Transconductance amplifiers & single-driver horns
~ by David Wright, of Pure Music

Revisiting the SET text (why horns suit SET amplifiers) after several years, I was struck by the clues spread throughout that might have pointed me towards transconductance amplifiers somewhat earlier. Unintentionally, SETs are almost a halfway house between conventional voltage amps and the subject of this article, current amplifiers (for that is what TC amplifiers are).

Back in the earliest days of audio, valves were the only amplifying devices available, so amplifiers developed which exploited their greatest strength. They were, and still are, the best voltage amplifiers, by which I mean that their output voltage is a "magnified" version of their input voltage, irrespective of load.

Loudspeakers. however, use the magnetic field created by a current flowing through the voice coil to make their sounds. It soon became apparent that when the voltage from the amplifier went up, the current through the voice coil didn't always follow it exactly. At high frequencies the inductance of the voice coil prevents sufficient current from flowing thereby reducing the audo output, and at the low frequencies the system resonance causes obvious non-linearities. A separate smaller speaker unit - a tweeter - was developed to overcome the former problem, but this in turn needed a crossover to ensure that only the high frequency signals went through it. Not only does the crossover waste power but the phase performance is very suspect.

Even with this added complication, if the speaker has been driven hard the voice coil will heat up and once again constrict the current flow. This is called "dynamic compression" and stops the "loud" from being as loud as it should be.

Fortunately, a few people continued to listen to single drive units covering as large a frequency range as possible, often using horn-type enclosures. There were occasional attempts to boost the extreme treble and bass of the single units by using tone controls, but these usually introduced precisely the phase shifts that the single speakers were themselves avoiding!

Suppose, however, that we could specify an amplifier which always delivered the correct current to the voice coil, automatically compensating for the effects of inductance or heating while allowing us to retain the simplicity and phase coherance that a single drive unit brings to speaker design. It exists, and is called a transconductance device. These have rarely received more than a footnote in textbooks and have more rarely been used for audio, yet they can deliver something very special. They allow the drive unit to operate in the way in which they were originally designed to do and not the way in which amplifier manufacturers have subsequently dictated.

The technically minded will have already determined that here is an amplifier with a voltage output which is proportional to the load and thus reflects the speaker's impedance. What then happens at the sharp impedance rise around the system resonance in a reflex type speaker? The answer is an unpleasant bass-boom, even worse than usual! Yes, these amplifiers can only be used with well-damped drive units: some infinite baffles, some panels and...horns.

To be realistic I must admit that most horns do have some resonance(s) at the bass end causing a gradual rise in impedance, but usually the amplitude of these is no more than that of the natural resonances found within any listening room and so what is heard is simply an increase in bass volume and extension. Maybe at last we shall hear the end of "everyone knows that horns are bass light, don't they?"

Another description of these amplifiers is that they put out a current directly proportional to the input voltage, a relatively straightforward process which can be accomplished in any number of different ways. I have built amplifiers using triodes and pentodes, transistors and mosfets, Class A and Class B, push-pull and single-ended. Each has its advantages and drawbacks, but subjectively the differences seem less than one might expect.

An incidental bonus is that the effects of cabling seem less marked also, and we have been experimenting with ultra-thin (and relatively high resistance) types.

However, I've been keeping the best 'till last. The usual sagging high-frequency response of the drive unit is dramatically improved and the high frequency energy directed into the room now stays almost constant as the frequency increases. Inevitably (those Laws of Physics, again) it is beamed quite tightly and so the polar reponse is modified considerably but the treble is there... all the sparkle you could want. Now one can hear the detail that was being lost before and since many of these high frequencies are mostly harmonics of lower ones the extra clarity is heard throughout the entire frequency spectrum. The subjective improvement is easily recognised. Synergy? Quite literally, they are made for each other. Super-synergy!

Please remember that this article is just a non-technical introduction to the topic and is intended only to suggest what is possible. We hope that the implications are obvious, but if not then please do not ask us about them, just arrange for a demonstration! See also Nelson Pass' inspirational website ( firstwatt) for a different slant (and some practical circuits).


Revisit Beauhorn.
To see the rest of the Beauhorn site, click on this link: HOME

Back to top