Posted on 17 July 2019

Now We are Cooking

Free Bodo's Power Magazines!




A new approach to induction heating design

Looking at a new ‘multizone’ design that allows efficient heating while reducing the number and power rating of converters and eliminating the need for relays or solid state switches. The demand for induction hobs - in which electromagnetic coupling between a ferromagnetic saucepan and concentrated coils in the cooker’s heating element are used to provide heat energy - continues to grow.

By Cesare Bocchiola, International Rectifier


While traditional induction hobs require very accurate positioning of the pan, multizone designs deliver efficient heating of the pan irrespective of its position. The challenge, however, is delivering a multizone design that remains commercially viable and minimizes component count.

Induction heating occurs due to electromagnetic coupling between the source (a coil driven by a power converter running at several tens of kHz) and a ferromagnetic pan. To work at full efficiency the relative position between the coil and the pan must be well defined - otherwise coupling becomes weaker, with the net result of increased converter reactive/active power ratio. For that reason, areas on the surface of the cooker are usually painted over with glass-ceramic substrate to indicate where to position the pan.

Multizone induction heating, on the other hand, allows efficient pan heating irrespective of position. This is typically achieved using a large number of small coils, each driven by one or more adjustable power, high frequency converters. An intelligent detection system works out the pan’s position and activates only those coils that can achieve full coupling. To increase cooking surface resolution (improving coupling), requires a larger number of coils, and thus the number of the converters can quickly increase beyond an economically viable level.

Among the requirements for a multizone design are the need to activate heating sources in any possible subset configuration (the pan could be any sort of shape, or in any orientation, but must still be heated across the whole contact area) and minimizing the component count and cost of the power conversion system. To date, designs have tended to rely on high power relays to connect/disconnect the coils from a relatively small number of converters, but this is costly. The goal, therefore, is to find a way of having sufficient coils, while reducing the number of converters (and their power ratings), and without using relays.

A new approach

One approach is to construct a modified switching matrix of elementary resonant converters. By proper connection of the converters to each other and to the resonant circuits composed of coils and resonant capacitors, any single coil can be activated and independently controlled. Converters are of the half-bridge resonant type, each operating at a fixed frequency (close to the coil-capacitor network resonant frequency). This has the advantage of automatically eliminating the interference between frequencies that can lead to acoustic noise. Fixed frequency operation is made possible by controlling the power of each coil via phase shift techniques, instead of duty cycle or frequency control methods.

Now let’s consider a novel wiring scheme that can minimize the number and size of the power converters.

Matrix design

Figure 1 shows a simplified version of the proposed switching matrix, which has been limited to a six-coil capacitor network for clarity. This matrix is composed of elementary converters and uses neither relays nor solid state switches. Converters are connected to each other through series resonant circuits, which are damped by the pan, whose loading acts an equivalent resistance. The circuits are laid out in such a way that, by turning on any pair of converters, the heating coil at the cross point is activated. The connection of the networks is such that the final design resembles a full-bridge configuration.

Simplified coil-converter matrix

Each constant frequency converter pair is operated at a constant duty cycle of 50%; with a reference signal used to ensure phase shifting between the two converters. This provides control of the power delivered to the coil. The low level signal network, controlled by an ON/OFF command and a power reference level, is provided by a central controller unit. Because of the phase shift control, when two or more converters are operated at the same time to control multiple coils, each converter pair only addresses its own series resonant circuit, and does not interfere with other converters. Unlike previously proposed solutions, coils are not placed in parallel to each other using relays, with the result that the circuits do not change their resonant frequency.

Independent control of coil power

Consider coils C01_02 and C02_02; their activation requires converters M01, M02 and N02 to be operated simultaneously (converters M01 and N02 driving coil C01_02, while M02 and N02 drive C02_02). Converter N02 is shared between the coils, but, because each converter operates at the same frequency and with a 50% duty cycle, simply choosing different phase shifts between converter pairs easily allows independent power control as shown in Figure.2).

Converter Pair Block Diagram

In this way, for a phase shift of 0°, the power delivered to the coil will be zero and for 180° shift the power will be the maximum achievable by the converters. Powers between the minimum and maximum are obtained by corresponding phase shifts between 0° and 180° Figure 3 shows the linear relationship between power and phase delay.

Coil power & current versus phase shift

Less converters and lower power ratings

In Figure.1, the converters are shown connected to the coil in a standard matrix configuration. Such a configuration, however, can present limitations.

In the standard matrix n1, n2,….., nN can be used to describe the converters in a row, and m1, m2,….., mM the converters in a column.. Each coil needs a different power level, which can be denoted by P1, P2, P3 and P4. Converter n2 will have phase Ph_n2 and m3 will have phase Ph_m3; the coil network between them will see the difference Ph_n2-m3. A power level of P2 is delivered to the coil network at the intersection between n2 and m7. Converter n2 is already activated, so it will just be necessary to activate converter m7, with a phase Ph_m7, such that Ph_n2–m7 will provide the power P2.

Again, to deliver power P3 to the coil intersecting n7 and m3 (with m3 already ON), it will be just sufficient to activate converter n7 with a phase Ph_n7, such that shift Ph_n7-m3 provides the power. Now trying to control P4 at the same time is not possible. In fact, n7 and m7 are already operating and their mutual phase shift is fixed by the power they are delivering to the other coils. A limit of a standard switching matrix is that only any upper (or lower) semi-diagonal submatrix of coils has elements whose power may be independently selected.

To overcome matrix limitations, it is necessary to employ an adjacent coil strategy. The basic idea is that adjacent coils must be independently turned ON/OFF, but when ON they do not really need to be driven at different power levels. While this seems restrictive from a theoretical point of view, it is not from a practical standpoint, because: With big pans covering more than one coil, adjacent coils share the same power level since they are coupled to the same pan. With small pans only covering one coil, adjacent coils are not driven at all. A modified switching matrix can be designed, realizing the benefit of adjacent coils, as illustrated schematically in Figure.4.

Modified switching matrix

The blocks represent converters, while circles denote coils. This modified matrix is based on two key rules. Firstly, adjacent coils must never connect to the same pair of converters, otherwise the heating resolution will be reduced. Secondly, in the case that two adjacent coils need to be turned on, requiring two pairs of converters, the same converters are allowed to drive another coil, which must be adjacent to the first two.

The general formula applying to this approach is: Ncoils = Nconverter * (Nconverter –1) / 2 Thus, the maximum number of coils that may be driven by the six converters shown in the diagram is 15. Each converter may drive up to five coils. Out of the 15 coils, only five are actually truly independent from each other, but by proper wiring, non-independent coils can be placed in a way that a small pan which needs a single coil can be independently controlled, while bigger pans covering several adjacent coils can be heated by activating these coils with the same power level.



VN:F [1.9.17_1161]
Rating: 0.0/6 (0 votes cast)

This post was written by:

- who has written 791 posts on PowerGuru - Power Electronics Information Portal.

Contact the author

Leave a Response

You must be logged in to post a comment.