December 21, 2019


Rythmik Audio Subwoofers

Custom Install products & Do It Yourself (DIY) issues

Advice on choosing a subwoofer

Sound Quality

Other issues

Rythmik Audio Subwoofers

Is Direct Servo stable?

When it comes to DIY servo subs, the first concern that comes to mind is the stability. Servo subs at one point had bad press because too many poorly executed projects were out there. These were too often done on a trial-and-error basis, without considering the control system or theories behind them. We can assure everyone that Direct servo system is absolutely stable.

A few of the features of the Direct servo system can further demonstrate that. First, the system is stable even when the amp is in clipping. Most accelerometer-based systems are not stable when the amp clips. So the protection circuit (called limiter) in those systems is a necessity, rather than a value added feature. In class AB and class H amplifiers XLR models, we add the switch to bypass limiter circuit, because such circuitry reduces linear dynamic range. Second, there is no relay in our amps. Some servo system cannot work without relays of instability during power-on. The power on noise of our system is just like that of a regular non-servo system.

One note about limiter. In class D amplifiers, the limiter is necessary because the clipping waveforms from class D amplifiers contain high frequency spectral components that may sound like voice coil rubbing. It can be confused with actual voice coil rubbing which by all means should be avoided. As a result, we suggest the limiter to be kept ON for these models. Furthermore, in our class D based amplifiers, the indicator will flash when the amplifeir is close to clipping. It allows the user cut down bass extension or volume to avoid non-linear dynamic range.


Do servo subs need a Linkwitz Transform?

Servo subs in general do not need Linkwitz Transform circuits. Our Direct servo subs don’t. The Direct servo subs are closed loop systems where a servo signal that monitors the movement of driver’s cone is fed back to the amplifier to achieve lower distortion, as well as a more coherent sound over a wide range of working conditions. LT-based subs are open loop systems. Flat response is achieved by predicting the transfer function between the amplifier output and the speaker output and pre-equalizing the amplifier output with the inverse of that transfer function. However, as the driver parameters change due to signal level, break-in, and fluctuating voice coil temperature, the actual transfer function between amplifier output and speaker output would also change. When that occurs, the output from the subwoofer will no longer be flat.


How is the frequency response measured?

We measure all subwoofer frequency response outdoors, using the near-field technique. For vented subs, measurements from the woofer and vented are taken separately. These two measurements are then summed after the levels of both measurements are adjusted so that they have same output at 1.4 times the enclosure/port resonance frequency.

measuring a driver


What is the damping factor between the driver and amplifier in Direct Servo subs?

Close to infinity. Keep in mind damping factor is not the same as the damping of bass extension. More accurately stated, damping factor is the ratio of speaker impedance vs amplifier output impedance. For most amplifiers, damping factor ranges from 20 (tube amplifiers) to 200 or even higher. In the past, higher values are preferred because they imply the frequency response of the amplifier is less sensitive to the speaker’s impedance. In a conventional (or non-servo) amplifier-speaker interface however, higher damping factor does not translate to real-world benefits because it does not include the impedance of speaker wires, or even the DC resistance of the voice coil for that matter. One should keep in mind, the voice coil DC resistance is not constant, rather it is temperature-dependent. A high power subwoofer driver means the voice coil can tolerate higher voice coil temperature. What is not often mentioned is that when the voice coil is at a higher temperature, the DC resistance also increases as a result. Copper wire has a temperature coefficient of 0.4% per degree C. A 125C (or 225F) temperature increase causes a 50% increase of the voice coil DC resistance, which in turn causes the Q value of the speaker response to change by a similar amount. In a Direct servo sub, the subwoofer response is completely independent of the voice coil temperature. Or to be more precise, the resistance on the path from amplifier output to speaker, including resistance of speaker wire, connectors if any, and voice coil resistance, does not change the frequency response at all. As a result, its damping factor is infinite, even by the most stringent definition.


Do I really need a flat response down to 14 Hz for music?

The simple answer is yes. The more accurate, more complicated answer is that it depends on the music you are listening to. On dance mixes and music with extremely high bass contents, one does not need 14 Hz extension because of what is called the “masking effect”. The masking effect occurs when a high amplitude signal is present simultaneously with other smaller amplitude signals. The perceived loudness of the smaller amplitude signal is drastically reduced. In addition, human ear’s sensitivity decreases in the lower frequencies. Therefore for this type of music, one may even want to set the extension setting higher with lower damping factor setting (in our A350/A370 amps). That will make the bass sound clean (by removing ambient information) and punchy. It essentially allocates more amplifier power to the signal that we hear and care about most. On the other hand, when no such signals are present (for example, in jazz, solo and other types of music) and masking effect is not strong, one can clearly hear the difference that a 14 Hz extension makes (vs say 28 Hz extension). It is not a difference that will have you jumping out of your seat say “there’s the bass”. Rather it is an improvement in two areas: ambience and dynamics.

Ambience: good recordings normally capture a lot of ambience information which is predominantly low frequency signals. It is generally agreed that the lower the sub plays, the deeper, taller, and wider the sound stage becomes (or one may say open and spacious).

Dynamics: during musical transitions, a lot of the non-harmonic content is produced. Most of this content is in the very low frequency band, which will be filtered out if the bass extension is not low enough.


How does Direct Servo achieve a flat response in so many different configurations? It sounds too good to be true.

Direct servo can be thought of as an adaptive equalization system. To be more specific, it performs a transformation on the physical T/S parameters into a set of virtual T/S parameters. By doing so, the amplifier adjusts its output so that the frequency response behaves as if the driver possesses these virtual T/S parameters. As a result, the Direct servo can be analyzed just like normal drivers and the technology can be applied to any known subwoofer configuration. All one needs to do is change the feedback network.


What is the amount of equalization present in Direct Servo?

The amount of equalization present in the sealed box configuration can be as high as 10 dB at 20 Hz if the extension is set to 14 Hz. That is also why we use a high power amp and driver so that it has enough headroom for equalization.


Why do non servo subwoofers seem to have a higher crossover frequency than servo subwoofers?

Our Direct Servo subwoofers utilize a frequency region that voice coil inductance has the least effect in impairing the cone control. That is the reason we limit our crossover to frequencies lower than non servo subwoofer. It is possible to achieve a crossover as high as non servo subwoofers, but performance is compromised as the effectiveness of the servo is reduced.

In practical use, many subwoofers lack adequate accuracy to be crossed as high as ours. Often one needs to cross low in order to “minimize the damage”, since most subwoofers have inferior sound quality to the bass of the main speakers.


Does the Direct Servo sensor make the sound slower?

Ask any one of our many satisfied customers and they will tell you emphatically “definitely not!” If by “fast bass” you mean more true to the original, then quite the opposite is true.

Our servo is based on sensing coil, and it does not have any additional mechanical assembly (unlike the accelerometer-based design, which is a mass attached to a pressure sensitive material to sense the inertia of mass acting on the material). There is no mechanical delay in our voice coil assembly. Also the location of the sensing coil is “co-centered” with the driver coil, immersed in the same magnetic gap. The moment the driver coils starts to move, the sensing coil will detect that and start reacting. In other words, the distance between sensing coil and driver coil is “zero”. This is very different to accelerometer-based approach, which place the sensor on the cone under the dust cap. There is a finite distance between the voice coil and the accelerometer, and as a result there is a response time. It is the same with another approach used in Germany that put the sensing coil in its own magnet gap. We don’t have these problems.

If the word “slow” is used with reference to the extra length on the voice coil, that is even less understandable. The voice coil is used both for driving and sensing. If the voice coil is slow for sensing, it will also be slow for the for driver coil? Does that mean all voice coil based speakers will be slow? The wire length used in voice coil is probably shorter than the transformer in a tube amp. Does that mean the tube amps will sound even slower? Better yet, get rid of the inductors in crossover networks.


What don’t we sell passive speakers or non-servo amplifiers?

The musicality of our subs come from the synergy of the amplifier and the driver operating under servo control. If you have only our amp, or just our driver, the synergistic effect is lost. That is why Rythmik is very unique and why we don’t bother getting into the component business.

Our subs sound very musical because they have very low memory effects. All passive drivers driven with non-servo amplifiers exhibit three major drawbacks: (1) the temperature of the voice coil changes the frequency response, (2) the hysteresis memory effect of the spider reduces the clarity of the sound, and (3) because the spider and surround are mechanical parts, the impact to frequency response changes both in short term and long term. Our servo operation improves all of the above and it achieves a level of sound quality beyond what the same driver and amplifier combination can achieve without servo control. Consequently, we prefer to sell customers the driver and servo amplifier combination which allows those components to reach their full potential.


How does push-pull technology compare to servo technology?

We still think our servo subwoofers offer more advantages:

  1. The idea of push-pull is good. But the driver with reverse cone has smaller cone surface (because the voice coil is exposed), so the distortion cancellation is not as good in practice. In practice, we found just 2x or 3x reduction on even order harmonic distortions. In order to get best distortion reduction, the driver with reverse cone needs to have 20% more excursion. In addition, push-pull only reduces even order harmonic distortions, which are more forgiving to start with. Odd order harmonic distortions are more detrimental and push-pull cannot address those. Our servo addresses all order of harmonic distortion. So it is more effective. There are other types of distortion that servo addresses while push-pull does not. For instance, the distortion caused by voice coil temperature and other mechanical nonlinearity that causes variation on amplitude response variation. When the cone puts less than what it is supposed to provide, the servo feedback will enable the amplifier to provide more power instantaneously. The opposite is also true (to stop standing wave or when the cone moves too much). In this case, the servo can even make the amplifier generate negative voltage, not just 0 voltage, to provide better braking power.The net result is you hear a much more dynamic and clearer sound.
  2. How does an 8″ driver achieve 20hz extension? It is through equalization. If one works with manufacture ring, then he will know there is a so-called unit-to-unit variation (some call it tolerance). The EQ is set up based on a “typical” driver model. So you may get a unit that deviates from this “typical” driver model and the EQ will not be ideal. In addition to this unit to unit variation, we also have batch to batch variation, and year over year variation (aging) and in the presence of all these variations, it becomes clear it is difficult use “one size fits all” approach. Servo is more advantageous because servo takes cone velocity into the feedback and it address all these 3 types of variations on the fly. This unit to unit variation problem feeds back to the first problem where a perfect cancellation can be even problematic in the presence of unit to unit variation.