“Let our rigorous testing and reviews be your guidelines to A/V equipment – not marketing slogans”
Facebook Youtube Twitter instagram pinterest

Hsu Research ULS-15 mk2 Subwoofer Measurements and Analysis

By

16_ULS_mk2_outdoor_testing.jpg

Outdoor ground plane testing of the Hsu ULS-15 mk2 

The Hsu ULS-15 mk2 was tested using ground plane measurements with microphone at a 2 meter distance in an open setting with well over 100 feet from the nearest large structure. The sub was tested with woofer facing the mic. The subwoofer’s volume knob was set to maximum, mode was set to EQ2, Q control was set to 0.7, and the low-pass filter was switched off. Weather for testing was recorded at 74° F with humidity at 50%.

17_ULS_mk2_frequency_response.jpg 

Hsu ULS-15 mk2 Frequency Response 

Looking at the frequency response, we can see that the ULS-15 mk2 has an extraordinarily flat response. The website product page for the mk2 states a 20-200 Hz, +/- 1dB, and this is essentially correct, at least for the EQ1 mode; this is an especially remarkable response for the fact that it accomplished with an analog amplifier. The EQ1 response uses a significant boost in deep frequencies to achieve such an impeccable response, and we can see a 12 dB lift centered at 23 Hz over the EQ2 response. It then falls off at a very steep rate, presumably to protect the driver from over-excursion. The EQ1 curve mimics that of a ported subwoofer, and, while this can be advantageous in some situations, it can present some problems in other situations. The subwoofer can only sustain that response as long as it has enough linear throw, and, as you turn up the volume, it can run out of excursion very quickly in deep frequencies, so this EQ1 curve is only useful up to a certain point. After that point, the driver is pushed past its linear excursion, and heavy distortion starts to set in, so the EQ1 mode is not where you want the subwoofer set if you intend to listen to content with deep bass at loud levels.

The EQ2 response more closely resembles that of a traditional sealed subwoofer and is also quite flat, achieving a +/- 1 dB window from 40 Hz to 200 Hz. I would recommend this mode for small rooms since room gain will naturally shore up the deep frequency response. I also recommend this mode for those occasions when the sub will be played loud with deep frequencies present in the material, since the driver is more restricted in deep bass, which will allow for more headroom. The response of either mode stretches up to nearly 200 Hz, allowing for a high crossover point if desired. This can be handy for those who want more dynamic range than can be had from their main speakers in mid and upper bass frequencies above the conventional 80 Hz crossover point, although it is recommended to have multiple subwoofers to do this in order to avoid localization of the subwoofer. Personally, I thought the system sounded much fuller with the use of a 100 Hz crossover, since the ULS-15 mk2 is so much more potent in that area than my tower speakers. Very few speakers, even large towers, will have mid bass dynamics of the ULS-15 mk2.

 18_ULS_mk2_Q_control.jpg 

Hsu ULS-15 mk2 Q Control Effect on Frequency Response 

The measured effect of the Q control is not dramatic but it does change the response. By de-emphasizing deep frequencies, bass can be made to sound tighter, since deep bass isn’t as readily perceived as higher frequencies. The Q control may also be helpful in adjusting the amount of room gain to desired levels.

19_ULS_mk2_crossover_FR.jpg 

Hsu ULS-15 mk2 Internal crossover response effects 

 

The internal crossover is based on a simple Linkwitz-Riley filter and uses a 24 dB/octave slope. Internal crossovers in subwoofers still have their place in systems without bass management, such as systems using integrated amps, or simple computer audio systems that use the computer as the pre-amp for active speakers. Since the crossover in the ULS-15 mk2 seems to affect the gain, level matching will have to be done after the internal crossover is set. Note the ULS-15 mk2 does lose its ruler-flat response with the crossover engaged in EQ2 mode.

Hsu ULS-15 mk2 CEA2010 Maximum Clean Output Measurements, referenced to 2 meters ground plane RMS
Test Frequency (Hz) Max Passing Measurement (dB) Total Harmonic Distortion (%) Harmonic Threshold Limiting
10 No Passing Result

12.5 90.7 15.0 3rd, 9th Harmonic
16 95.0 16.1 3rd Harmonic
20 99.4 18.7 3rd Harmonic
25 104.7 19.2 3rd Harmonic
31.5 110.4 19.6 3rd Harmonic
40 115.3 9.0
50 117.8 13.2
63 118.2 13.1
80 118.1 15.3
100 117.7 13.5
125 117.7 5.3

 

The above CEA-2010 measurements show the subwoofer’s clean peak SPL before heavy distortion sets in. Our measurements have been referenced to 2 meter RMS, which is 9 dB down from the standard requirement for the measurements to be shown at 1 meter peak. However most publically available CEA-2010 measurements are shown at 2 meter RMS, so we followed that convention.

The ULS-15 mk2 is able to post some excellent measurements. They are fairly close to Hsu’s own measurements. Hsu’s own measurements should be regarded as conservative, since their own tests are conducted near a freeway thereby raising their noise floor. Of note is the passing measurement at 12.5 Hz; 90 dB at such a low frequency is no small task, especially with only 15% THD (of 2nd - 5th harmonics). Also notable is the tremendous mid bass headroom, averaging around 118 dB at 50 Hz and above. It is no wonder Hsu has discontinued their mid-bass module, the MBM-12; the ULS-15 mk2 has substantially more mid-bass power than it. The third harmonic is the chief offender in nearly every measurement, and it is what prevents the subwoofer from achieving even higher SPL readings at frequencies where the amplifier has enough juice to push the driver out of its comfort zone. However, with a driver this formidable, its ‘comfort zone’ occupies a fairly wide dynamic range, and significant distortion does not set in until relatively loud levels. 

20_CEA2010_FFT_10-20.jpg      21_CEA2010_FFT_25-50.jpg
22_CEA2010_FFT_63-125.jpg

Frequency Breakdown of CEA-2010 Burst Measurements for the ULS-15 mk

The above graphs show the frequency spectrum of the increasing CEA-2010 bursts as reproduced by the ULS-15 mk2. Essentially, it depicts the behavior of the subwoofer reproducing short burst tones at successively louder levels, with each test tone raised by boosting the input gain by 1 dB until either no more output was to be had from the subwoofer or the subwoofer started making too many mechanical stress noises for the tests to be safely continued. The frequency marked above the graphs note the fundamental tone being tested, and this can also usually (but not always) be discerned in the graphs by the horizontal axis frequency point of the “main ridge,” the highest levels on the vertical axis. The noise below the fundamental (that random spikiness to the left of the main ridge) should be ignored. What should be looked at are the smaller ridges to the right of the fundamental. These are the distortion products of the fundamental, and it is here where we see how cleanly the subwoofer handles a given output level. These are mostly harmonics: whole number multiples of the fundamental.

Overall, the ULS-15 mk2 is a major improvement over the mk1, despite costing much less.

Some noteworthy features emerge from these graphs. We can see that there are levels of 10 Hz output that are not completely swamped by harmonic distortion, so the ULS-15 mk2 does permit some relatively clean 10 Hz output, to a point. This is even more true of 12.5 Hz, where there is relatively clean output until just past 90 dB, where the third harmonic rapidly increases. When pushed to large excursions, the driver ends up producing a lot of third order harmonic distortion. Since higher frequencies don’t need as much throw, we see much less distortion at nominal levels as the test frequency rises. One characteristic that can be seen are the different ways 2nd and 3rd harmonics rise. 2nd order harmonics maintain a more consistent increase that almost correlates with the increase in gain in the fundamental. The 3rd harmonic tends to erupt after a point, and this would be an indicator that the driver is reaching the end of its linear excursion.

It should be said here that anyone looking at all of these ridges of distortion and drawing the conclusion that this is somehow a highly-distorted sound or poor performance showing would be wrong. For the vast majority of the dynamic range of the ULS-15 mk2, distortion is inaudible. Until the loudest levels, there is a very large power delta between the fundamental and distortion products. Just because distortion products can be seen in these graphs does not mean it can be heard. Most of this distortion will be totally masked by the fundamental; for those readers who want to know more about the audibility of distortion at bass frequencies, we point you to this Audioholics article

23_ULS_mk2_long-term_output.jpg 

Hsu ULS-15 mk2 Long-Term Output Compression

Testing for long-term output compression was done by first conducting a sweep tone where 50 Hz hit 90 dB, and then we conduct further sweeps by raising the gain by 5 dB until no more output could be rung out of the subwoofer. As with the CEA-2010 measurements, the long-term measurements show the ULS-15 mk2 is a force to be reckoned with, roughly averaging 115 dB from 40 Hz to 150 Hz. We do see the natural response compress a bit at the highest two sweeps around 20 to 30 Hz and also from 200 to 400 Hz. The slope of the band of compression would seem to suggest that pushing the gain even more would have resulted in more output at lower frequencies but not anywhere else, and this is also hinted by the 12.5 Hz CEA-2010 burst measurement, where the angle of that compressed slope would have intersected 12 Hz at around 90 dB which correlates with the 90.7 dB CEA-2010 measurement at 12.5 Hz.

THD Graph ULS Mk2 Corrected

Total Harmonic Distortion of the ULS-15 mk2 as a percentage 

The total harmonic distortion measurements in the above graphs correspond to the long-term compression sweeps; they are the harmonic distortion measured during those sweeps. The distortion is presented here as a percentage of the fundamental, or how much of the sound produced by the subwoofer is distortion versus intended test tone reproduction. The graphs show the ULS-15 mk2 to be a very clean subwoofer until pushed to the very edge of its performance envelope, where a spike of distortion erupts from 15 Hz to 25 Hz. Distortion subsides below 15 Hz due to a filter that gives the low end a parabolic slope. The 15-25 Hz spike of distortion is the woofer over-driving the filter, and this can also be seen in the long-term compression sweeps, where the shape of the response changes from curved to flat in that frequency band at that output level. The distortion crest in the deepest frequencies in the lower power 106 dB and 111 dB sweeps is partly caused a lower signal-to-noise ratio at those frequencies and output levels; in more ideal testing conditions it would be lower like the same frequency range in the 115 dB and 116 dB sweeps. Overall a picture emerges of a highly linear subwoofer, especially ideal for music since it can pound out an average of 115 dB output levels from 40 Hz to 150 Hz with less than 10% THD. If you like your bass clean and loud, the ULS-15 mk2 is a solid choice.

25_ULS_mk2_HD_2nd-3rd.jpg     26_ULS_mk2_HD_4rth-5th.jpg
27_ULS_mk2_HD_6th-7th.jpg    28_ULS_mk2_HD_8th-9th.jpg

Hsu ULS-15 mk2 Harmonic Distortion Components as Percentage

The above graphs break down the composition of the harmonic distortion into individual harmonics. The most notable feature that can be seen is that the dominant distortion is the 3rd harmonic. Odd-order harmonics such as the third, fifth, and so on are typically the result of a symmetric non-linearity that affects both directions of a driver’s throw, so it stands to reason that the heavy spike in third-order distortion is due to the driver running past linear excursion. The shape of the low-end filter prevented the driver from over-extending below 15 Hz in our testing, so the spike of distortion ended there, but if we had pushed it even harder at low frequencies, we could probably have induced more distortion below that point.

29_ULS_mk2_group_delay.jpg 

Hsu ULS-15 mk2 Group Delay

Group delay is the measurement of how much time it takes for individual frequency bands of an input signal to be produced by the speaker. It can indicate that some frequency components are developing slower than others or are taking longer to decay. It is generally thought that 1.5 sound cycles are needed for group delay to be audible at bass frequencies, although there is an argument that group delay should remain under 20 ms to be completely unnoticeable, but that is likely meant for mid and upper bass frequencies. Either way, the ULS-15 mk2 presents nothing to worry about. Group delay stays under 20 ms down to around 30 hz, a frequency low enough that it is certain that group delay would not be discernable at or below it. The rise in group delay as frequencies descend hints at the subsonic filter guarding the low end of the response, but there is nothing measured here that would come close to being audible. Altogether, this is a good showing.

How much of an upgrade is the ULS-15 mk2 over the ULS-15 mk1?

30_ULS_mk1_mk2.jpg 

ULS-15 mk1 (left), ULS-15 mk2 (right) 

Since I happened to have a ULS-15 mk1 on hand, I decided to hold a mini-shootout between the mk1 and mk2 to see how much of an improvement, if any, was to be had. On the base specifications, I would not have predicted much of an improvement. The mk1 was a much more expensive subwoofer at $1099 and weighed a good deal more. Most of the weight difference occurred in the driver, where one would expect the additional weight to add up to a performance advantage; the mk1 driver weighs 41 lbs and the mk2 driver weighs 22 lbs. The extra weight of the mk1 seems to come from the massive top plate which is needed for strength and long stroke of the XBL^2 design.

31_ULS_mk1_mk2_drivers.jpg 

ULS-15 mk2 driver (left), ULS-15 mk1 driver (right) 

So how does the mk2 fare against the mk1?

32_ULS_mk1_CEA2010.jpg

Hsu ULS-15 mk1 CEA-2010 measurements and difference between the mk2’s measurements 

From 20 Hz to 40 Hz we see the two models are very close to each other in performance, but at 50 Hz and above the mk2 pulls away, as the mk1’s output begins to drop as it enters mid bass frequencies. At 63 Hz and above, the mk2 is more than fully doubling the mk1’s output. This is perhaps due to the split gap in the top plate that is characteristic of the mk1’s XBL^2 design which may lessen sensitivity around the rest position of the voice coil. This would reduce output as the frequencies increase, since less stroke is needed as frequency increases, so the voice coil isn’t moving out as far from its rest position. However, this does not explain the difference at 16 Hz. Furthermore, what isn’t shown is the difference at 10 Hz and 12.5 Hz, where the mk1 was not able to post a passing measurement. In fact, at these deep frequencies, the mk1 produced substantially more distortion than the mk2 for the same output level. This ran contrary to my expectation that the mk1 XBL^2 design would exhibit less distortion than the overhung mk2 at very high excursions.

Overall, the ULS-15 mk2 is a major improvement over the mk1, despite costing much less. It has much more composure at very deep frequencies and far more headroom at higher frequencies. This isn’t to say the original ULS-15 was a bad sub- far from it. Its 20 Hz to 50 Hz performance still rivals and exceeds more expensive sealed subwoofers. The mk2 slightly improved on all of its strengths and massively improved on its weaknesses.

 

Confused about what AV Gear to buy or how to set it up? Join our Exclusive Audioholics E-Book Membership Program!

Recent Forum Posts:

Epsonfan posts on September 06, 2020 13:18
Too many commie China loving libtards here who sold their souls to communist China just like Obiden , NBA and Hollywood.
BRAC posts on May 27, 2020 14:20
Awe
shadyJ, post: 1393701, member: 20472
Both are terrific subs. I think the Hsu subs are a bit more capable in deep bass and the Dynamo subs are a bit more capable in mid-bass. The Dynamo subs have a more sophisticated feature set and amplifier but the Hsu sub has a bit more robust driver. I would say that if you are connecting the sub to a system that doesn't have much in the way of bass management, go for the Dynamo since it has far more ways to control the performance. If you are hooking the sub up to a processor that does have a lot of bass management features, the Hsu sub may make more sense since it offers a bit more raw performance overall. But the Dynamo sub isn't that far from the ULS-15's performance.
Awesome. Really appreciate the help. I’m actually one of the few that prefers mid bass over deep bass for movies. Deep bass excites WAY too many rattles and vibrations in my room. Can be extremely distracting while watching a movie. Love that mid bass punch in the chest!
shadyJ posts on May 27, 2020 14:03
BRAC, post: 1393684, member: 91832
Need some help…I’m torn between a dual drive of these guys or 2x ML Dynamo 1600x. Both received great reviews here on the site. Which is the better performer for home theater?

Btw, I’m in Canada and can get the ML’s locally. The HSU’s would have to be imported. Cost works out roughly the same. Thanks!
Both are terrific subs. I think the Hsu subs are a bit more capable in deep bass and the Dynamo subs are a bit more capable in mid-bass. The Dynamo subs have a more sophisticated feature set and amplifier but the Hsu sub has a bit more robust driver. I would say that if you are connecting the sub to a system that doesn't have much in the way of bass management, go for the Dynamo since it has far more ways to control the performance. If you are hooking the sub up to a processor that does have a lot of bass management features, the Hsu sub may make more sense since it offers a bit more raw performance overall. But the Dynamo sub isn't that far from the ULS-15's performance.
BRAC posts on May 27, 2020 12:59
Need some help…I’m torn between a dual drive of these guys or 2x ML Dynamo 1600x. Both received great reviews here on the site. Which is the better performer for home theater?

Btw, I’m in Canada and can get the ML’s locally. The HSU’s would have to be imported. Cost works out roughly the same. Thanks!
ematthews posts on March 25, 2020 21:01
Matthew J Poes, post: 1377438, member: 85392
I know this is just feeding trolls here, but most don't understand, there is no such thing as made domestically. Nothing in our CE world can be made 100% domestically.

Only a very tiny fraction of amplifiers are made in America using american made modules. The amplifier circuits inside are still usually built in China and shipped as assembled modules. Those that aren't use “beans” made in China. It's impossible to do anything else. Those that are made mostly domestically (where only the beans are sourced from China) cost 1000's of dollars. Nothing most can afford.

I can't name names, I will probably get myself in trouble, but MANY companies that claim american made are not. I don't consider having all the guys fully assembled in China and shipped to the USA so they can be stuck inside a Chinese made box american made.

Those who think this kind of manufacturing is coming back to the US are nuts. It was never in the US. We've never been able to make those kinds of electronics in the US, by the time we shipped all that kind of production to China (70's) we gave up the ability, failed to progress, and left all the manufacturing development and infrastructure to China. We don't have the ability to create those kinds of facilities here and lack the workforce to man them. When I've asked about it (cost not withstanding) I've been told it would take a generation to even build a modicum of what China has. If a Foxconn built a facility to manufacture receivers, tablets, laptops, etc. in the USA and was totally vertically integrated, they would have to man it about 80% with Chinese workers. They would then need to work with the US education system to fundamentally change tech education to prepare folks to be able to work on the assembly lines, operate the machines, etc. It isn't that Americans are uneducated or stupid, quite the opposite. It's better described as wrongly educated. What we prepare people for isn't that kind of work. It's all relatively low paying, and would be here too. Whose going to get a tech degree to make $15 an hour.

That isn't america hating, that is just a fact of life. Reagan opened the door to trade with China and in that deal we shifted the manufacturing of consumer electronics from the USA to China. They ran with it and today have built the entire giant country into a huge powerhouse of CE production and logistics.

Best explanation I have read on this subject. Thanks for the good reply.
Post Reply