Greetings,
Not long ago I bought 4 small woofers (Vifa P17WJ-00-04) with the intention of building a D’Appolito design. While researching the design, I learned 3 things which persuaded me to build a conventional single-woofer design. The problem is I am still not confident that any of these issues are true and relevant. I would like to explain these issues to our audio DIY community to see if anyone has experience with them to see if I was mislead, or saved from disaster.
My questions are: Does this mean that most D’Appolito designs are flawed in this manner? Does this explain why MTM designs were all the rage in the 80’s but are now less common? Is keeping all of the qs of unmasked (by the tweeter) nulls >45° unreasonable criteria?
Mind you, reader, that this writing is after-the-fact. My single-woofer speakers are finished and I love the sound. The other two woofers found a home in the back of my car. Maybe, some day another reader or I will build a real good D’Appolito design based on your answers.
Not long ago I bought 4 small woofers (Vifa P17WJ-00-04) with the intention of building a D’Appolito design. While researching the design, I learned 3 things which persuaded me to build a conventional single-woofer design. The problem is I am still not confident that any of these issues are true and relevant. I would like to explain these issues to our audio DIY community to see if anyone has experience with them to see if I was mislead, or saved from disaster.
- I came across an article at www.briotechnology.com by Mithat Konar titled “Vertically Symmetric Two-way Loudspeaker Arrays Reconsidered.” Mr. Konar wrote that the original article by Joe D’Appolito (I have not read, because I can not find a free copy) does not address the issue of lobbing between the two lower frequency drivers. To chop a 15-page paper down to a few sentences:
- The woofers share a wider range of frequencies than do the tweeter and the woofers, thus lobbing errors occurring between woofers could create bigger problems than a D’Appolito design solves.
- The closer together the woofers are, the fewer and wider the lobes will be at a given frequency. This makes it possible to design a speaker with a large “sweet spot” directed at the listener.
- Also, the lower the frequency, the fewer and wider the lobes will be for a given woofer distance.
- Sections b and c together, mean that the number and width of lobes are determined by the distance between the woofers and their crossover point to the tweeter.
- The author then concludes that “Specifically, if off-axis nulls are to be avoided, (3,2) systems must be designed such that the lowest-frequency null in the low frequency driver’s summed response is masked by the output of the high frequency driver for all vertical displacements of significance”. He then chooses less than 45° as “all vertical displacements of significance”.
My questions are: Does this mean that most D’Appolito designs are flawed in this manner? Does this explain why MTM designs were all the rage in the 80’s but are now less common? Is keeping all of the qs of unmasked (by the tweeter) nulls >45° unreasonable criteria?
- I explained my plan to a sales person at Madisound, and he assured me that one can not put two 4W drivers in series and expect them to behave like an 8W driver. He said that doubling the impedance will reduce the bass, and this is why no one does this. I argued that if I choose an 8W tweeter, it shouldn’t matter, but he was insistent that I was mistaken without being able to explain why. I’m still skeptical. Does this mean all MTM designs put 8W woofers/mids in parallel making a 4W speaker? What is the theoretical basis for his argument?
- I read several posts on this site in which people had put 2 woofers into a vented cabinet, and found that one woofer was flapping like crazy, while the other was hardly moving. I think they said the one closer to the port moved more. This told me that, unless I put the woofers in separate enclosures, or make my enclosure perfectly symmetrical I might have the same problem. Yet, I’m sure I’ve seen commercial designs with 2 or more woofers in a non symmetrical cabinet. I decided I didn’t want to deal with the issue, at least not on this project. Does anyone have experience with the 2-woofers, 1-port design? Are there any articles/resources?
Mind you, reader, that this writing is after-the-fact. My single-woofer speakers are finished and I love the sound. The other two woofers found a home in the back of my car. Maybe, some day another reader or I will build a real good D’Appolito design based on your answers.
Most of this has some basis in fact, but no-one goes into the full story.
I built my 1st MTM in 1975, oblivious of the theorectical pluses & minus, i just liked the symmetry. At the time i had never seen another design using this config. My current feelings about MTM are embodied in this loudspeaker:
XO is ~300 Hz 1st order (XO is still under development). The low XO means wave lengths long enuff that the drivers can be deemed coincident.
Lobing & phase issues are real. The effort to push down the XO point to minimize these issues really stresses a conventional tweeter and put an invariably high order complex XO point right where the ear is most sensitive.
Series vrs parallel has to be taken in context with the amp used. The Madisound guy is clearily thinking in a world where an amp should have as low an output impedance as possible (ie a voltage amp). Voltage amps prefer speakers in parallel, The possibility exists that if the drivers impedances differ significantly one of them could be a power hog. The opposite is true with a current amp. The other thing that occurs is that 2 coils in series double the inductance, in parallel it is halved. With a voltage amp the HF will roll off sooner if the drivers are in series... maybe good, maybe bad.
I'm not a big fan of conventional ported boxes, the minuses being greater than the pluses. If my box has a hole in it (and many do), it will either be a high resistance port or a transmission line (TL used in its broadest sense = quarter wave resonator)
I'd suggest getting onto Madisound and pre-order a pair of Alpair6 (i'm thinking the metal ones, which will be scarce for a while -- the mid-tweeters i used are getting very scarce). It should mate well efficiency wise with the VIFAs giving you the opportunity to build baffle step compensation in by XOing around the baffle-step frequency.
dave
I built my 1st MTM in 1975, oblivious of the theorectical pluses & minus, i just liked the symmetry. At the time i had never seen another design using this config. My current feelings about MTM are embodied in this loudspeaker:
XO is ~300 Hz 1st order (XO is still under development). The low XO means wave lengths long enuff that the drivers can be deemed coincident.
Lobing & phase issues are real. The effort to push down the XO point to minimize these issues really stresses a conventional tweeter and put an invariably high order complex XO point right where the ear is most sensitive.
Series vrs parallel has to be taken in context with the amp used. The Madisound guy is clearily thinking in a world where an amp should have as low an output impedance as possible (ie a voltage amp). Voltage amps prefer speakers in parallel, The possibility exists that if the drivers impedances differ significantly one of them could be a power hog. The opposite is true with a current amp. The other thing that occurs is that 2 coils in series double the inductance, in parallel it is halved. With a voltage amp the HF will roll off sooner if the drivers are in series... maybe good, maybe bad.
I'm not a big fan of conventional ported boxes, the minuses being greater than the pluses. If my box has a hole in it (and many do), it will either be a high resistance port or a transmission line (TL used in its broadest sense = quarter wave resonator)
I'd suggest getting onto Madisound and pre-order a pair of Alpair6 (i'm thinking the metal ones, which will be scarce for a while -- the mid-tweeters i used are getting very scarce). It should mate well efficiency wise with the VIFAs giving you the opportunity to build baffle step compensation in by XOing around the baffle-step frequency.
dave
Second issue first: whenever you combine two woofers you have to choose series or parallel connection. Series connection will be 4 times the Z of parallel and 6dB less sensitive, although the efficiency (power per watt) will be the same in both cases. You need to make a choice based on the impedance you can handle and the sensitivity required to match the tweeter.
With series connection each unit gets half the volts, but as long as the impedance curves of both units match it will be an even split. The frequency response will not be altered (no loss of bass).
There can be odd behaviour if your woofers aren't well behaved. For example, DC offset, the tendency of a woofer to shift in or out with high drive levels (at some frequencies) can be worse when two woofers share a cabinet volume. If you aren't experiencing any odd performance with your system, then don't worry about it.
I think the D'Appolito analysis is correct and typical driver spacing will generally cause nulling at and below crossover. Generally you can't get the tweeter to crossover low enough to prevent it. I made a series of systems at Snell (the XA series) that worked at creating uniform vertical directivity. The only way to improve performance was to create a special cluster of two mids and a tweeter on a common face plate, that let me get the center to center distances to a minimum. It still took lower than usual crossovers to reduce nulling.
D'Appolito type layouts will give symmetrical radiation, a plus, but 2 1/2 way systems work well too. That is, using two woofers but with one woofer crossed lower and only the second woofer working up to the tweeter's crossover point.
David S.
With series connection each unit gets half the volts, but as long as the impedance curves of both units match it will be an even split. The frequency response will not be altered (no loss of bass).
There can be odd behaviour if your woofers aren't well behaved. For example, DC offset, the tendency of a woofer to shift in or out with high drive levels (at some frequencies) can be worse when two woofers share a cabinet volume. If you aren't experiencing any odd performance with your system, then don't worry about it.
I think the D'Appolito analysis is correct and typical driver spacing will generally cause nulling at and below crossover. Generally you can't get the tweeter to crossover low enough to prevent it. I made a series of systems at Snell (the XA series) that worked at creating uniform vertical directivity. The only way to improve performance was to create a special cluster of two mids and a tweeter on a common face plate, that let me get the center to center distances to a minimum. It still took lower than usual crossovers to reduce nulling.
D'Appolito type layouts will give symmetrical radiation, a plus, but 2 1/2 way systems work well too. That is, using two woofers but with one woofer crossed lower and only the second woofer working up to the tweeter's crossover point.
David S.
Hello Byron,
I have had similar thoughts about a set of Vifa PL18W** ‘s, also the 4 ohm version. What was the magic XO frequency for the hypothetical MTM you were considering but could not find a matching tweeter?
My twist to this is that I have the 7 inch woofers built into a PHOENIX style open baffle. If successful these speakers will be used for quasi-near-field in a small room.
Does this speaker being OB exclude it from the conversation?
At this point it is only an experiment in OB, Bi-Amp and equalization!
DT
All Just for fun!
I have had similar thoughts about a set of Vifa PL18W** ‘s, also the 4 ohm version. What was the magic XO frequency for the hypothetical MTM you were considering but could not find a matching tweeter?
My twist to this is that I have the 7 inch woofers built into a PHOENIX style open baffle. If successful these speakers will be used for quasi-near-field in a small room.
Does this speaker being OB exclude it from the conversation?
At this point it is only an experiment in OB, Bi-Amp and equalization!
DT
All Just for fun!
And i see some possible advantages in doing so:
- radiation pattern mismatch between the woofers
and the tweeter can be mitigated (in the vertical
plane at least)
- depending on the shape of the baffle, lowpassing one
woofer can be integrated into the baffle step correction
- if needed recessing the tweeter for group delay compensation
may be more easy if the tweeter is not mounted between the two
woofers but ontop of the woofer which is crossed over to the tweeter.
- gaining some decimeters in height for the tweeter
may be advantageous for small floorstanders to
get the tweeter more close to ear height.
magic X-over freq for 6.5" woofers
"What was the magic XO frequency for the hypothetical MTM you were considering but could not find a matching tweeter?"
Most of the tweeters I found had ~100mm faceplates. That, with 170mm woofers means a highest recommended X-over frequency of just under 1000Hz.
"What was the magic XO frequency for the hypothetical MTM you were considering but could not find a matching tweeter?"
Most of the tweeters I found had ~100mm faceplates. That, with 170mm woofers means a highest recommended X-over frequency of just under 1000Hz.
Thank you
Thank's for all the responses.
It sounds like, theoretically, I could have wired the 4 ohm woofers in series and mated them to an 8 ohm tweeter but possible complications, and the inability to get the woofers close enough together, would have made this inadvisable for a novice builder with no test equipment. So it looks like I made the right call.
FYI. The final project is somewhere in the sticky thread "System Pictures & Descriptions" #1863
Thank's for all the responses.
It sounds like, theoretically, I could have wired the 4 ohm woofers in series and mated them to an 8 ohm tweeter but possible complications, and the inability to get the woofers close enough together, would have made this inadvisable for a novice builder with no test equipment. So it looks like I made the right call.
FYI. The final project is somewhere in the sticky thread "System Pictures & Descriptions" #1863
Hello,
So you look in any of the Audiophile magazines and you see another version / model / brand of MTM. Does that mean that they are violating some rule of nature or do they have funny lobes?
Somehow I believe they feel that they are doing something correctly. How do the schools of thought differ?
DT
All Just for fun!
So you look in any of the Audiophile magazines and you see another version / model / brand of MTM. Does that mean that they are violating some rule of nature or do they have funny lobes?
Somehow I believe they feel that they are doing something correctly. How do the schools of thought differ?
DT
All Just for fun!
Building something vs going public with the idea via a patent application or writing an article about it are two entirely different things. Who know's, Joe could have built his first prototype in 1974...
Last edited:
Remember what a d'Appolito configuration is- it is NOT the same thing as MTM (those have been around for years before, as Dr. d'A clearly stated in his original paper), but a particular implementation of the MTM. The d'Appolito configuration required an odd order acoustic crossover, with specific driver spacing.
Interestingly, in later designs, d'Appolito used an acoustic 4th order LR, finding that it worked better. So paradoxically, these were MTM speakers from d'Appolito that were NOT d'Appolito configuration. In my last two sets of dynamic speakers, I found that indeed high slopes worked better, and that even with a 3k crossover, lobing was not an issue more than a meter or so away. The vertical dispersion did seem to be narrowed, which is a mixed bag. Great for seated listeners.
Interestingly, in later designs, d'Appolito used an acoustic 4th order LR, finding that it worked better. So paradoxically, these were MTM speakers from d'Appolito that were NOT d'Appolito configuration. In my last two sets of dynamic speakers, I found that indeed high slopes worked better, and that even with a 3k crossover, lobing was not an issue more than a meter or so away. The vertical dispersion did seem to be narrowed, which is a mixed bag. Great for seated listeners.
Although we commonly call it a D'Appolito array, Joe did not invent the array, which goes way back. He did write a paper that showed that with such an array, and with a crossover that combined the drivers with 90 degrees phase shift, you could partially fill in the directivity nulls.
I prefer to call them Symmetrical arrays. They would never be patentable: "obvious to practitioners of the art" the patent office would say.
David S.
I don't think it is semantics. We are clarifying that the invention was purely about the crossover relationship, while the physical configuration long precedes this.
Note that the essence is about the forced symmetry of the configuration. 90 degree phase shift between elements with a standard single woofer, single tweeter would give a lobe in one direction, null in the other. With the MTM arrangement it will, of course, be symmetrical with nulls equally placed around the listening axis.
I found with the Snell XA designs that an assymetrical crossover with a steep highpass and shallower lowpass gave more constant -6dB points.
David S.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.