Before the DDD driver

Welcome to the December 2021 German Physiks newsletter.

This month we would like to talk about the Walsh omnidirectional driver. This was used in the Ohm F loudspeaker and was the inspiration for our own DDD driver. The Walsh driver was an outstanding piece of design, but its performance was held back by the limitations imposed by the materials available at the time and the fact that Walsh’s design predated the introduction of computer modelling. Without the aid of computer modelling, optimising such a complex design would have been impossible.

In a future newsletter, we will explain how the German engineer Peter Dicks, by taking advantage of more modern materials and by creating a detailed computer model of the Walsh design, was able to analyse how it worked and then develop a new design that eliminated the Walsh driver’s weaknesses - low sensitivity, fragility, complexity and the consequent high cost of manufacture - to produce the exceptional DDD driver that is used on all German Physiks loudspeakers.

The Walsh driver

The Walsh driver was designed in the early 1960s by the American engineer Lincoln Walsh. His objective was to make a full-range driver that avoided the time delay and phase errors that occur with conventional multi-driver loudspeakers. His design, which was referred to as a “Coherent Wave Transmission Line Driver”, was awarded a US patent in 1969. Superficially it looked like a conventional driver with a very long cone that was standing on its end, rather than facing forward, as is generally the case with conventional drivers. The driver’s vertical orientation gave it an omnidirectional radiation pattern. The way that the cone was made to vibrate meant that it effectively acted as a pulsating cylinder and it was this that made it phase coherent over its operating range. Put simply, the phase relationships of all the frequencies in its operating range were accurately preserved.

Sadly Lincoln Walsh passed away in 1971, so he never got to see how much admired and influential his design would become. In the same year the right’s to Walsh’s design were purchased by the Ohm Acoustics Corporation of New York.  It appeared in various forms in a number of their designs, but the best known of these is the iconic Ohm F, shown above. This was produced from 1972 to 1984 and was very highly regarded at the time. It used a single Walsh driver to cover the range from 37Hz to 17kHz. The Ohm F still has a dedicated following today, mostly in the United States, where examples are most readily available.

Making a full range driver

Making a single driver cover the full audio range range is very difficult, as the requirements for a driver that works at low frequencies conflict with those for a driver intended to cover high frequencies.

At low frequencies you need to move a lot of air, so you need a driver with a cone that has a large surface area. This means that the moving mass of the driver will be high. This is not a problem at low frequencies, but it is at high frequencies. The inertia of a heavy cone prevents it from being accelerated fast enough to reproduce high frequencies. Any attempt to make it move fast enough will almost certainly result in the cone going into what is referred to as “break-up”. This is where the cone surface is no longer accurately following the input signal. The result will certainly be distortion and in time possibly mechanical failure. The key to making a driver cover a wide frequency range is to find a way of achieving a low moving mass in the higher frequency range.

Enter Lincoln Walsh. He designed a driver that superficially looked like a conventional driver stood on end. The cone was approximately 30 cm tall and was made in three sections as shown below.

At low frequencies, all three sections of the cone move together in unison, so here the Walsh driver is working just like the conventional pistonic drivers used in the vast majority of loudspeakers manufactured today.

The titanium foil at the top of the cone is very thin and therefore quite flexible. At a certain frequency, the force imposed by the voice-coil causes the titanium to flex, rather than for the voice-coil/cone assembly to move in unison. This results in a wave is being launched in the titanium foil, which travels down the wall of the top section, into the middle section and finally into the third section. This type of wave is called a bending wave. If you could see the cone in cross-section, what you would see is similar to what would occur if you had two people, each holding one end of a long rope horizontally and one person flipped their end of the rope once. We will ignore the wave being reflected when it gets to the other end for now. This is relevant to a later part of this discussion.

As the bending wave travels down the cone wall, because its velocity is much higher than the speed of sound in air, it causes a sound wave to be pushed outwards.

When the driver is working in bending wave mode, the mass of material that has to be accelerated, i.e. the moving mass, is that of the voice coil and the cone material in the wave. This is considerably less than combined mass of the voice coil and entire cone and it is this reduced moving mass that allowed the Walsh driver to cover frequencies up to 17kHz. It is also proof of the genius of Lincoln Walsh.

Achieving phase coherence

For the driver to be phase coherent, the phase of the sound wave radiated from all points down the cone wall has to be aligned, otherwise there will be frequency response anomalies caused by phase interference. Due to their comparatively high densities, the velocity of the bending wave in the titanium and aluminium sections is very much higher than the speed of sound in air. By the time the bending wave has reached the bottom of the aluminium section and is pushing out a sound pressure wave there, the sound pressure wave in the air that it generated at the top of the titanium section has travelled so that it is vertically lined up the the bottom of the aluminium section. Consequently, this produces the effect of a pulsating cylinder, making the Walsh driver phase coherent. That is the theory. Walsh’s use of the two different metals, with almost certainly different phase velocities* was possibly because he could not find a single material that had the characteristics he needed, and his two section approach produced the best compromise that he could find. It may also have been to give the cone sufficient strength.

As we said earlier, at the time Walsh was working on his design, computer modelling was not available, and without this it would not have been possible to accurately determine the required material’s characteristics, cone shape and size to make this work optimally. Walsh would have had to calculate as best he could with the resources available and then build prototypes to see what worked. Even with its imperfections, the Walsh driver worked extremely well and the Ohm F loudspeaker was very well received and successful.

*Phase velocity is the term for the velocity of a wave in a medium. Generally we only talk about the speed of sound for a wave travelling in air, where it can be perceived by our ears as a sound.

Reducing reflections

We did not talk about the paper section of the cone taking part in the bending wave radiation mode. This is because it serves a different purpose. As the bending wave travels down the two metal sections, it causes sound energy to be radiated, so its amplitude progressively decreases. When it gets to the bottom of the second section, there is still energy in the wave. This creates a problem. The cone is acting as a transmission line, which the bending wave travels along. If there is a sudden change in the characteristic of this transmission line, some energy will be reflected back from this point and interfere with the forward travelling bending wave and cause distortion. The amount of energy reflected back will depend on the degree of mis-match between the aluminium and paper sections. Walsh will have sought to minimise this, but the differences in the mechanical characteristics of these materials means that some mis-match was inevitable.

Any energy in the bending wave that is not reflected at the junction of these two materials will continue to travel to the bottom of the third section, where the open end of the cone will act as a major discontinuity and reflect the energy back up toward the top of the cone, interfering with the forward travelling bending wave.

It is therefore necessary to dissipate as much of this energy as possible before it gets to the bottom of the third section. The felted paper used for the third section is a comparatively lossy material, so the wave looses more energy as it travels though this than it would have lost in the first two sections. Consequently when it gets to the open end of the cone, the amplitude of the reflected wave is comparatively small.

If you look at the Walsh driver in the second image, you will see a number of short circumferential lines in the paper section of the cone. These are cuts in the paper and their function is to reduce the amount of energy going forward, but each cut will also cause a small amount of energy to be reflected backwards. Because these reflections are spread out over the cone, they are smaller in amplitude than the reflection that would be generated if the wave travelled to the end of the third section and was reflected from there. These reflections being generated being generated from a number of points along the cone surface also means that they are spread out over time, so are less likely to generate severe resonances. This technique was probably not 100% effective, so the residual wave travelling backwards together with any reflection from the junction of the aluminium and paper sections probably did give rise to some resonances.

Conclusion

The lack of modern high power magnets made the Walsh driver and thus the Ohm F comparatively inefficient. They needed a lot of power to make them really sing and produce the magnificent dynamics that they were capable of. Unfortunately, the Walsh driver was also quite fragile and easily damaged by over enthusiastic use of the volume control. It was said of the Ohm F that it needed, “250 watts to drive a pair of Ohm Fs, 300 watts to blow them up!” The complex construction of the Walsh driver made it time consuming and expensive to manufacture and also very difficult to repair. For this reason Ohm ceased production of the model F in 1984. Despite this, the Ohm F still has a dedicated following, especially in the U.S., where examples are most easily found. If you ever get the opportunity to listen to a pair, we suggest that you do so.

In the next newsletter we will talk about how Peter Dicks, also an extremely clever engineer, took inspiration from Lincoln Wash’s work to design the German Physiks DDD driver. This omnidirectional driver solved the problems in the Walsh driver and offered considerably improved performance.

To find out more about the DDD driver please click here.

• An online review of the Emperor integrated amplifier.

• A comparison of our DDD driver with another omnidirectional driver.