I've been digging into papers by Thiele/Small/Keele/Snyder et al and trying to absorb as much of the physics of bass reproduction as I can. I'm getting there.
My best resource for driver info remains "loudspeakerdatabase.com". More comprehensive recommendations would be welcome.
My question: how and why does Tang Band seemingly outperform the competition when it comes to small drivers for bass reproduction in small enclosures? In general, the main tradeoff seems to be max SPL, as they tend to have lower sensitivity and lower power handling capability than some of their competitors. Still, it seems as if their products offer a significant jump in bass capability compared to the competition. At the very least, they're "different".
Are Tang Band specs reliable? Does it all come down to a sacrifice of maximum SPL versus low frequency extension? Are other characteristics (distortion, etc.) lacking? Am I missing something?
My best resource for driver info remains "loudspeakerdatabase.com". More comprehensive recommendations would be welcome.
My question: how and why does Tang Band seemingly outperform the competition when it comes to small drivers for bass reproduction in small enclosures? In general, the main tradeoff seems to be max SPL, as they tend to have lower sensitivity and lower power handling capability than some of their competitors. Still, it seems as if their products offer a significant jump in bass capability compared to the competition. At the very least, they're "different".
Are Tang Band specs reliable? Does it all come down to a sacrifice of maximum SPL versus low frequency extension? Are other characteristics (distortion, etc.) lacking? Am I missing something?
First off, the lower cut-off frequency as shown by WinISD in a classic alignment (what WinISD calculates by default) is not relevant any more. DSP's are cheap and they can change the frequency response to whatever you want. Now it is about the max SPL and distortion.
Tang Band woofers mainly differ from others in the fact that they are capable of longer excursions. I guess there is little market demand for small (2 to 6") long excursion woofers, which explains why competitors do not offer similar woofers. For example, their long excursion woofers W5-1138 and W6-1139 either require a big box for the vent to physically fit (then why not pick a larger, lower excursion woofer?) or passive radiators (expensive).
Long excursion woofers can require a stiff suspension to keep the cone from rocking. This results in a higher fs than you would except. So there might be woofers that do not seem to dig deep, but actually are designed to do so.
Tang Band woofers mainly differ from others in the fact that they are capable of longer excursions. I guess there is little market demand for small (2 to 6") long excursion woofers, which explains why competitors do not offer similar woofers. For example, their long excursion woofers W5-1138 and W6-1139 either require a big box for the vent to physically fit (then why not pick a larger, lower excursion woofer?) or passive radiators (expensive).
Long excursion woofers can require a stiff suspension to keep the cone from rocking. This results in a higher fs than you would except. So there might be woofers that do not seem to dig deep, but actually are designed to do so.
Last edited:
Thanks, TBTL. Is there something about the TB woofers that precludes electronic EQ to further extend their low frequency response?
How would one go about determining linearity and distortion? Buy, try and measure?
How would one go about determining linearity and distortion? Buy, try and measure?
Not specifically for this brand. Power handling and cone excursion limit (which sometimes is larger than xmax) must not be exceeded. The power handling specification is the long term power handling, short term power handling is larger.
Yes, that is the only way. Most manufacturers do not publish distortion data. Sometimes others have measured them. There can be hints for low distortion in the datasheet, like 'shorting rings' or 'copper cap'. Still, you never know how it really performs.
Last edited:
That is not true. Almost none of it. First off, passive speakers are still a thing and especally in the DIY community still, by far, the most of the developed ones. Yes, DSPs became cheap and they slowly get a bigger piece of the cake. But even in the DIY magazines, the vast majority is passive or a passive sat/top and a subwoofer. And no, you can't change the frequency response to whatever you want since the drivers and the enclosure still has to fit together and physical limit are given, 20Hz will hardly be audible from a 2" driver, it still doesn't make sense. You still have to simulate the speakers capabilities excursion and power wise, be it with WinISD, Hornresp or AJ Horn or something else and these programs will also tell you the max SPL possible with excursion x at frequency y. If you ignore that, you'll get nowhere with a dsp, especally since power isn't everywhere available despite class D amps - think about mobile speakers, cell phones and other smart devices.
But you still cannot predict the exact destortion by distribution and magnitude, by a simulation. In the best case it's an estimation. And you can't 'fix' distortion by dsp. You still have to measure the distortion yourself because the materials and forces within a speaker are way too complex to simulate and even if you could do it, you won't get most informatons unless you disassemble a driver completely and analyze every bit and piece of it.
A. To get a better WAF. 😀 (or make the speaker mobile) And B. TB got several PR in their program which are not supercheap but still no luxus-class-price-tag on them. Besides that, there are a lot of drivers which also work fine in a sealed or filtered sealed enclosure (GHP). Not everything has to be BR.
That's only half of it. The suspension is the other way around. To achieve small enclosures and lower fs, the Mms is increased and the suspension has to be stiffer because of the higher moving mass and also to reduce the Vas (which is directly determined by the suspension and the cone surface), the stiff suspension rises the fs a bit again, that's true. The high mass reduces the efficiency though. Deep, loud, small - pick one or two, all three doesn't work.
Thanks for the great discussion. In the spirit of my original post, I'm curious why Tang Band seems to operate in a niche mostly unoccupied by any other manufacturer. Is there an inherent downside to their small woofers (other than the "deep/loud/small -- pick two" tradeoff)? Distortion, quality control, reliability, etc.? Or do they possess IP that prevents others from working in that space?
I find it almost impossible to believe that other manufacturers don't want to offer competitive small woofers.
I find it almost impossible to believe that other manufacturers don't want to offer competitive small woofers.
Most of them are very good. I would not buy every of them because I personally got a bit different taste.
TB is specialized for fullrange speakers. The subwoofers are a good fit for the smaller ones, be it a sat/sub combination or a FAST (fullrange an subwoofer technology), so these subwoofer drivers are a logical addition.
TB is specialized for fullrange speakers. The subwoofers are a good fit for the smaller ones, be it a sat/sub combination or a FAST (fullrange an subwoofer technology), so these subwoofer drivers are a logical addition.
No, they are not.
I have the W5-704, W5-1685, W6-789E and all measured with inferior specs. Disappointing.
Click HERE for TS.
Regards
Mike
More than the half of the listed self-measured parameters are WAY off. And there are drivers among them which I know have much closer parameters to the factory specs, like the Faital or Vifa. I don't claim the TB parameters to be correct but since so many are so much off, I suspect a methodical error (like the driver laying on the magnet, pole bore closed, equipment changed, measurement leads changed or unrelyable (croc clamps), no driver break-in etc.). Unfortunately there are none informations about the measurement enviroment.
TBH, I would not rely on these data. I wouldn't rely on some of the factory data either, though I still don't claim the TB parameters to be correct either.
But is your own measurements reliable? Small- or large signal used?
I measured tang band w4-992's some time ago and noticed the same, some values were different compared to manufacturer specs. Mostly Qms and VAS. Qts and BL was almost the same as specs says.
Fs dropped from 60Hz to 49Hz after burn-in (15minutes, 20Hz at full xmax).
Measuring TS parameters is not that simple anyway. There are different methods giving slightly different results etc..
Afterall, best would be to measure the driver anyway. Use either these or manufacturer (or borh) parameters and simulate. Then make the enclosure that you just simmed and measure it. If simulated FR corresponds to a measured FR, parameters might be correct.
I measured tang band w4-992's some time ago and noticed the same, some values were different compared to manufacturer specs. Mostly Qms and VAS. Qts and BL was almost the same as specs says.
Fs dropped from 60Hz to 49Hz after burn-in (15minutes, 20Hz at full xmax).
Measuring TS parameters is not that simple anyway. There are different methods giving slightly different results etc..
Afterall, best would be to measure the driver anyway. Use either these or manufacturer (or borh) parameters and simulate. Then make the enclosure that you just simmed and measure it. If simulated FR corresponds to a measured FR, parameters might be correct.
ICG, I agree with most of your points (and I think we do not contradict each other 😉 ).
I was talking about commercial applications for the mass market, where unit costs are everything and engineering (including DSP programming) is relatively cheap. I assume that speaker manufacturers only develop for this market and that the DIY market is a not-that-interesting secondary priority. So any driver developments are done for the former market, which does not need woofers to have a flat response without DSP*. Hence, not many speaker manufacturers design them. These woofers would only sell to the small DIY market.
For the average DIY person, the story is different indeed.
* For example, the 'JBL Xtreme' portable bluetooth speaker only reaches down to 150 Hz at its highest volume setting. At low volume levels, the bass is boosted to reach down to 70 Hz. See these measurements at about 2/5 of the page height: Oluv's Gadgets: Review: JBL Xtreme - how much bass can you handle?
I was talking about commercial applications for the mass market, where unit costs are everything and engineering (including DSP programming) is relatively cheap. I assume that speaker manufacturers only develop for this market and that the DIY market is a not-that-interesting secondary priority. So any driver developments are done for the former market, which does not need woofers to have a flat response without DSP*. Hence, not many speaker manufacturers design them. These woofers would only sell to the small DIY market.
For the average DIY person, the story is different indeed.
* For example, the 'JBL Xtreme' portable bluetooth speaker only reaches down to 150 Hz at its highest volume setting. At low volume levels, the bass is boosted to reach down to 70 Hz. See these measurements at about 2/5 of the page height: Oluv's Gadgets: Review: JBL Xtreme - how much bass can you handle?
Last edited:
In terms of FR, TB's plots are smoothed, but I find them to be reasonably close to drivers I've purchased, and I've used quite a few from them. There have also been shifted T/S parameters, but they are usually just shifted in a way that the box model does not change much from the spec sheet. I have found their Le to be WAY off at times in woofers.
That said, the Underhung woofer series 1720/1585/1721/2096 have been a known driver line for shift in parameters, and that they change as they break in.
I have a set of W8-2096, and I hope they don't change much under use. I did drive them fairly hard before I took T/S to see if they'd change much and they didn't very much. However, the specs were different unit to unit enough that the model was in question. It turned out about the same on paper. The Qms of these is VERY low; 2.1 and 2.0 respectively.
We'll see,
Wolf
That said, the Underhung woofer series 1720/1585/1721/2096 have been a known driver line for shift in parameters, and that they change as they break in.
I have a set of W8-2096, and I hope they don't change much under use. I did drive them fairly hard before I took T/S to see if they'd change much and they didn't very much. However, the specs were different unit to unit enough that the model was in question. It turned out about the same on paper. The Qms of these is VERY low; 2.1 and 2.0 respectively.
We'll see,
Wolf
From a professional point of view (so having a production line for a couple of thousand of products a year/mass market), I can tell that I can't recommend Tangband AT ALL!
Their T/S are literally all over the place through different batches and sometimes even within the same batch.
Not to talk about all the mechanical issues we had (glue layers letting loose and that kind of stuff) and specs that are far to rosy and to good to be true and inconsistent.
For a production line for high quality products all these points are unacceptable since in the end it means that the end-products will differ from each other.
So for me personally TB is an no go anymore from a professional point of view.
I guess they will be fine for hobby projects, as long as you double check and measure the T/S parameters as well as the frequency response.
Their T/S are literally all over the place through different batches and sometimes even within the same batch.
Not to talk about all the mechanical issues we had (glue layers letting loose and that kind of stuff) and specs that are far to rosy and to good to be true and inconsistent.
For a production line for high quality products all these points are unacceptable since in the end it means that the end-products will differ from each other.
So for me personally TB is an no go anymore from a professional point of view.
I guess they will be fine for hobby projects, as long as you double check and measure the T/S parameters as well as the frequency response.
any recommendations for a good 4 inch subwoofers? Needs to be ok for reflex (Qts <0.4) and xmax >3mm. Fs <60Hz.
I currently have 4x tb w4-992s and they are ok.. just looking for even better options.
I currently have 4x tb w4-992s and they are ok.. just looking for even better options.
Despite EVERYONE who does the measurements adamantly claims HIS measurements were perfectly and undoubtedly correct, they aren't reliable either. And temperature, air pressure, air moisture and general measurement errors and tolerances often lead to a lot different parameters. Even repeat measurements within the same enviroment lead to different results. Partly, because of the measurement tolerances itself, partly because some of the values are calculated and errors multiply. Check with a simulation tool if the parameters measured are consistent to each other (add up or are actually not possible). In the end, you get a ballpark number though and you can verify it at the built box. That is one of the main reasons why you should build a test enclosure or at least keep the option to be able to change the tuning afterwards, which is problematic if the port consists of a slot build of wood. In some of such cases an oscillating saw can save an enclosure from bein scrapped.
Thiele and Small parameters are always small signal parameters. Back then there were not such wide ranges of parameters like there are nowadays. With extreme (car) subwoofers with high Mms you'll often have to increase the voltage though because of the extreme losses of the stiff suspension and thick surrounds, they 'wake up' only at higher power input.
The parameters usually shift a lot while the break-in. That's also why measured parameters differ that much, some do it for a few minutes, others for a few hours, others take several days (!).
Yes, that leads to very big differences in the result, the parameters often look quite different. You cannot compare measurements from different sources directly. Some manufacturers provide correct parameters but do horrible response measurements, Eminence being a prime example there, worst response measurements in the industry and heavy smoothing, despite most of the drivers not in need of such 'tuning' of the measurements.
Yes, absolutely. It also helps a lot if you don't build the enclosure to absolute extreme, borderline from working. You can check that with your simulation, if you add 0,5 Ohm serial resistance (which is normal for contact resistances, coils at passive speakers add up on top of that) and the simulation changes dramatically, chances are your enclosure won't work with just small differences of the parameters from the datasheet and you're probably better off to go for i.e. a bit more enclosure volume or build a test enclosure.
If you can use a 5", I know a very good and cheap driver, the Omnes-Audio SW 5.01, works great in 6-7l, minimum enclosure volume for BR is ~5l, goes down to 40-ish. Was tested great in the latest issue of the German magazine Klang + Ton.
This is why LINK & LINK2.
I had first W6-1139SIF:s but immediately changed to W6-1139SI:s when I had possibility, just for the weight saving and small additional extra cabin volume. There is 2 elements both at the ends, where it has 1mm in every direction space. Would not be able to fit bigger. Very lightweight box with very big and long port inside with mild folds. Excellent drivers. and whole deal is 80cm long yet under 4kg, and plays very strong!