Preliminary IMD measurements for subwoofers (1 Viewer)

[Ilkka R]

Steve's question is a good one and that is one reason why original 70 Hz was chosen. 30/65 Hz could be possible, but 2f1-f2 yields a 5 Hz component.

I think the lowest possible would be 30/71 Hz.

 

[RogerW]

There probably isn't a perfect answer to this. The two tones chosen have to be within the range that most subs can reproduce without undue distress. Ideally, none of the IMD components should coincide with a harmonic of either of the excitation tones (a problem with 30/70). In practice, some of the IMD components are always liable to end up on top of each other, but if we are primarily interested in Total IMD% this should not matter too much. We don't want any IMD components below 10Hz or we can't measure them as accurately as we would like. And we don't want any IMD or HD components to be so close together in frequency that our spectrum analyzer lacks the resolution to discriminate between them. Finding a pair of frequencies that can fulfil all that lot is quite a tall order.

 

[Edward J M]

30/53 yields the following IM components:

83
23
7
76
113
136
166
46
189
143
129
37
16
99
196
219
67
182
173
242
249
69
272
226
152
14
295
203

And the following harmonics:

60106
90159
120212
150265
180318
210371
240424
270477
300530

+++++++++++++++++++++++++

30/65 yields:

95
35
5
100
125
160
190
70
225
155
165
25
40
135
220
255
55
230
185
290
285
105
320
250
200
10
355
215

60130
90195
120260
150325
180390
210455
240520
270585
300650

+++++++++++++++++++++

30/72 (suggested by Ilkka via email) yields:

102
42
12
114
132
174
204
84
246
162
186
18
54
156
234
276
48
258
192
318
306
126
348
264
228
24
390
222

60144
90216
120288
150360
180432
210504
240576
270648
300720

+++++++++++++++++++

While S/T SpectraPro is capable of 0.008 Hz resolution, the settings we're considering using provide a 1.5 Hz spectral line resolution, which should be capable of handling a few of the close ones.

Ed

 

[Warren_Sc]

Subwoofer C is the AVS-12 model Illka was banned for marketing.

...Sherv... pass the popcorn... I found you guys...

 

[Warren_Sc]

Wouldn't it be possible to just hold one tone (say 30hz) and move the other tone around (changing its frequency) and see which combinations give the highest level of IMD?

PS This is really great stuff.

 

[Edward J M]

Yes, this is possible. SpectraPro has a multi-tone generator which can output up to 10 tones simultaneously, or you can pick/choose which ones. It would only take a few seconds to check all of them.

However, which gives the higher %IMD and which gives what subjectively might be perceived as the more offensive IMD are two different things.

For example, with harmonic distortion, we find higher orders more audible/offensive than lower orders, and we also find odd orders more offensive than even orders.

The 30/53 case is interesting because it generates a lot of "in-band" and "near-band" IM components, with many of them falling in the

 

[dave alan]

This one looks the most interesting to me, also, FWIW.

Dave

 

[Robb Roy]

Ilkka,

Great work! It's nice to see some hard numbers on the subject. I've been so busy with 80+ hour work weeks that I haven't been following things too closely, but allow me to express my surprise that this got you banned at AVS. When did AVS become a front for the Flat Earth Society?

-Robb

 

[Ilkka R]

Thanks Robb.

AVS changed since they became the AV Scientology instead of AV Science. Luckily we still have other good forums.

 

[Edward J M]

Looks like we lost some good posts during the HTF server/database problems - that's too bad. :frowning: Anyone willing to repost would be appreciated.

Anyway......I ran some IMD tests over the long weekend on a review subwoofer. I did 30/53 and 30/72, both at 95 dB and 100 dB.

This particular subwoofer struggled to hit (and didn't quite make) the 100 dB mark. This is because its 10% THD limit and its output compression limit at 30 Hz are both just under 100 dB (2 meters ground plane). It can play louder than 100 dB at 53 Hz and 72 Hz, but 30 Hz was the limiting factor.

After looking at the spectral display for both and listening to the tones being played, I've got to swing toward 30/72 as having a more uniform distribution of IM components across the pass band, including a large-ish one at 12 Hz and a few between the two test tones too. Subjectively it also sounds more objectionable to my ears, but YMMV of course.

I'll have to calc IMD for both data sets to see which set of test tones provides a higher value. It will also be intersting to calc the combined THD from both tones to see if THD is higher than the IMD. At first blush just by glancing at the spectral display, I suspect (as does Ilkka who has seen this data) that this is indeed the case.

If the "higher THD than IMD" trend holds over time for several tested subwoofers, it will serve to deconstruct the notion that a woofer can have low THD but high IMD....and it will also allow THD to be used with confidence as a proxy for IMD - at least on a straight percentage basis. It would not address the subject of harmonic/IM weighting for relative audibility or offensiveness, though.

 

[Warren_Sc]

Hi-

What's the latest?

Thanks,

 

[Ilkka R]

My next measurement session is due to happen quite soon (in 2-3 weeks), so after that we will be having loads of comparable IMD data for many different subs. Including both low end and high end subs.

The fundamentals will be 30 Hz and 72 Hz. I will probably be testing IMD at three different levels, 90 dB, 100 dB and 110 dB. Maybe even higher if the subwoofer can do it.

 

[Tom Vodhanel]

>>>The fundamentals will be 30 Hz and 72 Hz. I will probably be testing IMD at three different levels, 90 dB, 100 dB and 110 dB. Maybe even higher if the subwoofer can do it

 

[Ilkka R]

Yes, naturally it will be 2 m GP. And I wouldn't be so sure about the other one... There aren't many, but there are some.

 

[Edward J M]

Wow - 10 dB is a big jump (a 300% increase in acoustic output). If the sub craps out between 10 dB intervals, maybe you could fall back on 5 dB intervals.

 

[Ilkka R]

It is a big jump, but I would like to cover as large amplitude range as possible, but I don't have time to use very small steps...for example 90, 95, 100, 105, 110 dB. But we'll see...

 

[Edward J M]

It sounds like a massive undertaking. Better bring the gas grills and plenty of food for a big cook-out.

Take plenty of pics - looking forward to seeing/reading everything.

 

[Warren_Sc]

I am looking forward to it...

...just make sure all of that festive stuff does not effect the noise floor...


Btw, if possible there should be time allowed to one test within your intervals to reduce potential questions. For example, maybe the test is 90, 100, 110, then if one frequency craps out you can squeeze in a 95 or a 105. Then again, I'll take whatever you can do. Thanks!

 

[Lee-c]

Interesting tests. Have the results of any of the outdoor tests come in yet?

 

[Ilkka R]

Yes, http://www.hometheatershack.com/foru...t-results.html

Check also other measurements, http://www.hometheatershack.com/forums/subwoofer-tests/