Intermodulation Distortion midbass shootout purifi anarchy plus maybe more

The following is a comparison of Intermodulation distortion (IMD) in 6-7" midbass drivers.

One of my lockdown speaker projects was to measure harmonic distortion and IMD in 14 10-12" subwoofers. See the thread here: A tale of 12" subwoofers, distortion and 15 dollars.

What I learned from this exercise is that the drivers that sounded subjectively more clear had lower IMD levels.

I thought I would continue the fun and measure a few 6-7" midbass drivers. I'm sure most people on this forum know, but IMD is similar to harmonic distortion except more than one tone is used. For my test I used tones at 40and 96hz as well as 50 and 210hz. Many years ago Paul Klipsch wrote a paper on how IMD is linked to cone excursion so I wanted to test the drivers in a situation where cone excursion is high and the IMD products are above the noise floor.

The test enclosure is a .5 cu ft (14 liter) sealed box with the bottom lip of the speaker just 2 inches above the floor. The microphone is on a stand 49"high with the base of the microphone 33" from the driver. A distance of about 1.5m from the speaker to the microphone. As I'm not measuring in an anechoic chamber my measurements are not comparable to anyone else's but comparisons can be made between the drivers I test.

The first two drivers I was able to test are the DIYSG Anarchy and the Purifi PTT6.5X04-NFA-01. As shown in the graphs the Purifi really shines in this IMD test. Yes, the anarchy is less than 1/5 the price and it does well for the money but the Anarchy seems to be affected by a lack of inductance control.
Erin's audio corner has klippel tested the purifi (XBl at 82%, Bl = 9.1mm, and for XC at 75%, Cms minimum was 6.9mm and the anarchy was klippel tested for the voice coil magazine (XBl at 82%, Bl = 8.5mm, and for XC at 75%, Cms minimum was 10mm). So a pretty close klippel result but the purifi outperforms the anarchy in all IMD and harmonic distortion tests I conducted.

For the 40 and 96hz test look for IMD products at 56,136,152,176,232
For the 50 and 210 test look for IMD products at
60,110,160,260,310,370

I will later convert the graphs into IMD percentage.
I am open to comments and recommendations of drivers that might do well in this test.
 

Attachments

  • Combined40IMD70db.jpg
    Combined40IMD70db.jpg
    242.7 KB · Views: 1,134
  • Combined40IMD80db.jpg
    Combined40IMD80db.jpg
    243.7 KB · Views: 1,088
  • Combined40IMD85db.jpg
    Combined40IMD85db.jpg
    241.3 KB · Views: 1,044
  • Combined50IMD70db.jpg
    Combined50IMD70db.jpg
    248.3 KB · Views: 1,011
  • Combined50IMD80db.jpg
    Combined50IMD80db.jpg
    229.2 KB · Views: 1,002
  • Combined50IMD90db.jpg
    Combined50IMD90db.jpg
    249.9 KB · Views: 533
  • IMG_2980[1].jpg
    IMG_2980[1].jpg
    880 KB · Views: 702
Last edited:
  • Like
Reactions: 2 users
The following is a comparison of Intermodulation distortion (IMD) in 6-7" midbass drivers.

I am open to comments and recommendations of drivers that might do well in this test.

Are your results consistent with Paul K's analysis? Then the way to do well in this test is to minimize cone excursion - use horn loaded large diameter drivers, which was his point and business plan.

Purifi no doubt has done an excellent job on HD but they are up against simple physics on IMD.
 
nc535, according to klippel bl and inductance nonlinearity, doppler and cone vibration are sources of IMD. In P. Klipsch's paper http://www.readresearch.co.uk/loudspeaker_papers/klipsch_modulation_distortion_article_1.pdf he concludes that horn loading is the best way to reduce excursion and therefore IMD.

However as drivers improve the difference between horn loaded bass and a direct radiator are getting closer. I have measured 30 and 72hz IMD with a high quality 12" subwoofer with a full copper sleeze to reduce inductance nonlinearity and also with a basic 8" woofer in a 7 foot folded horn that I built from plans.

Attached are the two IMD measurements in room with same location of microphone and speaker. Take a guess which one is the horn and which one is the sealed 12"?
 

Attachments

  • 1.jpg
    1.jpg
    134.3 KB · Views: 559
  • 2.jpg
    2.jpg
    135.8 KB · Views: 557
No guesses? It'd have to be the right driver for the horn usually high bl drivers work better. I'm also thinking cone vibration is pretty important with the increased compression of the horn as the eminence lab 12" which was designed for horn use had two dust caps to stiffen up the cone. I think the klipschorn uses a 15".
 
in Doppler distortion Purifi is probably also better as Doppler is purely excursion related and Purifi had higher sensitivity.
One example how big Doppler distortion can actually be. It is from relatively low sensitive full range, 70 Hz and 10 kHz are with same level.
 

Attachments

  • DopplerDistortion.png
    DopplerDistortion.png
    76.3 KB · Views: 449
Last edited:
turk 182, I'm sorry I only have one test box so I can't do an A/B test. However I did listen to a few select test tracks with the anarchy and then quickly changed to the purifi.

I'm always listening for good reproduction of the string bass as I played this in the college orchestra. String bass with the purifi is much more clean than the anarchy. Kick drum in the song "Circle of Peace" by Ziggy Marley sounds similar between the two but the bass guitar seemed more articulate with the Purifi.
 
Last edited:
Attached is the IMD percentage chart of the two drivers.
Shrub0,

A 6 dB increase in SPL requires a doubling of excursion, a linear ratio.
Bl (magnetic force factor with the voice coil at rest) reduces with excursion in a non-linear ratio.
Klippel analysis of Bl when reduced to 82% at "X" excursion will produce about 10 % intermodulation distortion (IMD).

The Purifi PTT6.5X04-NFA-01 stated "linear Xmax" is 9.8mm, while mechanically rated 14.5mm, due to the Bl being weaker than the mechanical formula (HVC – HG)/2 (Height of Voice Coil- Height of magnetic Gap) for Xmax predicts.
The Anarchy 704 Woofer 12mm Xmax rating is probably a mechanical rating, if its Bl curve and suspension compliance is similar to the Purifi, rated the same way it's "linear Xmax" would drop to 8.1mm.

Did you happen to measure the excursion of the drivers during the 85dB 40/96Hz or 90dB 50/210Hz tests?

Art
 
Last edited:
Persistence of vision makes it pretty easy to see the excursion with a tiny white or silver dot (I use a Silver Sharpie pen) near the surround. Look from the side, compare excursion to a ruler held at 90 degrees. Divide by 2 (Peak to Peak/2=Xmax), easier if you make 2mm hash marks on a white business card.
Shrub0,

You can also use a video camera /phone and look at it later without needing to get so close- though at 90dB I wouldn't worry.
With my vision, I had to lean into some subs to see the dot, but at 130dB+ level, vision starts to blur- wish I'd thought of using a video camera then..

Art