Toronto AES Section Logo
Toronto AES Bulletin

Review

Oct 1999


Toroidal Transformers - An Evening With Menno van der Veen

Good fortune smiled on the AES Toronto Section as we were able to schedule a meeting with Menno van der Veen between his consulting activities with Plitron Manufacturing in Toronto and his trip to the AES Convention in New York.

As Chairman of the 550 member Dutch Section of the AES, Menno first shared some practical information about his home section. They have developed some interesting procedures they call "The Laws of De Witt" to ensure that their ten evening meetings that are held each year are effective. Basically, these consist of using checklists to evaluate the quantity and quality of the introduction, theory, technical support data and applications of proposed meeting topics.

In celebrating its recent 25th anniversary, the Dutch Section printed a book containing its history and research and development activities. It is written in Dutch but summaries and captions are in English. Menno gave us a number of copies to be shown to the members of our section.

A synopsis of Menno's technical presentation follows:

TRANSFORMER INSTRODUCTION

Aside from the its usual application as a power transformer, toroids can provide superior performance over traditional E-I laminated core transformers in audio applications such as microphone and line level interfaces, 70 V distribution, step-up for electrostatic speakers and output for tube amplifiers. Menno confined his talk to the tube amplifier output application as tubes are at the centre of high-end audio attention. He began by providing a quick review of tube basics and brought attention to the fact that the output impedance of a tube is high and can be characterized by a voltage source in series with its plate resistance. Thus, output transformers are necessary to perform an impedance transformation to efficiently drive their low impedance loads, the associated speaker systems.

Menno showed a schematic and mathematical model of an output transformer and highlighted the parameters that are superior in the toroidal design and which contribute to higher bandwidth and lower distortion. These are higher mutual coupling between primary and secondary windings, reduced inter-winding capacitance, reduced primary and secondary resistance and capacitance, very large primary inductance and very small secondary inductance. He emphasized that these benefits are highly dependent upon the winding process, the selection of core material and the manufacturing and quality control processes.

TRANSFER FUCTION

Since there is interaction between the source, the transformer and the load, Menno displayed an equivalent circuit that included the output tubes, the transformer and the speaker. This showed that each component contributes to the overall frequency and phase response. The mathematical analysis showed that four important elements contributed to the band-pass characteristic of the output transformer transfer function. These are the transformer turns ratio, its insertion loss (heat loss in windings), its low frequency behaviour (first order high-pass) and its high frequency behaviour (second order low-pass). Menno stated that he performs custom toroidal output transformer designs for customers that provide him with the design specifics such as tube type, circuit topology, load characteristics and application information. This is successfully accomplished because he can then optimize the individual transformer parameters to achieve the desired design goals.

THAT PECULIAR TUBE AMPLIFIER SOUND

Menno displayed a transconductance curve of a typical output tube and highlighted the inherent nonlinearity. It was evident that gain was greatest and plate resistance smallest when signal levels were small. This explains why low-level details are heard more clearly and why the application of negative feedback to improve tube output stage linearity weakens this inaccurate but euphoric effect. In dealing with this type of distortion, Menno has defined this effect as DDFD (Dynamic Damping Feedback Distortion) and notes that it is inherent in all tube output circuits not using negative feedback. The fact that the speaker is also a complex load (not purely resistive) also contributes to frequency response errors caused by variations in signal level and damping. Experimentation by Menno has verified that these effects can be corrected by the application of negative feedback. As all transformers exhibit band-pass behaviour, group delay will be prevalent in the region where the low frequency response falls. Based on research by Preis and others, low frequency group delay is audible and therefore should be minimized. Menno achieves this by designing his transformers with very large primary winding inductance. This effectively moves the low frequency -3dB rolloff point to a fraction of a Hertz. This is much more easily accomplished with toroids.

Harmonic distortion will commence when the core begins to saturate. This will be of concern at low frequencies because the flux density in the core is inversely proportional to frequency. This low frequency distortion is responsible for creating the false psychoacoustic perception of increased bass by exaggerating the level of harmonics associated with absent or low level bass notes. Once again, Menno defeats this distortion by designing transformers with larger cores and very large primary inductance. His bass may seem softer but it is in fact more accurate. A graph of harmonic distortion versus frequency for a 70 V distribution toroidal transformer, being representative of all toroidal audio transformers, clearly showed that the transformer was virtually distortion free at all audio frequencies above 50 Hz. Distortion rose at 20Hz only to a remarkably low 0.005%.

NEGATIVE FEEDBACK

Menno then explained the role of the wide bandwidth toroidal transformer in creating a low distortion tube audio amplifier. If the full open loop gain of the amplifier can be made available at all audio frequencies, then negative feedback will effectively reduce distortion equally at all these frequencies. This desirable situation cannot be achieved unless the transformer low and high frequency -3dB rolloff points extend at least an order of magnitude beyond the lowest and highest frequency signals to be amplified. Since these extreme corner frequencies are achievable in toroids, such designs are feasible and practical using toroids.

The evening concluded with an enthusiastic question and answer session with Menno and a hearty round of applause for his exuberant and enlightening presentation.

by Jim Hayward, Committee Member

Previous | Next

| Home | Toronto Section | Membership | Resources | Archive | Search |
Originally posted: 3 Nov 1999
Last update: 3 Nov 1999
TorontoAES@vex.net