Regarding James W Johnson's Pi Speaker

Discussion in 'Archived Threads 2001-2004' started by Scott Simonian, Jul 2, 2001.

  1. Scott Simonian

    Scott Simonian Screenwriter

    Joined:
    Jun 20, 2001
    Messages:
    1,281
    Likes Received:
    0
    Trophy Points:
    0
    Wouldnt a BIG 15" woofer sound really lousy on most of the frequencies? I mean, wouldnt it sound muddy and undefined in the WHOLE mid-range frequencies?
    Im just not sure. I might be building my own full-range towers after I build my first DIY sub and would like the best performance.
    It seems like you are just looking for high SPLs. Im not sure. I could be interpreting the topic or your selection or just my knowledge on these things wrong.
    Hell, Im all for BIG, BADASS woofers !!!
     
  2. Greg Monfort

    Greg Monfort Supporting Actor

    Joined:
    May 30, 2000
    Messages:
    884
    Likes Received:
    0
    Trophy Points:
    0
    It depends on the driver's design. Study the 15" ~full range designs of the 40s-60s, such as Tannoy, Goodmans, Altec. WRT the Pi's, please re-read my response about XO points, etc..
    GM
    ------------------
    Loud is beautiful, if it's clean
     
  3. James W. Johnson

    James W. Johnson Screenwriter

    Joined:
    May 26, 2001
    Messages:
    1,055
    Likes Received:
    0
    Trophy Points:
    0
    I have asked the owner of Pi Speakers to respond, he is a really nice guy with alot of experience, his name is Wayne Parham.
    There is alot of good info at the Pi speaker open forum..http://www.audioasylum.com/forums/pi/bbs.html
    ------------------
    James' DIY speakers
     
  4. Greg Monfort

    Greg Monfort Supporting Actor

    Joined:
    May 30, 2000
    Messages:
    884
    Likes Received:
    0
    Trophy Points:
    0
    That's not a response of the speaker shown in your thread. Far from it.
    GM
    ------------------
    Loud is beautiful, if it's clean
     
  5. James W. Johnson

    James W. Johnson Screenwriter

    Joined:
    May 26, 2001
    Messages:
    1,055
    Likes Received:
    0
    Trophy Points:
    0
    No it is not far from it Greg.
    I am pretty sure that this response is from a 10 Pi pro series which use the same drivers as the 4 Pis. The cabinet volume of the 10 Pi is not much more than the 4 Pis. The main difference is that the bass driver in the 10 Pi is horn loaded.
    I have a feeling that many speaker designers are almost offended that someone would claim that such a simple system can perform so well.
    ------------------
    James' DIY speakers
     
  6. James W. Johnson

    James W. Johnson Screenwriter

    Joined:
    May 26, 2001
    Messages:
    1,055
    Likes Received:
    0
    Trophy Points:
    0
    Scott, you might want to read this..you will see that Wayne has been at this a long time and really knows his stuff.....http://www.pispeakers.com/PiAlign.doc
    ------------------
    James' DIY speakers
     
  7. Greg Monfort

    Greg Monfort Supporting Actor

    Joined:
    May 30, 2000
    Messages:
    884
    Likes Received:
    0
    Trophy Points:
    0
    That horn adds a lot of midbass, and a bit of LF gain. WRT BoxPlot, this shows a idealized half space response. Plot it in LspCad to see how it compares to a more realistic in-room response.
    I like BoxPlot for quickie comparisons, but that's all it's good for. Also, did he factor in VC heating, as it doesn't look like it.
    GM
    ------------------
    Loud is beautiful, if it's clean
     
  8. Scott Simonian

    Scott Simonian Screenwriter

    Joined:
    Jun 20, 2001
    Messages:
    1,281
    Likes Received:
    0
    Trophy Points:
    0
    Uh oh. I hope when I said "full-range", you didnt think I meant a one-way system. Geezuz, maybe you guys are just confusing me.
    Maybe... { : |
    Yeah, Im confused.
     
  9. Greg Monfort

    Greg Monfort Supporting Actor

    Joined:
    May 30, 2000
    Messages:
    884
    Likes Received:
    0
    Trophy Points:
    0
    No, I was responding to your comment that 15" drivers were incapable of ~accurately reproducing midbass, mids.
    GM
    ------------------
    Loud is beautiful, if it's clean
     
  10. James W. Johnson

    James W. Johnson Screenwriter

    Joined:
    May 26, 2001
    Messages:
    1,055
    Likes Received:
    0
    Trophy Points:
    0
    From Wanyne Parham, owner of Pi speakers.....
    >> Wouldnt a BIG 15" woofer sound really lousy on most of the frequencies?
    >> I mean, wouldnt it sound muddy and undefined in the WHOLE mid-range frequencies?
    >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
    The problem with larger drivers in the midrange is primarily caused by two things:
    1. Rolloff from the motor's inductive reactance and the cone's inertia (mass), and
    2. The cone's tendency to flex in a twisting pattern along it's diameter, which is often referred to as "breakup mode."
    Effects of the first - Rolloff - will result in a decreased amplitude at higher frequencies. This is described by their frequency response graph. And effects of the second are described by their distortion graph. It is also often evidenced by a peak or two in the midrange, if it hasn't already rolled off by inductance or cone mass, in which case the second potential problem is a non-issue.
    See the performance data of the JBL 2226 atwww.PiSpeakers.com/JBL_2226.pdf. The frequency response is reasonably flat to about 2Khz and its distortion is extremely low. While distortion rises in the top octave - due to cone flex - it is still far better than average. Specifically, it's about 10 times better.
    >> It depends on the driver's design. Study the 15" ~full range designs of the 40s-60s, such as Tannoy, Goodmans, Altec.
    >> WRT the Pi's, please re-read my response about XO points, etc.
    >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
    I don't know what this person wrote, but he may allude to bringing the crossover point lower to decrease "breakup mode" distortion. He may also describe what some have referred to as "beaming" (lower dispersion) at higher frequencies. If either or both of these were discussed, one might consdier the fact that both are accurate but not troublesome, in my opinion.
    The design choices for Pi Speakers shown on the website are those that provide very high quality, very high performance systems at reasonable costs and in relatively compact packages. Any design choices that require the band to be split to more drivers significantly increase the size or the cost, or both. But these kinds of systems are also available from Pi Speakers, and we discuss building them all the time for the DIY crowd. They're available.
    But I generally don't suggest them and here's why. Look back at the spec sheet of the 2226. See that it's distortion is about -45dB from 100Hz to 1KHz, and that it rises about 10dB between 1Khz and 2Khz? That is measured distortion at 60 watts, and it represents about 32,000 times less power than the fundamental, or about 0.003% distortion. Now when we look at the "hideous" distortion we get from these motors at 2Khz, which is still 35dB less than the fundamental, we're still only talking about 0.03% distortion at 60 watts. That's a whole lot better than the average driver, which clocks in at about 3%. Three percent is actually pretty good, and the JBL's worst is well under one percent.
    You can run four Pi-18's, and get around this "hideous" distortion by moving the crossover down an octave. That's not prudent with 1" drivers. Alternately, you could simply run the 2446 with the 2226. But the point is that the drivers aren't inexpensive, and I don't believe it's merited in the design. After all, we're talking about some really, really good performance figures. Less than 1% distortion - even in this critical region - way less.
    >> It does not get much better than this....(the response chart)
    >> That's not a response of the speaker shown in your thread. Far from it.
    >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
    This is a true statement, but the remark "far from it" is misleading. Comparing the performance of the four Pi, I would not consider the ten Pi's response curve to be "far from it." But the response curve of the ten Pi is what's shown above. Here's what is added to the four Pi's response curve to gain the response of the ten Pi:
    [​IMG]
    The curve shown above is the horn's participation in the system. It augments the bottom octave, while at the same time increasing output up to about 400Hz. You can see it's peak at horn cutoff - around 45Hz - and then its negative slope beyond that point.
    >> No it is not far from it Greg.
    >>>>>>>>>>>>>>
    I would agree with that statement.
    >> I am pretty sure that this response is from a 10 Pi pro series which use the same drivers as the 4 Pis.
    >>>>>>>>>>>>>>>>
    This is exactly the truth.
    >> The cabinet volume of the 10 Pi is not much more than the 4 Pis.
    The motor cabinet of the ten Pi's is exactly the same volume as the four Pi's. It is tuned at the same frequency, and with the same port. In fact, the motor cabinet of a ten Pi is a four Pi.
    Some might object to this usage of a conical horn. But those are the kind of guys that argue about how many angels dance on the head of a pin, and never build anything useful because they're too busy worrying about things that are almost completely unimportant. They miss the truly significant things. These are often the types that cannot work with machines having more than a a few moving parts, so they focus on loudspeakers because they have so few moving parts. They put them under a magnifying glass, but they are ill-equipped to properly understand what they see. Still, they'll talk all day long about all the angels they think they see.
    These kinds of horns work great as basshorns. Conical horns have a peak at cutoff with a negative slope. Match that 6dB peak at cutoff with the motor cabinet's -6dB rolloff point, and you've just built yourself a linearity enhancing device.
    But even without the horn, you've got a very good loudspeaker. Some would argue that it's more linear, and you can see why from the response chart showing the horn's participation. I see merit in both designs.
    >> The main difference is that the bass driver in the 10 Pi is horn loaded.
    >>>>>>>>>>>>>>>>>
    That's right. It's the only difference, except a minor change in the attenuation of the HF driver - It isn't attenuated as much because bass below 400Hz is greater on the ten Pi's. There's a few ohms more series resistance on the HF driver in a horn loaded Pi speaker than there is on the (otherwise exact same) bass reflex Pi speaker.
    .
    >> This is a response of the Four Pis...
    >>>>>>>>>>
    This is the response curve of the Professional Series four Pi, and it is published on the Pi Speakers website. This shows the lower few octaves. It shows what is produced by the four Pi, and that there is significant energy above 30Hz - with the system having a -6dB point at 50Hz. It is this which is summed with the horn's participation to create the ten Pi's response curve.
    >> The red curve shows frequency response at full output, green shows maximum power and black
    >> is response relative to unity. Units of measurement are shown for each, in their respective colors.
    >>>>>>>>
    This is true, except that the red curve is what Rick Carlson, author of BoxPlot, considers to be "safe response" at full output. Where impedance falls, the assumption is that power handling will also fall, so maximum output must be compensated or the structural integrity of the speaker might be compomised. The truth is that the red curve should look just like the black curve, and the difference between the two curves should be noted as a possible failure mode. But the red curve is a pretty good indication of what a person can expect to be "maximum safe response."
    You'll notice that 30Hz, the system at full power begins to fall off more rapidly than the black curve. It's not that at full power the speaker will rolloff more rapidly, it's that at full power, the speaker is more vulnerable below 30Hz. In this particular design, the difference is marginal and can be overlooked. You can see that the two curves are pretty well matched. But I've seen designs that had pretty significant regions with this kind of impedance anomoly - strictly a function of system tuning and quite common in optimally tuned cabinets containing speakers having high Thiele/Small Q or low Pi Q values. Note that saying "high Thiele/Small Q" or "low Pi Q" is saying the same thing - they're essentially reciprocals of each other.
    I don't happen to incorporate those kinds of low (Pi) Q motors, but it's important to understand nonetheless. If you're running a speaker having high Thiele/Small Q values, like a subwoofer or other high Q motor, you'll see a very sharp falloff in this "maximum safe power" curve. This does not indicate that the speaker becomes any less linear at high power levels. It indicates that such a system may begin to distort heavily, fatigue, "bottom out" the voice coil, or fail entirely at high power levels.
    >> That horn adds a lot of midbass, and a bit of LF gain.
    >>>>>>>>>>>>>>>>>>>>
    That is evident by the horn's participation chart. Peak at cutoff with a negative slope.
    >> WRT BoxPlot, this shows a idealized half space response.
    >>>>>>>>>>>>>>>>>>>>
    Actually, the curve is one that's been predicted as free air response. It isn't Pi/2 and it certainly isn'y some kind of "idealized Pi/2." Where did this writer come up with the idea of "idealized half space?"
    >> Plot it in LspCad to see how it compares to a more realistic in-room response.
    >>>>>>>>>>>>>>>.
    One could certainly do that. I'm not familiar with LSPCad. I do have a copy, but I'm not really sure why I downloaded it. That's not entirely accurate - I do know why. I found PiAlign on an "audio shareware" site a few years back, and decided to look at some of the other offerings on that site. So many things had been written in the two decades since PiAlign was written, and I just wanted to look at them. Lots of neat stuff, but I didn't really need the tools, so I didn't take the time to learn them.
    Sort of back to that thing about the systems with only two moving parts. Just how many calculation tools do we need? PiAlign was a big help over hand calculators and slide rules, but we did in fact send a men to the moon with slide rules and I did in fact design several Pi Align'ed speakers with hand calculators. The speakers were probably easier than the moon thing. And PiAlign really made it easy.
    The other, newer programs have more attractive GUI's. But really, the only two I've found that I use regularly are Rick Carlson's BoxPlot and David McBean's Hornresp. And really, I only use these because they accurately predict response curves, just as accepted formula show they should. So I only use them to plot response curves, because it's easier than doing it by hand. Slide rule or calculator.
    >> I like BoxPlot for quickie comparisons, but that's all it's good for.
    >>>>>>>>>>>>>>>>
    I think this fellow must have a little "penis envy" of Rick Carlson. BoxPlot is my favorite, and is the most useful of all the T/S programs I've found.
    >> Also, did he factor in VC heating, as it doesn't look like it.
    >>>>>>>>>>>>>>>>>>>>>>>>>>>>>>
    This is relatively unimportant, as compression is shown in the performance specifications for the drivers. Measured data is more accurate than a computer prediction, because it's measured. I'm pretty comfortable having a computer simulation of a Helmholtz resonator, a inductive/capacitive resonator, or an acoustic filter chamber. These are relatively simple systems and easy to simplify into an accurate model. But heat flow is a different story. Honestly, the only accurate modeling of compression from voice coil heating could be done on FEA software, and this still wouldn't be as good as the measurement aquired data.
    For your four Pi's, you'll notice compression specs given for each driver, taken at three power levels. If memory serves, maximum compression is 4dB. That's why the 2226 has maximum outputof 120dB instead of 125dB - 10log600 is 124.8dB, but you must remove 4dB for "VC heating" - compression - and so we're calling 'em 120dB speakers.
    Check out "Finite Element Analysis" software sometime to see what I mean. Those are the kinds of programs that come close to modeling this type of heat flow, and it's not really appropriate to expect any sort of approximation of compression from a program that's given only T/S specs. The program would require a pretty detailed 3D model of the speaker motor, along with the construction materials used.
    Anyone who expects this kind of a loudspeaker program to give power (heat flow) results accurate to that level is pretty self-decieved. One should probably disregard his comments, or at least consider them to be questionable until verified.
    Wow! That was a bunch of stuff! Bunch of bunk, Yes?
    Take care!
    Wayne
    ------------------
    James' DIY speakers
    ------------------
    James' DIY speakers
     
  11. James W. Johnson

    James W. Johnson Screenwriter

    Joined:
    May 26, 2001
    Messages:
    1,055
    Likes Received:
    0
    Trophy Points:
    0
    Sorry Jack its my fault because I copied and pasted his e-mail directly to the thread.
    I will ask him if he would like to register at this forum.
    ------------------
    James' DIY speakers
     
  12. James W. Johnson

    James W. Johnson Screenwriter

    Joined:
    May 26, 2001
    Messages:
    1,055
    Likes Received:
    0
    Trophy Points:
    0
    He'd be a valuable addition. >>>>>>>
    Yes he would [​IMG]
    I should also add that Wayne is one of the nicest guys I have ever met on the internet.
    Hell I feel guilty that I almost strayed away from building the four Pis in leu of the Unity Horns.
    He gives is plans away for nothing!!!, How many speaker designers with his kind of experiece are doing that?
    It is set in stone for me , I will be building the Four Pis [​IMG] I cannot wait to hear them!
    Today I begin gathering the supplys that I will need.
    ------------------
    James' DIY speakers
     

Share This Page