Nowadays there's no short of information about driver's harmonics distortion. For example when we compare Scanspeak 15WU/8741T00 vs SBA SB15CAC30 from a famous German DIY magazine, it's pretty easy to see both are great drivers, but SB15 seems to be a slightly better in all regards. After all, SB15CAC30 is better at 3rd harmonics(HD3) in all frequencys, and equal or better 2nd harmonics(HD2) thanks to a more stiff cone. So conclusion is pretty clear right? As long as you are ok with 5mm xmax, SB15CAC30 should be a better driver:
Until I'm interested to do a 30hz+255hz(4:1) test my self. Test condition is not ideal but made as fair as possible. I test free raw driver with nearfield MIC, same distance for both driver, with about 3.5mm~4mm one way excursion on SB15 and match the bass tone(not voice tone) on 15WU. After all, voice tone is 12db lower in voltage and should be handled easily for both drivers. Both drivers are my own collections so there's no need for bias here. Here's the result:
SB15:
15WU/8741:
Not only SB15 shows much higher HD2, HD3, IMD2, IMD3; but the waveform is visually distorted, which is a bit shocking here. Considering the price, 15WU in this particular test is clearly another class or two higher than SB15, which it should be, but considering SB15 metal cone drivers doing so well in almost all DIY audio tests and is regarded so highly, and 15WU is generally regarded not so well by DIY community, this is still a bit surprising.
So I decided to do similar tests for my 4 inch driver collections, but with 80hz+680hz(4:1). I choose these frequencies because 80hz is higher than most driver's FS in this group, can produce reasonable loud SPL in home use. I pushed the drivers to around 2.5~3mm XMAX one-way so that it stayed within most driver's linear excursion range(except W4-2142). I use Linkwitz ranking system to give each driver an average score by their ranking in each category, instead of a subjective ranking. Most ranking are quite straightforward, but I want to add another one: Amplitude stability. I'll explain how it works:
In this graph, I compared the voice tone amplitude for the ones from bottom and top bass tone respectively, and derive a score by its min/max. In this example, bottom voice tone is 62 pixel height, and top one is 57 pixel height, thus got a score of 92% in amplitude stability. This is still very rough, but is easy to do, and at least provide a bit of information. This test should be mostly indicative of BL curve symmetry, but also probably reluctance force modulation. To be honest I'm not sure.
Another note: In IMD rankings, I bias more on IMD components on the LEFT. on the right it should be easily masked just like HD2.
Here are the results:
2: W4-2142: Performs very well, also underhung driver despite smaller XMAX.
3: W4-1720: Huge underhung motor, very bad HD but decent IMD. Honestly I expected it to perform much better in this test.
4: 830854: Very good driver, and it performs slight better than the NE series from the same company.
5: NE123W: Good performer, I didn't expect it to do well in IMD test due to poor-ish Klippel tests in this series. But holds up well.
6: W4-1337SD: Another good performer. I expected it to do better, but IMD3 is not very good.
7: WF120BD01: Wavecor has excellent reputation in almost all non-linear tests, and ZaphAudio regarded it as an high-end all-rounded driver, but in this test it's a bit disappointing. IMD2 is particularly high.
8: SB12PAC25: Exceeded my expectation, considering this driver has no shorting rings, very close to WF120BD01 which has a fancy motor.
9: W4-1052SD: Ok performer. Considering the HD test is pretty good, IMD performance is a bit disappointing.
10: CF120: Worst in this group, a bit shocking considering this driver has copper cap in motor and has long throw. Actually I feel this driver is not bad at all, maybe the 80hz bass tone caused some unfortunate air turbulence out of the tiny hole on the back the driver. Higher or lower frequency it performs better.
That's all. Thank you!
Edit: Peerless 830854 has been added.
Until I'm interested to do a 30hz+255hz(4:1) test my self. Test condition is not ideal but made as fair as possible. I test free raw driver with nearfield MIC, same distance for both driver, with about 3.5mm~4mm one way excursion on SB15 and match the bass tone(not voice tone) on 15WU. After all, voice tone is 12db lower in voltage and should be handled easily for both drivers. Both drivers are my own collections so there's no need for bias here. Here's the result:
SB15:
15WU/8741:
Not only SB15 shows much higher HD2, HD3, IMD2, IMD3; but the waveform is visually distorted, which is a bit shocking here. Considering the price, 15WU in this particular test is clearly another class or two higher than SB15, which it should be, but considering SB15 metal cone drivers doing so well in almost all DIY audio tests and is regarded so highly, and 15WU is generally regarded not so well by DIY community, this is still a bit surprising.
So I decided to do similar tests for my 4 inch driver collections, but with 80hz+680hz(4:1). I choose these frequencies because 80hz is higher than most driver's FS in this group, can produce reasonable loud SPL in home use. I pushed the drivers to around 2.5~3mm XMAX one-way so that it stayed within most driver's linear excursion range(except W4-2142). I use Linkwitz ranking system to give each driver an average score by their ranking in each category, instead of a subjective ranking. Most ranking are quite straightforward, but I want to add another one: Amplitude stability. I'll explain how it works:
In this graph, I compared the voice tone amplitude for the ones from bottom and top bass tone respectively, and derive a score by its min/max. In this example, bottom voice tone is 62 pixel height, and top one is 57 pixel height, thus got a score of 92% in amplitude stability. This is still very rough, but is easy to do, and at least provide a bit of information. This test should be mostly indicative of BL curve symmetry, but also probably reluctance force modulation. To be honest I'm not sure.
Another note: In IMD rankings, I bias more on IMD components on the LEFT. on the right it should be easily masked just like HD2.
Here are the results:
Peerless NE123W:
Wavecor WF120BD01:
TB W4-1757SB:
SB Acoustics SB12PAC25:
TB W4-1052SD:
TB W4-2142:
TB W4-1337SD:
Dayton CF120:
TB W4-1720:
Peerless HDS 830854:
Here's a score table:
| Amplitude Stability | Bass HD | Voice HD | IMD2 | IMD3 | AM Ranking | Overall | |
| W4-2142 | 96.20% | 7 | 6 | 3 | 2 | 2 | 4 |
| NE123W | 95.30% | 9 | 3 | 4 | 6 | 3 | 5 |
| W4-1720 | 96.90% | 10 | 2 | 5 | 3 | 1 | 4.2 |
| CF120 | 90.70% | 6 | 9 | 7 | 10 | 6 | 7.6 |
| W4-1337SD | 90.50% | 5 | 5 | 6 | 8 | 7 | 6.2 |
| W4-1052SD | 87.80% | 8 | 4 | 8 | 9 | 8 | 7.4 |
| SB12PAC25 | 82.90% | 3 | 10 | 9 | 4 | 9 | 7 |
| W4-1757SB | 93.80% | 2 | 1 | 2 | 1 | 4 | 2 |
| WF120BD01 | 81.80% | 1 | 7 | 10 | 5 | 10 | 6.6 |
| 830854 | 93.30% | 4 | 8 | 1 | 7 | 5 | 5 |
And my IMD-biased rankings and comments:
1: W4-1757SB: Clear winner in this test. Great low distortion underhung driver.2: W4-2142: Performs very well, also underhung driver despite smaller XMAX.
3: W4-1720: Huge underhung motor, very bad HD but decent IMD. Honestly I expected it to perform much better in this test.
4: 830854: Very good driver, and it performs slight better than the NE series from the same company.
5: NE123W: Good performer, I didn't expect it to do well in IMD test due to poor-ish Klippel tests in this series. But holds up well.
6: W4-1337SD: Another good performer. I expected it to do better, but IMD3 is not very good.
7: WF120BD01: Wavecor has excellent reputation in almost all non-linear tests, and ZaphAudio regarded it as an high-end all-rounded driver, but in this test it's a bit disappointing. IMD2 is particularly high.
8: SB12PAC25: Exceeded my expectation, considering this driver has no shorting rings, very close to WF120BD01 which has a fancy motor.
9: W4-1052SD: Ok performer. Considering the HD test is pretty good, IMD performance is a bit disappointing.
10: CF120: Worst in this group, a bit shocking considering this driver has copper cap in motor and has long throw. Actually I feel this driver is not bad at all, maybe the 80hz bass tone caused some unfortunate air turbulence out of the tiny hole on the back the driver. Higher or lower frequency it performs better.
That's all. Thank you!
Edit: Peerless 830854 has been added.
Last edited:
Isn't waveform fidelity (of the waveform measured from the driver's acoustic output) a direct consequence of the nonlinearity profile getting applied during the transduction process from electrical signal to final acoustic output?
i.e., isn't what happens in frequency domain itself telling us what is going to happen in time domain? (That the waveform is going to get distorted with higher levels of distortion. I agree that it may be more intuitive to see what happens to the waveform in the time domain from the time domain plot itself).
In that case, wouldn't having a ranking system that weighs separately both time domain and frequency domain "views" of the same phenomenon happening, a little unfair, in the sense that the ranking may now be doubly punishing a driver?
Thanks
Vineeth
i.e., isn't what happens in frequency domain itself telling us what is going to happen in time domain? (That the waveform is going to get distorted with higher levels of distortion. I agree that it may be more intuitive to see what happens to the waveform in the time domain from the time domain plot itself).
In that case, wouldn't having a ranking system that weighs separately both time domain and frequency domain "views" of the same phenomenon happening, a little unfair, in the sense that the ranking may now be doubly punishing a driver?
Thanks
Vineeth
No it's not. A raw driver could have a very non linear frequency response. A finished loudspeaker will alter the response to hit a particular target. Without at least equalizing the amplitude of the drivers to be the same at least at 80hz+680hz these measurements are not particularly useful.
The test with the SB15CAC30 shouldn't have had the low tone at 30Hz. No competent designer would run that driver so low.
The test with the SB15CAC30 shouldn't have had the low tone at 30Hz. No competent designer would run that driver so low.
No, I use 4:1 for voltage amplitude. Drivers has different sensitivity for voice tone, for example W4-1720 has low sensitivity around 680hz, so I judge voice HD based on voice acoustic amplitude.
I think about that too. In fact amplitude stability seems to have correlation with IMD2, but I just provided it as an extra information.
Last edited:
Equalizing 680hz is nice to have, but too time consuming for me. The reason behind I already explained, is that 680 tone input is 12db lower than 80hz tone, it's not stressful for the driver at all. All distortion is judged relatively, IE. voice tone 2nd HD is measured by its 1360hz component vs the corresponding 680hz acoustic component, not a same single reference across all drivers.
The 30+255 test was inspired by Purifi and Hificompass. They even run 30hz+255hz for 4 inch drivers. I think the main reason for the test is to reveal motor non-linearity without access to Klippel. We already has nearfield HD test in Hificompass, but it's hard to tell if it came from suspension, or motor. With 30+255 tests, we have more information to separate those causes for distortion. Think about a BL curve, IMO it should not matter what frequency you run Klippel test to obtain an BL curve.
Difference is I run them as naked driver. Without baffle it makes distortion looks much worse, since stronger cancellation for lower frequencies. But relatively condition is the same for all drivers, and they all have similar basket size. For example, SB15 and 15WU are all 15cm drivers. I believe they are comparable in this particular condition and test.
Last edited:
Yes, 100%. This driver is often used as a midrange. When used as a true mid-woofer in a small 2-way, the bass response will extend down to 70 Hz at the lowest.
As I explained before, running an 30+255 test has nothing to do with whether we will use it down to 30hz or not.
In fact, if you make a 2 way passive bass-reflex with SB15NAC, you can only decide where you tune the driver, you can't decide if it will play 30hz tone. It's decided by input signal, because most home audio people have no high-pass filtered playback system. And the higher you tune the speaker, the higher infrasonic excursion (and modulation) it will occur. It's the opposite to protect driver excursion at very low frequencies. The same with sealed, but it's much better.
Go to any audio show with a turntable as audio source, you will see those passive systems ALL have funny excursions at maybe 1hz. It's not even in mastered audio signal, it's just the nature of turntable.
In fact, if you make a 2 way passive bass-reflex with SB15NAC, you can only decide where you tune the driver, you can't decide if it will play 30hz tone. It's decided by input signal, because most home audio people have no high-pass filtered playback system. And the higher you tune the speaker, the higher infrasonic excursion (and modulation) it will occur. It's the opposite to protect driver excursion at very low frequencies. The same with sealed, but it's much better.
Go to any audio show with a turntable as audio source, you will see those passive systems ALL have funny excursions at maybe 1hz. It's not even in mastered audio signal, it's just the nature of turntable.
Last edited:
OP is running the 30hz tone just to get a specified excursion, before adding a midrange tone. This is standard practice when doing this sort testing, I don't get the pushback from others?
we do the 30Hz+255Hz IMD test to reveal position dependent amplitude modulation effects in both the motor and soft parts, ie mainly Bl(x) and Sd(x). The 30Hz is a bit arbitrary, could be lower as well. The main point is that the deep tone controls the excuriosn and then we have the 255Hz voice tone superimposed which has low excursion (excursion drops with frequency squared when abbove fs). The test is heavily inspired by HifiCompass. We can also do multitone analysis but the results can be quite hard to interpret. For the two tone 30Hz+ voice tone we can check how close we are at the Doppler limit (the intermodulation caused by the motion of the mebrane).
We have a Klippel but developed our own analyser that is more precise and gives more information in order to separate the distortion sources (eg. the 'Cunningham' distortion aka flux or reluctance modulation)
Cheers
Lars (co-founder at Purifi)
We have a Klippel but developed our own analyser that is more precise and gives more information in order to separate the distortion sources (eg. the 'Cunningham' distortion aka flux or reluctance modulation)
Cheers
Lars (co-founder at Purifi)
Isn't the amplitude modulations of IMD due to natural summing?
In that case, deviation from a pure mathematical summing of 2 frequencies at 4:1 is what should be measured.
Amplitude stability is a moot term not needed for...anything.
In that case, deviation from a pure mathematical summing of 2 frequencies at 4:1 is what should be measured.
Amplitude stability is a moot term not needed for...anything.
summing is linear if the system is linear so the spectrum of two tones is simply the two tones. You are right that if you look at the envelope of the sum of two tones then it looks modulated (beat pattern) but that is not amplitude modulation. (AM). AM shows up as side bands.
Sorry I don't understand what you mean.
Linear, mathematical summing of 2 frequencies will not have voice tone amplitude differences between the crest and trough part of the bass tone. They will be equal, hence 100% amplitude stability.
But I admit amplitude stability is a random term came up in my mind, nothing academic. IMD2 side band seems somewhat correlated, but I feel IMD2 is not very reliable, for example 830854 have almost no IMD2 side band on the left, yet "amplitude stability" is still easy to derive.
Last edited:
Funny, my desk speakers use the TB 1720s as midbass in a 2-way, and I was EQ'ing them against my corrected headphones. I had to pull almost 5dB out of the 170hz range to get it to sound similar. I wonder how much of that extra noise was HD vs signal. I didn't dislike it, either. I'd put a sub with these and go to town all day. Glad to hear they can handle a wide range, even while digging low.
A bit out of topic but -20db of distortion will not affect objective loudness in meaningful way, in the range of 0.1db at most. But distortion component will affect subjective loudness, how much is hard to say. I will say -20db of low order distortion in bass range is not very audible, as most music content has lots of harmonics itself.
I think room acoustics is always a biggest factor for desktop EQ. A huge peak between 100~200hz is very typical.
D1sco you got good ears. About 4-5dB dip with a Q1-2 centred around 150Hz is about right!
Yes.
This is how the pro's (as in professional- it is your job to sit at a desk 8+ hours a day) do it:
Reference:
https://www.genelec.com/s360a#section-technical-specifications
@bikinpunk has got some data as well with what the prosumer monitors do it in regards to EQ when using them on a desk/console:
https://www.erinsaudiocorner.com/loudspeakers/kali_in-5/