Feed Forward
Advanced Distortion Supression Technology
Feed Forward™ Distortion Suppression Technology
No matter how big or how powerful a powered subwoofer may be it is always possible to push it beyond its limits to the point where it becomes distorted, makes objectionable noises or even damages itself. It is always a challenge when designing a powered subwoofer to come up with a method of limiting the maximum output in such a way that avoids excess distortion or damage but preserves as much of the dynamic capability and maximum output capacity of the subwoofer as possible. Distortion and, eventually, damage to a subwoofer comes from one of three major sources:
- Too much power from the amplifier causes over heating of the voice-coil, higher distortion and ultimately failure of the woofer driver.
- Too little power available from the amplifier causes high distortion in the signal fed to the woofer driver, generates excess heat and may eventually burn out the amplifier.
- Excessive movement of the driver cone caused by over driving at low frequencies leads first to high distortion, then to objectionable mechanical noises and finally to mechanical damage to the woofer driver as it literally tears itself apart.
Usually powered subwoofers are designed with amplifiers that have more than enough power to over drive the woofers mechanical system. So, #2 above is normally not a problem. Advances in voice-coil technology have also made voice-coil burnout much less likely and it is common practice to match woofer drivers and amplifiers to make burnout very unlikely. That leaves #3 as, by far, the most challenging problem. Mechanical distortion and damage can occur almost instantaneously regardless of how well matched the woofer driver and amplifier may be and may be triggered by any number of complex signal combinations which can be very difficult to predict. As a result, most attempts to protect the system or reduce distortion limit the output of the woofer far below it’s theoretical maximum output in an effort to guarantee a safe level of operation.
The difficulty comes from the fact that even though a woofer may have flat frequency response it takes different amounts of power from the amplifier to produce the same amount of sound at different frequencies. This means that combinations of signals at different frequencies may overload a woofer while the same signal level at a single frequency would be okay. For example, let’s consider a hypothetical 10” powered subwoofer in a 2.5 cu ft. ported box. Figure 3 shows the frequency response and impedance of the system. As you can see, the frequency response (Red) is quite flat to below 30Hz, but the impedance is far from flat. It rises from around 4 ohms at 30Hz to 36 ohms at 20Hz and 60 ohms at 50Hz. (please note that the impedance curve has been offset upward by 40 Ohms for clarity) That means this woofer will draw only 1/9th as much power from the amplifier at 20Hz compared to 30Hz and only 1/15th the power at 50Hz. So, if the amplifier has enough power to drive the system to maximum output at 30Hz it will have way too much for the system to handle at 20Hz and 50Hz.
Figure 4 shows the maximum acoustic output of our hypothetical woofer versus frequency assuming a 500 watt amplifier. As you can see, the maximum output at 30Hz is about 6dB greater than at 40Hz to 50Hz and 25dB greater than max output at 20Hz. Broad band limiting would dictate that we roll off the low frequency response of the Figure 3–Frequency Response (Red) and Impedance (Blue) for Figure 4–Max. hypothetical subwoofer acoustic output at 10’ with 500 watt amp amplifier below 30Hz to protect the woofer at the lowest frequencies even though the woofer can deliver significant low bass output (See Figure 3). We would also have to limit the amplifier to a safe level well below the max output in the 40Hz to 50Hz range. So, at normal listening levels we would lose all of the low bass capability of the woofer and at high levels we would lose at least 6db of dynamic range since the limiter setting is determined by the frequency range with the lowest maximum output. It is no surprise that this form of compression or limiting tends to make a subwoofer sound muffled, lacking in extension and lacking in punch.
Another commonly used method is to apply limiting or correction based on the actual movement of the woofer driver cone as measured by a position sensor. These are known as servo feedback systems. The position sensor is coupled to circuitry that compares the input signal to the cone movement and provides a correction signal to the amplifier input signal. In theory the correction signal modifies the amplifier output signal in a way that will correct for any mechanical distortion in the acoustic output. And, in practice servo feedback systems often produce excellent test results with steady state test tones. However, real program material is dynamic and the servo system cannot correct for anything until after something has already gone wrong. So, the more heavy handed the servo correction algorithms the more sluggish the system sounds as it tries to measure, compare and correct in a dynamic signal environment. In addition, the correction signal may go the wrong direction by asking the woofer to correct itself by moving beyond its mechanical limits. Therefore, broad band limiting is still required to protect the woofer driver against correction signals that might seriously overload the system. So, although a servo system can produce impressively low harmonic distortion numbers for steady state test tones, it’s ability to handle the dynamics of real world audio signals can be seriously compromised due to the fact that it is correcting for something that has already gone wrong and, of course, still requires the broad band limiting we discussed above.
From a music or movies performance point of view the best limiting or protection system would be none at all. Obviously we can’t take that risk. But, second best would be a system that limits only that which absolutely needs to be limited and which does its job proactively rather than reactively. Polk’s proprietary Feed Forward distortion suppression system anticipates problems and fixes them before they happen. It uses our ability to predict how the woofer will respond to complex signals to respond with the absolute minimum intervention to minimize distortion and chance of damage without compromising dynamics or transient response and while actually increasing maximum output. In essence the digital processors of the DSWPRO subwoofers are programmed with a mathematical model that knows in advance how the woofer acoustic system will respond to the output signal from the amplifier. The processor monitors the amplifier input for signal combinations that would cause distortion or damage and alters only that part of the signal required to maintain low distortion and safe operation. It does this before the signal ever gets to the woofer driver so there is no feedback correction time delay and no need for broad band limiting. Also, since Feed Forward is proactive rather than reactive, the limits can be set much higher and closer to the theoretical maximum output of the woofer at each frequency without risk of excessive distortion or damage. For example, in our hypothetical woofer of Figure 4 the Polk Feed Forward system would know in advance that the output capability is lower at 50Hz. Even if limiting at 50Hz were required, Feed Forward would still preserve full dynamics above and below that frequency. It would also preserve low bass capabilities at normal listening levels while providing appropriate protection from low frequency mechanical damage at higher levels. As a result, when the program material calls for greater dynamics Feed Forward allows the system to deliver without an artificial cutoff. Unlike other distortion reduction or protection systems Feed Forward is virtually undetectable in operation. It delivers all of the excitement and impact or subtlety of the program material, as the case may be, while maintaining lower distortion and higher maximum output levels.
