necessarily prove the push-pull drive is what did it."
This simply is not true. The LABhorn provides the baseline to compare against. As I mentioned in the last post, the front/rear chamber differences cannot do this. The front chamber provides low-pass, but not enough to reduce distortion anywhere close to what is measured. As for the rear chamber, the LABhorn actually optimizes the chamber dimensions to reduce excursion at the expense of response. Making the rear chamber larger actually increases distortion, if done alone. That's one of the things the naysayers pointed to before the measurements proved them wrong.
Further, I changed front/rear chamber sizes in each of three versions, and there was no significant difference in distortion. The differences are mostly in the response curve.
What changes distortion in push-pull drive is the symmetry of the load. If you make one side different than the other, then distortion reduction isn't as good. I've measured this too. The prototype couldn't be built with as exact tolerances as the production models, and the chamber sizes weren't exactly equal. Distortion was still way lower than the LABhorn, but production models were lower still. This is because the precision of the chamber size matching was better.
That's also why you can't just take one woofer and turn it around for a comparison. The magnet displaces air volume on one side, making the chamber much smaller. You have to build chambers that are physically smaller on the woofer face side to maintain equal volumes between the two.
But none of that matters. The LABhorn distortion curve provides a good baseline to test against. The 12Pi basshorn is an order of magnitude lower than the LABhorn, especially at the lowest frequencies, which is prima facie evidence that push-pull drive works. I think I've said that some four or five times now, but it is worth repeating, especially in light of the repeated "comparative tests" drivel.
That's right. Of course, you have to build it right for best results, as Djim said.
There was a guy on PSW that did exactly that, a vented cabinet with dual LAB12 woofers configured push-pull. The amount of distortion reduction was clearly evident in his measurements, as is to be expected.
Now onto the cooling plugs:
There is a common misconception that forced air convection cooling is all that works. I think this is probably due to the successful implementation of cooling vents for woofers, starting in the 1970s or so. Tends to give people a "seat of the pants" view that narrows their focus to an "air cooling" mindset.
I also was subject to this view when I first started developing the cooling system for the 12Pi hornsub. I could see that the failure mode was almost always thermal, which started an early argument that the first limit was excursion. Some arguments were fierce, early on, but I believe all have finally come to realize this isn't true. The LAB12 driver is actually very difficult to damage from over excursion, because of the xmax/xmech relationship, and because of its rarely low thermal limits, especially in basshorns. It can be damaged from overexcurison, but actual failures from this mode are rare. But I digress, the point is, I could see the failure mode was usually thermal, and I set about trying to vent the air out of the rear chamber, to keep the rear chamber cool.
What I learned is this almost totally didn't matter. A sealed rear chamber does heat up, but not nearly as much as the core temperature, which is the real problem. Skyrocketing core temperatures will melt the voice coil adhesive, whether or not the rear chamber air temperature is 100° or 50°. That's a secondary issue.
That's why "sound byte" arguments like this are so annoying:
It may seem "obvious" but it is completely false, illustrated with an analogy that appeals to common sense, but that actually is a false illusion. What's important in the space heater is the room temperature, but what's important in our case is the voice coil temperature. The analog would be accurate only if we were talking about the heating element. Now tell me, Mr. Welter, do you really think the heating element is cooled by any significant amount when the room is cool? Or does it glow just as red hot whether in a 50° room as it does in a 100° room?
These are some of the kinds of arguments I've faced, over and over and over again. To tell the truth, it gets real old. The data is out there, look at it. Take advantage of it. Use it in your own designs. Or you can ignore it too, if you want. But don't use flat-earther arguments anymore. They're wrong and they're annoying.
You know, there was a time when I argued the other side, but in a slightly different configuration and for a different reason. I've mentioned this a few times in this thread, but it is worth mentioning again because it goes back to the "sound bite" thing, so commonly used on messageboards.
When I first encountered the LABhorn, it was in the context of using it for home theater. The suggestion was made that boosting the deepest bass (below cutoff) would make it awesome in that environment. It rolls off at 30Hz, but to get down to the sub-20Hz range, EQ could be employed. Some people that are pretty visible in the industry promoted this idea, saying "even below cutoff, it still provides some horn loading."
I argued that was a bad idea, because a LABhorn at 20Hz is just a direct radiator in an undersized sealed box with a distortion amplifier stuck on the front of it. The only "horn loading" a LABhorn provides at 20Hz is to horn load the distortion. Of course, that was heresy, and I guess some people would think of me as being a naysayer. But I think there's a big difference between arguing minutia and arguing a legitimate point.
I guess that stigma still remains in some people's memories. I did "bitch" about the LABhorn, particularly when used for home theater. I mean, I understand the spectacle of having huge basshorns in the house. But beyond that, just talking about benefits and drawbacks, I would prefer using direct radiating subs, tuned for the intended passband, low enough they weren't massive distortion generators.
The "other side" argued that the horn loading reduced drive requirements to very low power levels, making the distortion insignificant. Through the passband, this is most certainly true. But below the cutoff frequency, it most certainly is not. When presented a 20Hz signal, pretty much the only thing coming out of the horn is distortion, well over 100%.
I suggested that it would be better to have four direct radiating subs, say LAB12 woofers in 3ft3 boxes tuned to 20Hz or so. They would produce adequate output, and could be distributed throughout the room to mitigate room modes.
Comparing that with LABhorns, the vented boxes would need roughly 100 watts where the hornsubs would only need 10 watts (through the passband). While the horn distortion would be lower in the passband, the 10dB boost required to bring up the ~20Hz range would put the stop band levels at 100 watts. The distortion created from that is well over 100%. All the LABhorn is doing with a 20Hz signal is to make (fairly loud) 40Hz and 60Hz harmonics. I suppose you could call that "some horn loading" but it's not what I would consider to be a good thing. So I guess some might call that "bitching".
To me, this is a completely different argument than the one we've had here in the last few pages. Ironically though, it is similar, in that I was arguing for a vented box. But what I was arguing for was using multiple relatively small subwoofers, tuned appropriately for the passband. In the last few pages, we've been discussing a vented box tuned to cover the exact same passband, from 30Hz up. That's a different matter entirely. The vented box has one advantage, which is its size. In all other respects, it cannot compete.
Two features that are sometimes argued about are the push-pull drive and the cooling plugs. Unbiased people don't generally argue, they just read and make up their own minds. But people with agendas sometimes do. The argumentative person is usually someone that makes a subwoofer, mostly commercial manufacturers and some DIYers, all wanting to have the most successful, most popular subwoofer system. So to them, it is really important to win the "popular opinion" vote. If they aren't running push-pull drive and cooling plugs, they will probably attack those features or in some other way try to minimize the system in the reader's minds. So I wanted to write a few more words, to put some of the most common arguments in their proper light.
The first one, the push-pull drive:
Before the 12Pi prototype was built and measurements posted, naysayers simply said it wouldn't be particularly effective. But once they saw the measurements, showing the 12Pi to have order of magnitude lower distortion, the arguments changed to "Sure, I see the lower distortion, but that doesn't necessarily prove the push-pull drive is what did it."
This simply is not true. The LABhorn provides the baseline to compare against. As I mentioned in the last post, the front/rear chamber differences cannot do this. The front chamber provides low-pass, but not enough to reduce distortion anywhere close to what is measured. As for the rear chamber, the LABhorn actually optimizes the chamber dimensions to reduce excursion at the expense of response. Making the rear chamber larger actually increases distortion, if done alone. That's one of the things the naysayers pointed to before the measurements proved them wrong.
Further, I changed front/rear chamber sizes in each of three versions, and there was no significant difference in distortion. The differences are mostly in the response curve.
What changes distortion in push-pull drive is the symmetry of the load. If you make one side different than the other, then distortion reduction isn't as good. I've measured this too. The prototype couldn't be built with as exact tolerances as the production models, and the chamber sizes weren't exactly equal. Distortion was still way lower than the LABhorn, but production models were lower still. This is because the precision of the chamber size matching was better.
That's also why you can't just take one woofer and turn it around for a comparison. The magnet displaces air volume on one side, making the chamber much smaller. You have to build chambers that are physically smaller on the woofer face side to maintain equal volumes between the two.
But none of that matters. The LABhorn distortion curve provides a good baseline to test against. The 12Pi basshorn is an order of magnitude lower than the LABhorn, especially at the lowest frequencies, which is prima facie evidence that push-pull drive works. I think I've said that some four or five times now, but it is worth repeating, especially in light of the repeated "comparative tests" drivel.
That's right. Of course, you have to build it right for best results, as Djim said.
There was a guy on PSW that did exactly that, a vented cabinet with dual LAB12 woofers configured push-pull. The amount of distortion reduction was clearly evident in his measurements, as is to be expected.
Now onto the cooling plugs:
There is a common misconception that forced air convection cooling is all that works. I think this is probably due to the successful implementation of cooling vents for woofers, starting in the 1970s or so. Tends to give people a "seat of the pants" view that narrows their focus to an "air cooling" mindset.
I also was subject to this view when I first started developing the cooling system for the 12Pi hornsub. I could see that the failure mode was almost always thermal, which started an early argument that the first limit was excursion. Some arguments were fierce, early on, but I believe all have finally come to realize this isn't true. The LAB12 driver is actually very difficult to damage from over excursion, because of the xmax/xmech relationship, and because of its rarely low thermal limits, especially in basshorns. It can be damaged from overexcurison, but actual failures from this mode are rare. But I digress, the point is, I could see the failure mode was usually thermal, and I set about trying to vent the air out of the rear chamber, to keep the rear chamber cool.
What I learned is this almost totally didn't matter. A sealed rear chamber does heat up, but not nearly as much as the core temperature, which is the real problem. Skyrocketing core temperatures will melt the voice coil adhesive, whether or not the rear chamber air temperature is 100° or 50°. That's a secondary issue.
That's why "sound byte" arguments like this are so annoying:
It may seem "obvious" but it is completely false, illustrated with an analogy that appeals to common sense, but that actually is a false illusion. What's important in the space heater is the room temperature, but what's important in our case is the voice coil temperature. The analog would be accurate only if we were talking about the heating element. Now tell me, Mr. Welter, do you really think the heating element is cooled by any significant amount when the room is cool? Or does it glow just as red hot whether in a 50° room as it does in a 100° room?
These are some of the kinds of arguments I've faced, over and over and over again. To tell the truth, it gets real old. The data is out there, look at it. Take advantage of it. Use it in your own designs. Or you can ignore it too, if you want. But don't use flat-earther arguments anymore. They're wrong and they're annoying.
You know, there was a time when I argued the other side, but in a slightly different configuration and for a different reason. I've mentioned this a few times in this thread, but it is worth mentioning again because it goes back to the "sound bite" thing, so commonly used on messageboards.
When I first encountered the LABhorn, it was in the context of using it for home theater. The suggestion was made that boosting the deepest bass (below cutoff) would make it awesome in that environment. It rolls off at 30Hz, but to get down to the sub-20Hz range, EQ could be employed. Some people that are pretty visible in the industry promoted this idea, saying "even below cutoff, it still provides some horn loading."
I argued that was a bad idea, because a LABhorn at 20Hz is just a direct radiator in an undersized sealed box with a distortion amplifier stuck on the front of it. The only "horn loading" a LABhorn provides at 20Hz is to horn load the distortion. Of course, that was heresy, and I guess some people would think of me as being a naysayer. But I think there's a big difference between arguing minutia and arguing a legitimate point.
I guess that stigma still remains in some people's memories. I did "bitch" about the LABhorn, particularly when used for home theater. I mean, I understand the spectacle of having huge basshorns in the house. But beyond that, just talking about benefits and drawbacks, I would prefer using direct radiating subs, tuned for the intended passband, low enough they weren't massive distortion generators.
The "other side" argued that the horn loading reduced drive requirements to very low power levels, making the distortion insignificant. Through the passband, this is most certainly true. But below the cutoff frequency, it most certainly is not. When presented a 20Hz signal, pretty much the only thing coming out of the horn is distortion, well over 100%.
I suggested that it would be better to have four direct radiating subs, say LAB12 woofers in 3ft3 boxes tuned to 20Hz or so. They would produce adequate output, and could be distributed throughout the room to mitigate room modes.
Comparing that with LABhorns, the vented boxes would need roughly 100 watts where the hornsubs would only need 10 watts (through the passband). While the horn distortion would be lower in the passband, the 10dB boost required to bring up the ~20Hz range would put the stop band levels at 100 watts. The distortion created from that is well over 100%. All the LABhorn is doing with a 20Hz signal is to make (fairly loud) 40Hz and 60Hz harmonics. I suppose you could call that "some horn loading" but it's not what I would consider to be a good thing. So I guess some might call that "bitching".
To me, this is a completely different argument than the one we've had here in the last few pages. Ironically though, it is similar, in that I was arguing for a vented box. But what I was arguing for was using multiple relatively small subwoofers, tuned appropriately for the passband. In the last few pages, we've been discussing a vented box tuned to cover the exact same passband, from 30Hz up. That's a different matter entirely. The vented box has one advantage, which is its size. In all other respects, it cannot compete.
Last edited:
Since there is so much hair splitting going on, isn't the 12pi 26" wide versus 22.5" for the Labhorn? That would make it a >13% larger subwoofer and provide a small performance advantage as well. Larger horns covering the same operating bandwidth usually have smoother response and slightly higher sensitivity.
Art I look forward to seeing your results. I may do a quick push pull versus standard mounting distortion test myself. I have a common chamber, dual opposed, sealed enclosure that I could use. Since it is sealed and a common chamber the air volume seen will be the same and it will be a simple test to do. I won't get to it for a couple of weeks though.
Wayne, it is generally held that push pull works for the even harmonics but not the odd. Do you have any public data for the harmonic make-up of the 12pi distortion? Is it predominantly 3rd? I would expect the odd order harmonics to still be present in roughly the same level as with standard driver orientation, or does push pull also work to reduce the odd harmonics? The subs that I have tested extensively with push pull driver mounting are Danley DTS10's. They exhibit the usual skyrocket below cutoff of any other non sealed subwoofer. Obviously the 12pi has a sealed chamber on the drivers which keeps excursion under control below the horn loading where the tapped horn behaves more like a vented enclosure below loading but the distortion shown in the 12pi test below loading still seems remarkable given the 15-20 db of harmonic amplification the horn provides. The 3rd harmonic of 15-20hz puts it right in the horns wheelhouse. A harmonic -40 db down representing 1% distortion being amplified by 15db to -25db from the fundamental would raise the distortion to 5.62%.
Art I look forward to seeing your results. I may do a quick push pull versus standard mounting distortion test myself. I have a common chamber, dual opposed, sealed enclosure that I could use. Since it is sealed and a common chamber the air volume seen will be the same and it will be a simple test to do. I won't get to it for a couple of weeks though.
Wayne, it is generally held that push pull works for the even harmonics but not the odd. Do you have any public data for the harmonic make-up of the 12pi distortion? Is it predominantly 3rd? I would expect the odd order harmonics to still be present in roughly the same level as with standard driver orientation, or does push pull also work to reduce the odd harmonics? The subs that I have tested extensively with push pull driver mounting are Danley DTS10's. They exhibit the usual skyrocket below cutoff of any other non sealed subwoofer. Obviously the 12pi has a sealed chamber on the drivers which keeps excursion under control below the horn loading where the tapped horn behaves more like a vented enclosure below loading but the distortion shown in the 12pi test below loading still seems remarkable given the 15-20 db of harmonic amplification the horn provides. The 3rd harmonic of 15-20hz puts it right in the horns wheelhouse. A harmonic -40 db down representing 1% distortion being amplified by 15db to -25db from the fundamental would raise the distortion to 5.62%.
That's right, actually, it's 45"x45"x28", which is 5-1/2" wider than the LABhorn. One of my goals was to smooth the response curve, especially its large dip right above cutoff. So I used a slightly larger mouth and different front and rear chamber sizes to make response smoother. I wanted a horn that could be used in fewer numbers and still get good response. It still fits through all doors (even doors in homes are 30" wide), and is a nice truck pack dimension.
Right again, push-pull only cancels even harmonics. The odd harmonics are still there, but you're right, the sealed rear chamber helps excursion not rise as much. Still does rise though, and quite a bit. My measurements are all THD+N, mostly because of time constraints, but I would assume as you have, that the distortion that's there is predominantly third harmonics. Higher harmonics are getting shifted far enough out of the passband to be attenuated by the bandpass nature of the horn.
As an aside, I'm not sure if you knew this but Danley was initially skeptical, but as you've noted, he has now has begun using push-pull drive himself.
Last edited:
I tried to edit and add some info, but I was just past the 30 minute limit. It occurred to me that some additional comments might be helpful:
It is remarkable, thanks. There is a large and obvious difference in the amount of distortion produced by the 12Pi hornsub and all the other subs we measured at each of the Prosound Shootouts.
But even given those large differences, it isn't really unexpected. The 800 watt THD level of the 12Pi hornsub at cutoff is about 15%, which is undoubtedly mostly the third harmonic. The rear chamber is sealed, so excursion doesn't rise nearly as much as a BR or TL/TH box below cutoff. Excursion still rises though, which definitely causes the bottom end distortion to start coming up. One would expect the third harmonic of the woofer at ~30Hz in a 20 liter sealed chamber to be somewhere in the vicinity of -25dB to -30dB, so with the horn acting upon this level, it would probably rise to -20dB to -15dB, which is exactly what we're seeing.
Above 35Hz or so, excursion is reduced by horn loading, and we also have push-pull drive to further reduce even harmonics. An octave above that, excursion is even lower, and we start seeing distortion fall even more. Towards the top of the passband, distortion is lowest, because we have several things in play: Horn loading, push-pull drive, the acoustic low-pass of the front chamber and horn folds and the natural tendency of excursion to be reduced as frequency rises.
Compared with other systems, we see some of the same issues, but some are different. The excursion of a vented box or a transmission line remains low through the passband, but rises dramatically below cutoff. Systems like that lose control fast. Tapped horns have an additional difficulty, in that they tend to amplify the harmonics below cutoff. They also seem to have a lot of artifacts just above the band of interest - still in the "passband", but unwanted. I think that probably contributes to some additional distortion, if you view them with THD+N rather than a series of H2, H3, H4 harmonics. I haven't studied them in great detail, but I am not entirely sure if those high frequency artifacts are harmonically related. If not, they would show up on a THD+N measurement, but maybe not on a chart of harmonic series.
In any case, push-pull drive reduces even harmonics. Some may argue that even-harmonics are more euphonic than odd-harmonics, and that may be true. It's another discussion. But I think if I can reduce harmonics, it's probably a good thing. With two drivers, it seems worthwhile. In my case, in the 12Pi hornsub, the difference was remarkable, especially at low frequencies. It is kind of weird to hear the thing go dead quiet down low, when you're used to hearing sweep after sweep of other horns that don't do that.
It is remarkable, thanks. There is a large and obvious difference in the amount of distortion produced by the 12Pi hornsub and all the other subs we measured at each of the Prosound Shootouts.
But even given those large differences, it isn't really unexpected. The 800 watt THD level of the 12Pi hornsub at cutoff is about 15%, which is undoubtedly mostly the third harmonic. The rear chamber is sealed, so excursion doesn't rise nearly as much as a BR or TL/TH box below cutoff. Excursion still rises though, which definitely causes the bottom end distortion to start coming up. One would expect the third harmonic of the woofer at ~30Hz in a 20 liter sealed chamber to be somewhere in the vicinity of -25dB to -30dB, so with the horn acting upon this level, it would probably rise to -20dB to -15dB, which is exactly what we're seeing.
Above 35Hz or so, excursion is reduced by horn loading, and we also have push-pull drive to further reduce even harmonics. An octave above that, excursion is even lower, and we start seeing distortion fall even more. Towards the top of the passband, distortion is lowest, because we have several things in play: Horn loading, push-pull drive, the acoustic low-pass of the front chamber and horn folds and the natural tendency of excursion to be reduced as frequency rises.
Compared with other systems, we see some of the same issues, but some are different. The excursion of a vented box or a transmission line remains low through the passband, but rises dramatically below cutoff. Systems like that lose control fast. Tapped horns have an additional difficulty, in that they tend to amplify the harmonics below cutoff. They also seem to have a lot of artifacts just above the band of interest - still in the "passband", but unwanted. I think that probably contributes to some additional distortion, if you view them with THD+N rather than a series of H2, H3, H4 harmonics. I haven't studied them in great detail, but I am not entirely sure if those high frequency artifacts are harmonically related. If not, they would show up on a THD+N measurement, but maybe not on a chart of harmonic series.
In any case, push-pull drive reduces even harmonics. Some may argue that even-harmonics are more euphonic than odd-harmonics, and that may be true. It's another discussion. But I think if I can reduce harmonics, it's probably a good thing. With two drivers, it seems worthwhile. In my case, in the 12Pi hornsub, the difference was remarkable, especially at low frequencies. It is kind of weird to hear the thing go dead quiet down low, when you're used to hearing sweep after sweep of other horns that don't do that.
I would REALLY appreciate it, if someone could re phrase the details going on here, about 2nd third harmonics, distortion values, and the rest of it, etc,
into a language i could understand and appreciate easily.
why do these small differences matter so much?
put a pair of 12s or 15s into a well designed (and built) box and we get great bass.
why so much discussion on the 5% here and there??
into a language i could understand and appreciate easily.
why do these small differences matter so much?
put a pair of 12s or 15s into a well designed (and built) box and we get great bass.
why so much discussion on the 5% here and there??
Because once you have heard, used and built enough cabs, figuring out how to get that last little 7% performance improvement out of a design is what is most interesting.
yes, agreed, and is a discussion I'm interested in.
As someone brought up in the 80's on stacks of double 18's... my ear wants to hear at least a little distortion. (the key word here is 'little') A PA sound completely free of distortion is in someway 'lacking' to many ears. However this gets us back to the whole tube vs transistor amplifier argument. Tubes had even order harmonics that transistors lacked. All that was left in a transistor amp was the odd order harmonics. This to many left an amplifier sound that was lacking (the 2nd order harmonic) and harsh. (3rd was still there) However as ears and minds changed, the sound of a transistor amplifier became more 'normal' and the word harsh was not used as much.
I think the same is true of the push pull drive in the 12pi. As your ear gets used to the sound, it becomes what you want to hear, and anything else stands out in stark contrast. (regardless of 1% more or less, here or there measurements) It's partly why I love the sound of tapped horns, I get just a little distortion that I don't get in a front loaded horn cabinet, while not getting bombarded with distortion like an over driven double 18 cabinet. It's my 'happy medium' that gets me efficient spl and the sound I like. (and yes, a TH like anything else when driven nuts will distort like crazy, as my over 100% distortion smaart graph shows)
I don't want to speak for either art or wayne -- but I'm pretty sure they both have their 'normal' and that those normals are in no way the same. And that is at least part of the debate.
The results of my testing of Lab 12 in ported and sealed cabinets , normal front load and push pull are interesting. Had a perfect day for testing,(other than smoke from all the nearby fires) noise was very low, never had to wait for wind to die as in most testing here.
Although 1st, 3rd, 5th (odd order) harmonics were greatly reduced by push pull loading, the second harmonic sometimes increased, resulting in the THD not signifigantly changing, as can be seen in the 20 Hz 49 volt test results below.
It appears the 12Pi has similar THD approaching 100% around 20 Hz to the PP or standard loaded Lab 12 in a sealed enclosure, so I’d expect the harmonic structure to be similar.
Since the frequency response and output of both the PP and standard cabinets were almost identical, an A/B listening test was simple, just swap the speaker cord and short out the unused cabinet.
At lower power levels, where the Lab12 is quite clean, no difference could be detected. When pushed at a power level where distortion could be noticed, the most distracting artifact was the audible vent noise from the reversed cone of the PP.
The lack of the second harmonic, an octave, made the now predominant third harmonic distortion more apparent in the PP cabinet. This is music related thing, the third harmonic , being a perfect fifth, may sound OK with some compositions, while sounding “wrong” with others.
Although the technical side of me says reduction of distortion of any sort is a good thing as far as a reproduction system is concerned, both my old ears, and a college freshman apprentice thought the push pull cabinet sounded less “musical” when pushed hard.
And Wayne, please don't take what I have written here out of context, at levels equivalent to your tests, and at what any reasonable user would use the 12Pi at in it's passband, distortion simply would not be audible.
Art Welter
Attachments
not for me,
for me the music itself is the most interesting, if i can get 97% of the performance from a simple box, and not have to build a very complicated box.
im building a simple box and enjoying the music.
You can clearly see the reduction of the even harmonics in Art's measurements. The fundamental is 20Hz, and the even harmonics are 40Hz, 80Hz, and so on. I can see a 30dB reduction in the second harmonic, for example. That's what push-pull does for you.
What's nice about a hornsub with push-pull drive is the natural bandpass nature of the system tends to reduce the third harmonic and higher. For example, the third harmonic of 30Hz is 90Hz, which is still in the passband, but by the time you get to 50Hz, the third harmonic is moving outside the band. So the third harmonic is attenuated for about half the range the 12Pi basshorn is used in. All the higher harmonics are too. Basically, when you deal with the second harmonic in a hornsub, you've got most of the distortion beat.
About the subjective stuff, the audibility and "euphonics", I don't get into the matter of what's audible and what's not, or what's euphonic and what's not. I do think that's an interesting topic though, and I definitely do not discount it. I am certain that the spectrum and distribution is as important if not more so than the single-metric numbers. But I haven't addressed that in any of my measurements. They're just raw measurements. I am not sure it is as easy to quantify what people "like" as it is to measure raw performance values. That's a subjective thing, and takes controlled statistical tests with groups of people.
I will say this, just my own subjective opinion. This is what I perceive.
1. In main speakers, lower distortion speakers tend to sound clearer to me, all other things being equal. However, the audibility is a function of SPL, and below a certain level, I cannot detect the distortion.
2. At high power level, a speaker with a little more distortion will be more fatiguing, making me want to "turn it down" much sooner than a less distorted speaker.
3. At low power levels, a speaker with a little more distortion will be more fatiguing over a long period of time. At low power levels, I cannot tell that it is going to fatigue me at first, even for a few hours, but after a few days at a trade show, for example, the higher distortion speaker will fatigue me, where the lower distortion speaker does not.
4. In subwoofers, distortion is much less noticeable, but higher distortion tends to sound louder and fatter. (Sorry for the subjective terms, just trying to describe what I hear)
5. Just like the mains, a high distortion woofer will be fatiguing, even if it cannot be easily detected.
I gained these opinions after many years of using my own speakers which often come in a stock or upgraded version. The upgraded version usually has a midwoofer with a shorting ring, creating less distortion. Their response curves are very similar between the stock and upgraded versions, but the upgraded model has lower distortion.
My approach has been to reduce distortion where I can.
What's nice about a hornsub with push-pull drive is the natural bandpass nature of the system tends to reduce the third harmonic and higher. For example, the third harmonic of 30Hz is 90Hz, which is still in the passband, but by the time you get to 50Hz, the third harmonic is moving outside the band. So the third harmonic is attenuated for about half the range the 12Pi basshorn is used in. All the higher harmonics are too. Basically, when you deal with the second harmonic in a hornsub, you've got most of the distortion beat.
About the subjective stuff, the audibility and "euphonics", I don't get into the matter of what's audible and what's not, or what's euphonic and what's not. I do think that's an interesting topic though, and I definitely do not discount it. I am certain that the spectrum and distribution is as important if not more so than the single-metric numbers. But I haven't addressed that in any of my measurements. They're just raw measurements. I am not sure it is as easy to quantify what people "like" as it is to measure raw performance values. That's a subjective thing, and takes controlled statistical tests with groups of people.
I will say this, just my own subjective opinion. This is what I perceive.
1. In main speakers, lower distortion speakers tend to sound clearer to me, all other things being equal. However, the audibility is a function of SPL, and below a certain level, I cannot detect the distortion.
2. At high power level, a speaker with a little more distortion will be more fatiguing, making me want to "turn it down" much sooner than a less distorted speaker.
3. At low power levels, a speaker with a little more distortion will be more fatiguing over a long period of time. At low power levels, I cannot tell that it is going to fatigue me at first, even for a few hours, but after a few days at a trade show, for example, the higher distortion speaker will fatigue me, where the lower distortion speaker does not.
4. In subwoofers, distortion is much less noticeable, but higher distortion tends to sound louder and fatter. (Sorry for the subjective terms, just trying to describe what I hear)
5. Just like the mains, a high distortion woofer will be fatiguing, even if it cannot be easily detected.
I gained these opinions after many years of using my own speakers which often come in a stock or upgraded version. The upgraded version usually has a midwoofer with a shorting ring, creating less distortion. Their response curves are very similar between the stock and upgraded versions, but the upgraded model has lower distortion.
My approach has been to reduce distortion where I can.
Last edited:
I mentioned earlier similar findings as Art describes LINK But the knowledge of theoretical 'correct' can influence your hearing if you are not realising.
I did a interesting test with some of the sound techs and a few colleagues. We are testing push-pull TH’s over the last year and in a first ‘blind’ tests without showing anyone THD graphs, 8 people choose the cab with higher distortion.
I repeated the same test one week later. I told them I changed things (which I didn’t) and showed them the THD plots first before the test. Then half of them (those who have more theoretical knowledge) 'suddenly' had changed their minds and stated the lower THD sounded better...
For Jbell, that was the flute theory that seems missing in PPTH 😉
Last edited:
- Status
- Not open for further replies.