Method and system for determining resonant frequency

Abstract

A method includes providing an LCR circuit including an inductive element and a capacitor, and applying a voltage to the LCR circuit, where the applied voltage induces a response between the capacitor and the inductive element of the LCR circuit; determining a plurality of resonant frequencies of the response at a plurality of time points respectively; a responsive output signal is generated and a resonant frequency is selected from a plurality of resonant frequencies based on the output signal, where the setting of the frequency of the applied voltage depends on the selected resonant frequency, and then the above steps are repeated.

Description

Technical Field

[0001] Various aspects of the present invention generally relate to methods and systems for determining a resonant frequency, and more particularly to methods and systems for using an inductive element in an aerosol generating device. Background Art

[0002] Aerosol generating devices are known. For example, tobacco heating devices form aerosols by heating an aerosol generating substrate (such as tobacco) but not burning it.

[0003] Common devices use an induction heater to generate an aerosol from a suitable medium, which is then inhaled by the user. Typically, the suitable medium needs to be heated significantly before an aerosol is generated for inhalation. Therefore, the heaters of such devices reach high operating temperatures. Safety of users of such devices is of paramount importance.

[0004] Many systems have been designed so far by observing the resonant frequency f res Typically, these systems rely on directly measuring the resonant circuit’s f using an external signal. res , and then observe the circuit's response by accumulating the pulses over a fixed time using a digital counter, or by using a timer that measures the interval between each pulse.

[0005] WO2020 / 260884A1 describes an induction heater and a temperature measurement method for an aerosol generating device. For such devices, there is still a need to provide improved methods and systems. Summary of the invention

[0006] The present invention provides a system and method for estimating the resonant frequency of a susceptor using an inductive element for inductive heating; the resonant frequency can be used as a representative variable, such as the temperature of the susceptor, in an inductive heating system. The system and method use an estimated resonant frequency (ERF) because the actual resonant frequency (ARF) is not known at any given point in time. The system and method may use multiple factors, including a discrete array of measurements and an approximation of a least squares fit, to generate the ERF. The ERF is highly correlated with the ARF and is time responsive, allowing the overall system to accurately control rapid temperature fluctuations.

[0007] The susceptor may be part of the removable consumable. A susceptor is an element that can be heated by induction heating, a process known in the art. In a preferred embodiment, the induction heater is designed to use only elements that exhibit significant changes in electrical characteristics due to heating as susceptors, rather than surrounding heating mechanisms.

[0008] In a first independent aspect, a method according to claim 1 is provided. The method further comprises setting the frequency of the applied voltage to a value representing the ERF based on the output signal to estimate the ARF and repeating the method steps according to the first independent aspect. The method is thus an iterative method. Advantageously, signal averaging over time can be used to improve system performance.

[0009] In a subordinate aspect, the output signal comprises a voltage (or current) signal, and the method further comprises determining a maximum value of the voltage (or current) signal, wherein the ERF is selected as the ARF among a plurality of frequencies corresponding to the determined maximum value. Thus, the method estimates the ARF of the composite induction heating system by sampling a range of excitation frequencies and monitoring the output (e.g., current flowing in the circuit) in response. For example, the frequency value at which the system voltage (current) peak occurs represents the ERF of the system, assuming that the first derivative of the response curve is approximately linear.

[0010] In a further subordinate aspect, the step of determining the maximum value comprises fitting a polynomial function to the output signal data. In a further subordinate aspect, the method further comprises minimizing the order of the polynomial function using a coefficient of determination (CD). Alternatively, determining the maximum value comprises using a least squares approximation method.

[0011] In a further subsidiary aspect, the iterative method includes setting the frequency to a value above the ERF in a second iteration and setting the frequency to a value below the ERF in a third iteration. Advantageously, this can limit the rate at which the power source and circuit need to source / sink, thereby improving overall efficiency and reducing electrical shocks on the device, extending its useful life.

[0012] In a dependent aspect, the ERF is selected to use a peak detection circuit (PDC), the PDC comprising a PDC capacitor, the method further comprising the steps of:

[0013] During the sampling time, the PDC capacitor is charged;

[0014] measuring a plurality of voltages in the output signal and calculating an average value of the plurality of voltages during a sampling time;

[0015] Discharge the PDC capacitor;

[0016] The frequency of the applied voltage is set to ERF.

[0017] In a subsidiary aspect, the method further comprises inductively heating a susceptor using the inductive element to aerosolize a substance in a heating mode.

[0018] In a subsidiary aspect, the output signal is used to provide a temperature measurement of the susceptor. In a further subsidiary aspect, providing the temperature measurement includes determining the resistance of the susceptor. Thus, the frequency value can be directly converted to a temperature value, or can be used to find the resistance value of the secondary conductor represented by the susceptor.

[0019] In a subsidiary aspect, the method comprises repeating the method steps according to the first independent aspect to provide a plurality of ERFs, determining a maximum value of the ERFs, and if the maximum value of the ERFs is within a predetermined range, setting the frequency of the applied pulses to a maximum value. In a further subsidiary aspect, the method further comprises repeating the above steps to ensure that the set ERFs are within the predetermined range.

[0020] In a subsidiary aspect, the method further comprises determining a change in the ERF (e.g., based on a change or shift in the output signal). By detecting and quantifying a shift in the actual resonant frequency (ARF) and / or response amplitude, the inductive heater device can simultaneously heat the susceptor and detect the temperature reached by the susceptor through the circuit, eliminating the need for using an external sensor (e.g., a thermocouple or thermistor).

[0021] In a subordinate aspect, selecting the ERF comprises determining a change in the resistance of the circuit, the method further comprising determining the temperature of the susceptor. Advantageously, this allows the use of susceptors made of aluminum foil and / or other common metals such as nickel or (stainless) steel and having a thickness of multiple skin depths, allowing the quality factor (QF) value of the system to be lower, indicating lower losses in the system.

[0022] In a further subsidiary aspect, there is provided a method of controlling the temperature of said susceptor in real time, using the method defined herein.A parameter reflecting the temperature of the secondary conductor can be measured in real time and advantageously used as a control variable in a closed loop temperature control system.

[0023] In one example, the method allows the use of passive tubular conductive susceptors as temperature sensors for heating processes without modification. This allows the heater to have higher performance, with more accurate temperature determination and better reliability, while the overall cost is significantly reduced. This makes it replaceable if it becomes contaminated and / or damaged during use. In addition, the method can be used to manage the heating of various materials, in the form of tubes, plates and rods, which may or may not release particulate matter, aerosols or vapors.

[0024] In another independent aspect, a system according to claim 16 is provided.

[0025] In a subordinate aspect, the voltage generator comprises a switching device for generating pulses by switching between a positive voltage source and a negative voltage source. In a further subordinate aspect, the switching device comprises an H-bridge, preferably a half H-bridge (HHB). The function of the H-bridge is to efficiently convert direct current into alternating current.

[0026] On the subordinate side, the voltage generator includes a damping resistor and a turn-off diode. Advantageously, the inclusion of such a network makes the signal smoother and prevents voltage spikes.

[0027] In a subordinate aspect, the inductive element comprises a flat ribbon conductor, for example made of copper or other suitable metal. For example, the susceptor may be made of aluminum.

[0028] In a subsidiary aspect, the susceptor is a helically wound susceptor. In a further subsidiary aspect, the susceptor has a conical shape.

[0029] The inductive element can be considered as the first circuit of the transformer, while the inductor constitutes the second circuit of the transformer. In a subordinate aspect, an electrical connection is provided around the second circuit of the transformer (i.e. the inductor). The electrical connection can be direct or include a capacitor.

[0030] In a subordinate aspect, the system includes a peak detection circuit (PDC) for determining a voltage maximum from a plurality of voltages in the output signal. In a further subordinate aspect, the PDC includes a unity gain operational amplifier (OA). In a further subordinate aspect, the PDC includes a switch whose impedance can be switched between ~0Ω and ∞Ω. This reduces the time for the voltage to reach equilibrium during the sampling time. In a further subordinate aspect, the PDC includes a ceramic capacitor, preferably made of C0G / NPO material; this improves stability over time and ambient temperature. In a further subordinate aspect, the PDC includes a bipolar field effect transistor, preferably with a low on-resistance and a very high off-resistance.

[0031] In certain embodiments, a current sensor is provided for measuring current through the inductive element. In further embodiments, a control module is provided for determining system performance based on the output signal.

[0032] In a further aspect, computer readable instructions are provided which, when executed by a computing device, cause the computing device to perform the method according to the first independent aspect.

[0033] In a further aspect, an aerosol providing system for generating an aerosol from an aerosolizable substance is provided, the aerosol providing system comprising any of the features of the system according to the above independent aspects, the aerosol providing system (or delivery system) being configured to perform an operation after receiving an output signal from an output circuit. It should be understood that the system and method according to the present invention can be used in various applicable delivery systems.

[0034] An example of a susceptor device for receiving a consumable product and for use in an aerosol delivery system, wherein the susceptor device is a spiral winding.

[0035] In a further preferred example, the susceptor device is made of aluminum.

[0036] In an example of a susceptor device, the susceptor is conical, preferably with a cone angle of 1-10°. Since the consumable is usually cylindrical, when the consumable is fully pushed into the susceptor, the lower half of the consumable will contact the susceptor to achieve good thermal contact.

[0037] Using a spiral winding structure is advantageous because it provides a way to construct a controllable, low-angle cone that is as strong as a cylinder.

[0038] Further dependent aspects are provided in the dependent claims and may be applicable to any of the independent aspects. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Embodiments of the present invention are now described with reference to the accompanying drawings, in which:

[0040] Figure 1 This is a schematic diagram of an LCR "resonance circuit";

[0041] Figure 2 (a) to (c) are plotted graphs of the relationship between the driving voltage and the current in the LCR resonant circuit;

[0042] Figure 3 (a) and (b) are plotted graphs of output according to an exemplary embodiment of the present invention;

[0043] Figure 4 (a) to (c) are examples of circuit layouts;

[0044] Figure 5 is an example of an induction heater that can be used in embodiments of the present invention;

[0045] Figure 6 is a circuit diagram that can be used in an embodiment of the present invention;

[0046] Figure 7 (a) and (b) are exemplary switch devices that can be used in embodiments of the present invention;

[0047] Figure 8 An exemplary circuit according to an embodiment of the present invention;

[0048] Fig. 9 (a) and (b) are examples of peak detection circuits (PDC) that may be used in embodiments of the present invention;

[0049] Fig.10 Plotted graph of an example measurement sequence for PDC;

[0050] Fig.11 is a flowchart of an algorithm according to an exemplary embodiment of the present invention;

[0051] Fig.12 (a) and (b) are example outputs of a pulse width modulation (PWM) system;

[0052] Fig.13 (a) and (b) are schematic diagrams of the construction of a spiral winding susceptor that can be used in an embodiment of the present invention;

[0053] Fig.14 is a graph of comparative resonance curves for an example spiral wound susceptor at five different temperatures;

[0054] Fig.15 for Fig.14 A more detailed plot of the 200°C curve;

[0055] Fig.16 Plotting an example of calculated values ​​of the temperature of a helical winding susceptor foil as a function of the estimated resonant frequency (ERF) of the system;

[0056] Fig.17 A graph is drawn to illustrate an example parameter fit for closed loop control of an induction heater temperature profile;

[0057] Fig.18 A graph is drawn to illustrate the origin of the peak voltage in an example series resonant circuit near resonance and its relationship to the circuit resistance;

[0058] Fig.19 Plotted graph of an example calibration curve of susceptor temperature versus measured peak ADV voltage. DETAILED DESCRIPTION

[0059] 1. Finding the resonant frequency in a series LCR resonant circuit

[0060] Figure 1 An exemplary LCR "resonance circuit" is shown, which includes a resistor (R t )、Inductor(L t ) and capacitor (C t ), these components are connected to a voltage source (V sw ), the voltage source oscillates in the form of a bipolar square wave with a frequency of f d (=1 / τ), where bipolar means that the output voltage oscillates symmetrically between positive and negative in one cycle. V sw An alternating current (AC) will be generated in the resonant circuit. Even though the circuit consists only of an inductor and a capacitor, there are still losses in the circuit due to the finite resistance of the component leads and the printed circuit board, which appear as resistance R t .

[0061] As is widely documented in the literature in this area ( https: / / en.wikipedia.org / wiki / RLC_circuit), this configuration has the following characteristics: If the frequency is the actual resonant frequency ARF (f res )=1 / (2π√(L t C t )), then the current in the circuit |I t | reaches the maximum value. ARF(f res ) represents the inductor L t The energy stored in the form of magnetic field and capacitor C t As the circuit oscillates, the two components exchange energy during each half cycle, leaving only the resistor to regulate the voltage supplied by V sw Therefore, from all aspects, due to L t and C t Any effect caused by the presence of disappears, and the current I t Follow V sw , where the former can be calculated by the following equation: t =V sw / R t .

[0062] The resonant characteristics of a resonant circuit also depend on the amount of losses, which is determined by R t Relative to L t and C t To measure this effect, the quality factor (QF) can be defined as Q = (1 / R)√(L / C). QF can be described as a measure of the loss in the system, taking into account the reactive element (L t and C t ) to the energy lost in each complete cycle of the driving square wave. Therefore, Q=100 means that ~1% of the energy is lost per cycle.

[0063] Apart from Figure 1 R shown in t , L t and C t In addition, the system also includes C t and L t A circuit "tap" at the connection point consists of two series resistors R s1 and R s2 And low value capacitor C smpl Composition, C smpl The capacitors are preferably in the picofarad (pF) range, forming a classic potential divider, where V smpl =V t ×R s1 / (R s1 +R s2 ), where V t YesL t and Ct The voltage at the node formed by the connection point is s1 and R s2 value, which effectively changes V smpl The amplitude is adjusted from 0 to V t . C smpl The presence of will block any DC current from the resonant circuit, which means that V smpl will be a symmetrical sine wave centered around GND (ground voltage).

[0064] Figure 2 (a) shows an example value of V when τ = 20 sw Relative to I t and V smpl In this example, although V sw It has square wave characteristics, but the current flowing in the circuit presents a sinusoidal waveform. This is due to the bandpass frequency filtering characteristics of the LCR circuit, which attenuates the high-order Fourier terms in the square wave. As the Q value increases, the filtering effect will be better.

[0065] If the driving frequency f d Set to ARF(f res ), then the reference driving voltage V sw When t The peak and valley values ​​of V occur at τ / 4 and 3τ / 4, respectively. To obtain a finite voltage reading, V must be measured relative to ground (GND). smpl Since the only implementation is C t Fully charged and I t = 0, so it will shift 90° to I t (τ / 4). Figure 2 From the graph shown in (a), V sw and I t are in phase throughout the cycle. Due to this offset, when |I t |When reaching the maximum value, V smpl =0; when |I t When |=0, |V smpl | reaches its maximum value. Therefore, in this case, V smpl The zero crossing points of are located at τ / 4 and 3τ / 4, which constitute the criterion for resonance.

[0066] Figure 2 (b) and 2(c) show that when f d Does not match ARF(f res ), due to L t , C t or R t What happens when the value changes. Figure 2 (b), when fd <f res When C t The impedance of is dominant, so the current leads the voltage, similar to the case of a pure capacitor connected to an AC source. Figure 2 (c) shows that when f d >f res In this case, the inductance L t dominates, causing the current to lag behind the driving voltage.

[0067] exist Figure 2 In both cases (b) and 2(c), V sw and V smpl The curves of V smpl The magnitude is also smaller than Figure 2 (a) shows the amplitude. For an untuned resonant circuit, if V sw and V smpl , as f d Approaching f from either direction (high frequency to low frequency or vice versa) res , V smpl Relative to V sw The amplitude will increase. d =f res When V smpl reaches its maximum value. This is because when C t and L t Resonance occurs when the complex impedances are equal, and due to C t and L t With a phase difference of 180°, they cancel each other out, leaving only R t impedance.

[0068] Therefore, in this case, V smpl To determine ARF(f res Specifically, if the circuit is excited by a set of N discrete frequencies, and ARF(f res ) is roughly in the middle of the frequency array, then V for each frequency increment can be obtained and plotted smpl For best results, each V smpl The variance of the data points should be as small as possible relative to their magnitude, which may require taking many measurements of the parameter and averaging them.

[0069] Despite this accuracy, this numerically derived value cannot be the true value of ARF. Therefore, from now on, this quantity will be called ERF, and ERF and f res The claimed values ​​of will be considered the same and interchangeable.

[0070] One way to estimate the peak value (ERF) of the data is to fit a polynomial curve to the data set in the form smpl With the excitation frequency f d The associated form is:

[0071] V smpl = af d n + bf d n-1 + cf d n-2 +…+constant(1)

[0072] While polynomial fits work well, it should be understood that other mathematical functions based on logarithms, exponentials, and trigonometric functions can be used. In addition, there are many ways to fit a set of data to a mathematical function; in this case, the least squares approximation (LSA) method ( https: / / en.wikipedia.org / wiki / Least_squares ).

[0073] If we take the derivative of equation (1), we get the following equation:

[0074] dV smpl / df d = naf d n-1 + (n-1)bf d n-2 + (n-2)cf d n-3 +……… (2)

[0075] V smpl When f reaches its maximum value d value, that is, f d =f res , can be found through the following relationship:

[0076] dV smpl / df d = naf res n-1 + (n-1)bf res n-2 + (n-2)cf res n-3 +……= 0 (3)

[0077] In order to find ERF(f res ), we must find the roots of equation (3). This can be achieved using a variety of existing methods (https: / / en.wikipedia.org / wiki / Root-finding_algorithms).

[0078] Figure 3 (a) and Figure 3 (b) shows the adoption of this technology. Figure 3 (a) shows the behavior of the current flow through the series LCR circuit, with values ​​around f res =2.07MHz and enumerates at discrete frequencies with an interval of 2.5KHz. Figure 3 (b) Then at f res The narrower 40KHz windows on either side show more detail. Since the peak shape is relatively symmetrical, a good fit can be achieved using a second-order polynomial of the following form, as shown by the dashed line:

[0079] V smpl = af d 2 + bf d + c (4)

[0080] In order to find f res Values:

[0081] dV smpl / df d = 2af d + b = 0 (5)

[0082] Therefore, it follows that:

[0083] f res = -b / 2a (6)

[0084] In this case, although the points are generated by the equations, the fitted line (dashed line) is derived using the least squares method and is therefore independent. Using the coefficients and methodology outlined, equations (4), (5), and (6) produce f res The overall value agrees with the calculated value to five significant figures. It should be noted that increasing the order of the chosen polynomial generally improves the derived f res This will increase the computational cost because the computational cost of solving the polynomial coefficients and roots is nonlinear. Therefore, an effective strategy is to use the lowest order that can achieve an acceptable coefficient of determination (CD = R 2 ; See https: / / en.wikipedia.org / wiki / Coefficient_of_determination It should be understood that any feasible method can be used to fit the polynomial coefficients and find the roots of the solution. Once the ERF (f res ), you can reset the scan parameters using that value as the center point of the scan and start the process over again.

[0085] As mentioned above, this process can be automated and programmed into a small microprocessor, for example, to implement a real-time feedback mechanism. Once the resonant condition is met, the system can continue to monitor in case component values ​​change or other external factors may change f res If the system runs continuously for a long time, signal averaging can be used to improve performance.

[0086] Furthermore, this method is equally applicable to any generalized LCR circuit layout, Figure 4 Three examples of these are shown in Figure 1. Note that the topologies shown are only a subset of the possible LCR combinations, including many combinations of inductors, capacitors, and resistors in series and parallel. In general, parallel topologies maximize the impedance of the circuit and minimize the impedance at f. res Furthermore, if you add multiple component classes, such as two inductors, a capacitor, and a resistor, and can vary the values ​​of these components, the options are endless, although only a small fraction will be useful in real applications.

[0087] 2. Resonant induction heating circuit

[0088] One application of the above technology is in induction heaters, an example of which is an induction heater such as Figure 5 shown. Figure 5 An exemplary induction heater is shown, where L t It is composed of a coil 18 wound around a cylindrical metal sheet such as aluminum or stainless steel called a susceptor 16. t It can be regarded as a transformer with inductance L prim The primary 18 of the sensor 16 is a transformer with an inductance L sec In this case, L prim With multiple coils N prim , while L sec As configured in this example, there is only a single turn.

[0089] To maximize the heating effect of the susceptor, the preferred susceptor 16 is made resistive by composition, shape, or both (e.g., metal selection and thickness) so that prim The induced current flowing in the susceptor 16 can induce a significant current in the susceptor 16, thereby generating significant ohmic heating.

[0090] Figure 5 The example in FIG. 1 includes a conductor 18 of rectangular cross-section having a high aspect ratio of width to thickness to fully utilize the coil's ability to conduct current. This is because of a phenomenon known as the "skin effect" ( https: / / en.wikipedia.org / wiki / Skin_effect), so any frequency of alternating current is actually limited to flow within a depth from the surface of the conductor 18, which is called the skin depth δ. In the actual implementation of the induction heater, the 4.5 turns of the coil 18 are made of copper. For an alternating current of 2MHz, δ is about 50μm. For a 1mm diameter and a total cross-sectional area of ​​0.79mm 2 The circular cross-section conductor of the copper tube only has 19% (0.15mm) of the AC current. 2 Alternatively, if the primary coil is preferably made of a copper strip with a width of 2.5 mm and a thickness of 100 μm according to an embodiment of the present invention, about 100% of the copper participates in the conduction, corresponding to a cross-sectional area of ​​0.25 mm 2 . The strip achieves a 62% reduction in resistive losses while requiring only 31% of the copper compared to the round case. Therefore, the flat strip conductor (FSC) 18 is more energy-efficient and resource-efficient by using a metal strip with a thickness of 2δ instead of a round cross-section wire to form the primary coil. It should be noted that many induction coils use a special winding wire commonly known as "Litz Wire" (LW) ( https: / / en.wikipedia.org / wiki / Litz_wire ). Although LW is used in many applications, it is cumbersome and costly to use in consumer products compared to FSC and is produced by relatively few suppliers.

[0091] Figure 6 shows the equivalent circuit produced by this configuration, where R sec is the ohmic component of the secondary impedance, Z sec is the inductance of the secondary impedance. Using complex notation, Z sec =R sec +j(XL sec +X Csec ), where XL sec =jωL sec , X Csec =-j / (ωC sec ), j = √-1. In order to maximize R sec To dissipate the energy in the circuit, the complex components of the impedance must cancel each other out, which in turn maximizes the current in the circuit, meaning that f res =1 / (2π√(L sec C sec )) driving circuit. According to the transformer equation, the impedance Z in the sensor 16 sec It is amplified to N in the primary circuit. prim 2 (See for example http: / / hyperphysics.phy-astr.gsu.edu / hbase / magnetic / refload.html ).

[0092] The purpose of the susceptor 16 is to extract energy from the primary circuit and convert it into heat, thereby significantly raising the temperature of the susceptor, typically to about 200°C, so as to be able to heat an object placed within it to a similar temperature. Using the above relationship, if L prim Composed of 4 turns, it will make Z sec The nominal value of R is increased by 16 times. Since the goal is to minimize losses in the primary circuit and the transformer, a well-designed circuit will only dissipate the available energy in the secondary circuit. sec , Figure 6 R is omitted t . Return to reference Figure 1 , at this level of analysis, R sec and L prim / L sec The combination of can be equivalent to R t and L t .

[0093] When the receptor 16 heats up, the ARF changes. In practice, there are many potential reasons for this effect, such as sec Overall, this effect has several consequences, including:

[0094] - If the initial f d =ERF, but due to heating, ERF shifts, and f d If left unchanged, the efficiency of energy transfer to the susceptor 16 will be reduced, resulting in a slower rate of heating and possible thermal shock to the primary circuit.

[0095] - If this variation is tracked during the heating phase, its value can generally represent the susceptor 16 temperature (see Section 6 below), eliminating the need to connect and monitor a discrete temperature sensor. Therefore, being able to measure ERF wirelessly during heating (and cooling) has several advantages for the management of inductive heaters.

[0096] Reference Figure 6 In order to provide sufficient current to the susceptor 16 forming the secondary circuit, it is necessary to provide electrical connectivity around the secondary circuit, for example:

[0097] -In R sec The left hand side (LHS) terminal and L sec The right-hand end (RHS) terminals of the 1000A are connected directly with a wire or similar. This connection is suitable for I sec , including direct (non-oscillating) current. This type of connection is found, for example, in domestic stove induction heaters.

[0098] - Alternatively, for AC, a capacitor can be used instead of a wire. Preferably, a capacitor large enough is chosen so that I sec The peak voltage developed between the capacitor plates does not exceed the dielectric strength of the insulation between the plates. Failure to meet this condition may reduce the efficiency of susceptor heating and / or cause arcing, damaging the capacitor structure and affecting the overall efficiency of energy transfer between the primary and secondary circuits.

[0099] In order to achieve an adequate level of heating within the secondary circuit, the circuit needs to have a minimum value of admittance to allow sufficient power to be transferred from the primary circuit to the secondary circuit. If there are few ohmic or capacitive connections around the circuit that prevent a sufficient amount of I sec Flow, the heating of the susceptor 16 will be limited or even prevented.

[0100] 3. Driving the resonant circuit

[0101] exist Figure 1 In, V sw It is defined as a bipolar square wave generator capable of providing sufficient current to heat the susceptor 16 . Figure 7 (a) shows a circuit form to achieve this goal, namely an H-bridge (HB) switch arrangement as defined in the art (see https: / / en.wikipedia.org / wiki / H-bridge and https: / / ieeexplore.ieee.org / document / 5411854 The function of an H-bridge is to efficiently convert DC power to AC power, albeit in the form of a bipolar square wave, by synchronously switching a series of transistors to generate a bipolar square wave. One of the most common uses of an H-bridge is in driving electric motors, where the four transistors can effectively control the direction and speed of rotation of the motor through a precisely synchronized set of control pulses that are sent to the bases (or gates) of the transistors.

[0102] Reference Figure 7 (a), the so-called "full H-bridge" (FHB) consists of two pairs of field effect transistors (FETs), namely FET11 / FET12 and FET21 / FET22, each of which is connected to a DC power rail (V+) and ground (GND). As shown in the figure, the FHB manages a resistive load (R L )’s energy supply.

[0103] In this example, the FET is an n-channel device, but p-channel devices can also be used, or even P-channel and N-channel devices in the same package. It is important to note that other transistor technologies such as bipolar devices can work equally well in this application. In fact, the pairs are manufactured on the same piece of semiconductor, so they have similar responses to voltage levels applied to the gate connection (Gxx) and internal heating when switching large currents. With careful design and feasible component matching, a single transistor package can replicate the functions described below, but it will inevitably be more expensive and occupy more space.

[0104] To generate an oscillating voltage on RL, in the first half of the oscillation cycle, a positive voltage is applied to gates G11 and G22, forming a low resistance path (about 0.01Ω), while G12 is grounded and G21 is connected to the right hand end of RL, which puts them in a high impedance state with an impedance of the order of GΩ. This will cause current to flow from P1 to P2 through RL. In the second half of the oscillation cycle, a positive voltage is applied to G21 and G12, while G11 is connected to R L This will induce current to flow from P2 to P1, reversing R L The above process is then repeated to L A continuous bipolar square wave current is generated on the FHB. A characteristic of the FHB is that the voltage swing is ±V with an amplitude of |2V|.

[0105] Although these circuits can be made from relatively simple and small devices, most of which are about the size of a match head, the FHB is a relatively complex circuit, and R L With duplicate components on both sides, each pair of FETs requires independent but synchronized signals to ensure that the AC in RL remains symmetrical. In addition, to avoid wasting energy and damaging the FETs due to the so-called "shoot-through" (ST) phenomenon, the four gate signals must be precisely timed with each other, using the so-called "deadband", which adds complexity ( http: / / www.irf.com / technical-info / appnotes / an-1032.pdf ), because control must be done between transistors within the same package and between two packages.

[0106] To solve these problems, you can use Figure 7 (b) The simplified system shown in Figure 2 only uses half of the components in the FHB and is called a half H-bridge (HHB). Figure 7 (b) is shown. At the right hand end, P2 is connected by two capacitors C in series. 21 and C 22 Instead, RL is coupled to their common point (CP). The capacitor can be of any value, but if RL is replaced by a reactive load ZL , then its capacitance should be greater than Z L The upper limit is determined by size and cost. Multilayer ceramic capacitors are best, type marked X7R or Y5V. For pure DC voltage, C 21 and C 22 The voltage across the terminals is evenly distributed so that the voltage at CP is 1 / 2V+. Due to the characteristics of capacitors, if there is any AC component in the signal as well, CP will be at 0V / ground potential. For this example configuration, CP is always at 0V since HHB has no DC component.

[0107] In order to generate an alternating current, if in the first half of the oscillation period, G 11 Apply a positive voltage and set G 12 Connected to ground, the current will flow from V+ through FET11 and RL, and then through C 22 In the next half cycle, if G 11 Connecting to R L The left hand end of the C 21 , R L And flows to ground through FET12. L From this perspective, the total voltage swing is ±1 / 2V+, giving a peak-to-peak amplitude of |V|.

[0108] 4. Construction of the resonant circuit heating system

[0109] Assuming the induction heater can accommodate the lower voltage swing, the advantage of the HHB switch system is that it cuts in half the number of active components (such as FETs and drive electronics) required to sequence and control the switch voltage G. 1x In addition, since the capacitor passively prevents the occurrence of ST phenomenon, the management of ST is also simplified. This almost cuts the BOM cost and PCB space requirements in half, because the capacitor C 21 and C 22 It is cheap and occupies little space, and simplifies the control signal. Overall, fewer active components improve the reliability of the system, while reducing heat dissipation, further improving reliability and making the system easier to integrate into the product.

[0110] Back to Figure 1 The inventors propose that HHB is suitable for electrically excited resonant circuits. Figure 2 It can be seen that in response to V sw square wave characteristics, HHB will produce a relatively smooth, high-quality sine wave output V smpl , and there is no discontinuity under resonant and non-resonant conditions and the additional noise is negligible.

[0111] Figure 8 Shown is a combination of Figure 1 The sampling connection and Figure 7 As mentioned above, in order to establish oscillation in the resonant circuit, it is necessary to connect V g1 and V g2 Applying gate voltage alternately triggers FET1 and FET2. Figure 8 The left hand side shows V g1 and Vg2 characteristic waveforms, these waveforms are offset in time, but at V g1 Close and V g2 There is a gap in the time axis between turning on and vice versa, called the “dead time”, which prevents both FETs from being in a low resistance state for a moment, preventing large and usually uncontrollable currents from flowing through the FETs from V+ to GND, commonly referred to as “shoot-through” ( http: / / www.irf.com / technical-info / appnotes / an-1032.pdf ), which may damage the circuit and the power supply. For high-speed resonant circuit excitation, the duration of the dead time should be as short as possible. In this case, V g1 and V g2 The TPS28226DR device (not shown in the drawing) manufactured by Texas Instruments will be generated and automatically synchronized.

[0112] The gate drive voltage is then passed through a Schottky diode (D gx ) and a low value (about 5Ω) resistor (R gx ) in parallel to the input network applied to the FET. In this example, the resistor slows down the positive-going edge to prevent it from ringing. Conversely, the diode increases the rate at which charge is pulled out of the gate when the gate is turned off. Together, this addition improves the switching characteristics by dampening voltage oscillations (see Fundamentals of MOSFET and IGBT Gate Driver Circuits, SLUA618 Application Note, p. 19, www.ti.com / lit / slua618 ). This is because the input gate of the FET acts as a capacitor that needs to be charged / discharged to switch the transistor between conducting and non-conducting states. Without this addition, when operating at high speeds (in this case >1MHz), the gate charging may become unstable, causing V gx Oscillations eventually lead to unstable current flow and increased noise in the resonant circuit.

[0113] The inventors evaluated several different FETs and found that Nexperia TM BUK9K5R1-30E device ( https: / / www.nexperia.com / products / mosfets / automotive-mosfets / BUK9K5R1-30E.html) is the best device to form HHB. Using BUK9K5R1-30E can switch 40 amps from 30V supply.

[0114] In addition, V smpl (And from this, it can be inferred that V sw ) is not important, since analytical measurements rely on differential methods that exclude constant offsets. Nevertheless, for good practice, it is usually necessary to ensure that the system inputs are operating normally, so Figure 8 The circuit includes a monitoring output V trig Ideally this would be at Rt / FET1 / FET 2 However, the inventors found that the waveform at this point is seriously distorted. g2 / R g2 Network Post V g2 The smoothness of V trig Derived from the point between the input damping network and the gate connection of FET2.

[0115] 5. System measurement and control to determine f in real time res

[0116] To determine ERF in real time, embodiments of the present invention employ the techniques discussed in Section 1 above: d =1 / (2π√(L t C t ))=f res When the impedance |Z t | reaches the minimum, current |I t | reaches its maximum value. By measuring V smpl Representative I t To detect this condition, a peak detection circuit (PDC) is used. A PDC is an electronic device that iteratively measures the peak voltage of the input signal (tinyurl.com / 6xhjynww).

[0117] Fig. 9 (a) shows a circuit consisting of a diode (D1), a capacitor (C1), and a discharge resistor R 1 The PDC is composed of 1 Not charged, any application to V in A positive signal will force current into C1 through D1 until the voltage across the capacitor equals V in If V in remains at a constant level, the voltage across C1 will also remain at a constant level. in drops to 0V, then through C 1 and R 1After a period of time determined by the RC time constant, the charge will drop to 0V, effectively resetting for the next peak detection cycle. Fig. 9 The circuit layout in (a) is equivalent to the input circuit of most analog-to-digital converters (ADCs) integrated into many microcontrollers. Therefore, it is advantageous that these systems can be transformed into PDCs by simply adding external diodes and appropriately sized resistors.

[0118] In fact, Fig. 9 The PDC circuit in (a) has some disadvantages mainly in terms of speed and flexibility. To solve these problems, Fig. 9 (b) shows an alternative PDC circuit where V in It is buffered by U1, a unity-gain operational amplifier (OA), whose output passes through diode D1 and then charges capacitor C1. Using an op amp, V in Isolation from C1 and low output impedance allows the capacitor to charge quickly, generating V in The circuit impedance is irrelevant.

[0119] In addition, Fig. 9 In the PDC circuit (b), R 1 is replaced by a switch whose impedance can be switched between approximately 0Ω and ∞Ω, which allows collection of all 1 The charge during sampling is reduced, thereby reducing V out The time required to reach equilibrium. When the system needs to be reset for a new reading, the SW can be turned off 1 , so that C1 is fully discharged within a microsecond time scale. With this arrangement, if the voltage to be sampled is applied to V in After a period of time, measure the voltage V across the capacitor. out , which represents the voltage during the sampling period V in During this sampling process, it is best to set V in With V smpl Disconnect to avoid random fluctuations and noise added to V during sampling. out This is accomplished by adding a second control line (V in To obtain the best repeatability, first disable the control -VIDC by turning off switch SW 1 Drain the charge in C1, making V out ≈0V. Then turn on SW 1 to allow the PDC to begin sampling V again i For maximum stability over time and ambient temperature, the capacitor should preferably be a stable ceramic capacitor such as C0G or NP0, SW 1Bipolar field effect transistors with low on-resistance and very high off-resistance are preferred.

[0120] Fig.10 A waveform of a measurement cycle is shown, illustrating Fig. 9 (b) Characteristics of the circuit. At about 2.06MHz ERF (f res )(V in , dotted line), the voltage across the capacitor (V out , solid line) and the discharge control voltage V sw (dashed line) near the driver V smpl The AC voltage, when V sw Set SW when voltage is set to zero 1 To use PDC to find V smpl The following four-step method can be used to determine the magnitude of

[0121] Step 1-C1 charging: refer to Fig.10 The time displayed on the X-axis is t = 30μS, SW 1 Setting it to open circuit turns VIDC off, allowing charge to accumulate from U1 to C1, which is caused by the charge applied to V in V smpl As more pulses are applied to V in , V out Gradually rises, approaching V smpl When this happens, the increment caused by the subsequent pulse is smaller than the previous one. In an embodiment of the present invention, V out The semi-stable value is reached after about 60 pulses. To improve stability, the device can be programmed to 1 Using the previous values ​​as a guide, feed the PDC with 60 cycles to make V out A stable voltage appears on the C. To some extent, using a fixed number of pulses to 1 The collected charge defines a reference baseline. In another embodiment, SW 1 Keep it open and ignore V for a fixed time in It is composed of discrete pulses. The duration of this phase is 30μS.

[0122] Step 2 - Measure V out :At t = 60μS, the next step is to use an ADC (not shown in the figure) to measure V out . The device is V out Continuous readings are taken and stored in memory. When multiple samples are averaged, a better signal-to-noise ratio can be obtained. Before collecting samples, to ensure that V outstability, VIDC is set to “ON”, which prevents V smpl is applied to V in The duration of this phase is 42μS.

[0123] Step 3 - Data Averaging and Discharging the Sampling Capacitor: At t = 102 μS, the samples are summed in the accumulator. Since the output of the ADC is usually an integer, to speed up the averaging process, it is best practice to set the number of samples to a power of 2, P (e.g., 64, where P = 6). This provides the opportunity to average the total by shifting the binary number right by P bits (towards the least significant bit), which is equivalent to dividing by 2. P , which makes the processor efficient. During this period, a signal is applied to switch SW1, causing it to close, thereby storing 1 The charge in V provides a path to ground. In fact, after about 1μS, V out Returning to a value close to 0 V, the PDC is now ready to start measuring the next data point. The duration of this phase is 5 μS.

[0124] Step 4 - Set next fd value: At t = 107 μS, the next set of fd values ​​is sent to the programmable clock oscillator, causing it to generate new square wave values ​​through HHB. Once the clock is stable, the program returns to step 1. The total duration of this stage is about 2 μS.

[0125] As mentioned above, in order to obtain a res ), a series of N amplitude-frequency averages must be collected to form a balanced V smpl With f d series, where f d From below ERF(f res ) starts, passes through and exceeds ERF(f res ), which is roughly located at the middle point to assist the LSA calculation. In this way, the selected ERF (f res ), although this may vary due to external factors. Overall, using 20 data points, the ERF (f res ) can be completed on a time scale of 6 milliseconds. Therefore, this technique is advantageous for providing frequency tuning data, which can be used as a proxy for sensor temperature at frequencies >100 times / second during heating.

[0126] When performing a discrete frequency sweep, as f d The value is close to ERF(f res ), V smpl increases, since a corresponding amount of energy is stored in the resonant circuit, which must come from V swAs mentioned above, the rate at which energy is transferred into and out of a resonant circuit is controlled (and limited) by its Q. Since the step-by-step frequency sweep process needs to be performed quickly, on a time scale of less than 2 milliseconds, with minimal distortion, to allow the resonant circuit to fully reach equilibrium at each new frequency, it is necessary to minimize V smpl Since the profile of the sample series is designed to be roughly symmetrical, such as Figure 3 As shown in (b), the following method can be adopted.

[0127] Back to Figure 3 (b) If the first sample is acquired at 2.045 MHz, then execute Fig.10 The second sample can then be taken at 2.085MHz, which will typically give a V that is roughly equal to 2.045MHz. smpl The third fd value can be 2.048MHz or 2.083MHz, both of which are closer in absolute value to the derived ERF (f res ), which will make V smpl A small increase in the fourth f requires more energy to be transferred to the resonant circuit. d The value is 2.083MHz or 2.048MHz, both of which are energy neutral. This process can be continued until it converges to ERF (f res )'s approximate location.

[0128] The sample readings measured during this process are then sorted in ascending frequency order in the microprocessor and passed to the LSA to estimate the current ERF (f res ) value. Among the 20 samples collected during this process, V smpl In another embodiment, a similar algorithm can be used, first from the previously selected ERF (f res ), the second ERF is selected after the previously selected ERF (f res ) a small increment above the previously selected ERF (f res ) in small increments below the starting selected ERF. Once the ERF values ​​are generated, the array of points can be filled in order of frequency or in order of decreasing signal amplitude as it either descends from the previous ERF peak or rises towards the current peak amplitude value. Over the remaining 17 steps, the current drawn from the power supply will be gradually reduced regardless of whether the second ERF is above or below the starting selected ERF. Overall, this implementation of the stepped technique can be used to limit the charging rate that power supplies and circuits need to source / sink, which will improve overall efficiency, reduce electrical stress on the device, and extend its life.

[0129] If, on the other hand, the 20 samples are simply stepped linearly from below to above ERF (f res ), the rate at which energy flows into the resonant circuit will be <f res is twice as fast as the above method, and then in about 10 steps>f res In both cases, energy needs to be supplied to or extracted from the resonant circuit at twice the rate as described above, which may result in V smpl With f d The data is distorted, so that the selected ERF (f res ) values ​​are skewed.

[0130] It should be noted that through V smpl The amplitude of ERF(f res ) technology is possible because the HHB generates a smooth and stable sine wave characteristic for I t , which is in V smpl is reflected in both the resonant and non-resonant states. In particular, in the FET 1 and FET 2 Add R to the gate input gx and D gx network (i.e., damping resistor and turn-off diode), at V gx Smooth state changes and prevent I t voltage spikes on the V smpl Without these networks, these spikes would unduly interfere with the values ​​obtained from the PDC. Other drive circuit configurations may not produce such a clean signal as the HHB with its input network, so using a PDC in an alternative bridge configuration may not always be feasible.

[0131] Fig.11 is a flow chart showing the control algorithm of the device in an embodiment of the present invention. When the device is turned on, an initial position scan of 100 samples will be performed over a 1 MHz bandwidth to determine the "rough" position of the ARF. The scan range will be wide enough to determine the ERF (f res ) value, and based on this, the system will store ERF(f res ) should appear in the upper and lower frequency range. If ERF(f res ) is out of range, the system will terminate the program and stop using the device for safety reasons.

[0132] On the other hand, if the ERF value is within the tolerance range, a second 20-point scan, called a “wide scan”, is performed using the new ERF (f res) value as the center point and a bandwidth of 400KHz. Again, the system will use a finer set of frequency tolerances to evaluate whether the ERF is valid. If not, the system will return to the 1MHz bandwidth and repeat the previous step. Otherwise, the system will perform a final "tracking scan" (200KHz bandwidth and 20 points) to obtain a usable derived ERF (f res ) and passes it to the temperature control system, which can be a thermostatic (on / off) device or a more complex proportional-integral-derivative controller (PID-C) (https: / / en.wikipedia.org / wiki / PID_controller).

[0133] A common difficulty when measuring the temperature of a sensor through a heating system is that the measurement process will inevitably increase the temperature from the nominal value, resulting in measurement deviation. Therefore, it is necessary to minimize this effect. During operation, when the system is heating or maintaining a stable high temperature (>200°C), the inherent heating effect on the temperature determination can be considered less significant. On the other hand, the heater is not turned on most of the time. In this case, the system must be able to provide stable temperature readings under all conditions, so detailed analysis is required.

[0134] According to the above steps, using Fig.10 The time shown in res The time required to calculate the value is composed of the following components:

[0135] Step 1: 30μS - Open SW1, allowing C1 to charge to V smpl Peak

[0136] Step 2: 42μS - Take 64 ADC readings

[0137] Step 3: 5μS - Close SW1 to reset C1 and average the results

[0138] Step 4: 2μS - Set the oscillator clock to the next f d value

[0139] Total 79μS

[0140] If N = 20, data points can be accumulated in 1.6 mS. Since the curve fitting algorithm of the LSA technique selects ERF (f res ) takes about 6mS. Data collection can be performed independently using direct memory access (DMA) technology. At this time, the overall speed will be limited by the LSA algorithm.

[0141] In actual heating systems, PID-C usually controls the temperature curve by adjusting the power applied to the heating circuit using a pulse width modulation (PWM) system. https: / / en.wikipedia.org / wiki / Pulse-width_modulation In this scheme, the square wave excitation (V sw ) (about 2MHz) is in turn replaced by a second lower frequency square wave f pwm Modulation, where the duty cycle (DC) of the peak and valley values ​​can be adjusted between 0% and 100%, corresponding to the heater being off and operating at full power, respectively, with the values ​​in between acting as pseudo-analog control of the heating process. In this way, PID-C can control the heater more finely than a simple binary (on-off) thermostatic control algorithm.

[0142] f pwm The choice of is often driven by external factors, one of which is the audible response of human hearing, which is often quoted as 20-20KHz, with a peak around 3.5KHz ( https: / / en.wikipedia.org / wiki / Equal-loudness_ contour ), the sensitivity drops rapidly below 300Hz and above 20KHz. If the electronic device modulates a strong current, an audible sound (such as "power hum") may be produced, which may be annoying to the user. Since the time required to collect 20 unique data points is about 1.6mS, if f pwm >20KHz, its period cannot be accommodated, so in the following discussion, the PWM time base frequency will be set to about 100Hz with a period of 10mS.

[0143] For example, to create a temperature profile for a channel, there are four different control stages:

[0144] Phase #1: Idle at room temperature, waiting for the segment to activate. During this phase, the heater is not required, but it is important to monitor the ambient temperature to create a starting temperature when the heater activates. In this case, a temperature reading can be taken every second. This can be achieved by reducing the time base to 1Hz, and setting the duty cycle to 20%. This will reduce the overall heater duty cycle to approximately 0.16% of the average heater on time, resulting in negligible heating.

[0145] Phase #2: Heating the filament to operating temperature. At this point the user has loaded the filament into the susceptor and started the heating cycle, with the goal of getting the system to operating temperature as quickly as possible. In this case, the duty cycle is usually set to about 90-100%.

[0146] Phase #3: Maintaining a constant operating temperature. Typically, the maximum power of the heater is greater than the power required to maintain the maximum operating temperature, so the duty cycle is reduced to 30-50%.

[0147] Phase #4: Cooling down after the session. The simplest way to achieve this is to set the duty cycle to 0%. In some cases, it is desirable to control the cooling curve, but this still requires a lower duty cycle than phases #2 and #3. After completing this phase, the system returns to phase #1.

[0148] Of all the four phases, the most critical to the temperature measurement system is phase #2, where the rate of temperature increase is the highest, so careful integration of PWM and temperature measurement is necessary to achieve precise control in this phase while not interfering with the rapid heating.

[0149] According to the above, since the susceptor 16 is heated in the temperature measurement method, it is practical to merge the temperature measurement pulse into the leading edge of the PWM pulse. Fig.12 The diagram in (a) shows a 100 Hz, 90% duty cycle (9 ms on / 1 ms off) PWM with frequency / temperature data acquisition integrated into the first 1.6 ms of the waveform. In phase #2, data acquisition occupies the first 16% of the available duty cycle, leaving the remaining 84% of the duty cycle available for heating.

[0150] Fig.12 The graph in (b) shows that PWM can easily reduce the frequency / temperature determination to a 20% duty cycle, indicating that the method can be used unaffectedly in all heating and cooling stages. In addition, using these 20 data points, the selected ERF (f res ) must be evaluated by a microprocessor. Current testing on a Microchip Inc. PIC32MX1 shows that this can be done in about 6 milliseconds, fitting neatly into the remaining 8.4 milliseconds of the PWM cycle time. Overall, by using this format, the temperature of the sensor can be estimated at a rate of 100 Hz during all stages of device operation.

[0151] 6. Construction of a practical external receptor for aerosol tube consumables

[0152] according to Figure 4 , the induction heating device includes a high aspect ratio (length / diameter) cylindrical metal foil susceptor 16, into which a similarly shaped object is placed to generate heat under controlled conditions. Using rectangular materials and bonding along straight edges to form a cylinder, although simple, may have limited strength in the case of high aspect ratios. In contrast, if the same rectangular strip is spirally wound, the angle at which it is wrapped around the mold will change the overlap, thereby changing the rigidity and overall strength. This is why paper straws are manufactured in this way, allowing the structure to be made using similar manufacturing techniques: see https: / / tinyurl.com / 59dn883a For reference.

[0153] Fig.13(a) shows the construction of a thin spiral wound susceptor using foil 20 bonded to high temperature resistant paper, which can withstand temperatures up to 200°C. In this example, the foil thickness is 12 μm. Similar materials are produced by various suppliers, and in this article, Bacofoil TM Product (UPC 5023139217700) produced all the results. To form the spiral structure in this example, a strip of material was cut and a drop of glue 21 was deposited on one side, which could be on the foil or paper side 22, which was Fig.13 An example is shown in (a). Then, a cylindrical mold is placed on the long side of the strip at an angle greater than 45° so that the sheet is tightly wrapped around the mold to form a bonded rigid tube 25, as shown in FIG. Fig.13 (b) shows the foil on the inside and the paper on the outside. The angle of the mold bar relative to the edge of the thin film strip can be used to control the pitch and thus change the amount of overlap during assembly. It should be noted that there will be a double layer of foil at the overlap, which will cause electrical interaction with the magnetic field of the induction heater. If the thickness of the foil is less than the skin depth δ, then this overlap may cause double heating, which can be used to adjust the amount of heating in the filament.

[0154] As previously mentioned, according to a preferred embodiment, the consumable heated by the susceptor is cylindrical and has a constant diameter along its length. In order to optimize the heat transfer between the inner foil of the susceptor and the outer surface of the consumable, the consumable and the workpiece should be pressed tightly to provide efficient thermal contact and promote heat transfer between the two. In addition, by definition, the consumable should be removable from the susceptor. Since both parts (i.e., the susceptor and the consumable) are cylindrical, one way to achieve this is to make the susceptor slightly tapered, so that when the consumable is inserted into the susceptor, the consumable will initially contact the inner wall of the susceptor. When the consumable is fully pushed into the susceptor, the lower half of the consumable will contact the susceptor, thereby providing good thermal contact between the two. In practice, the internal taper angle required to achieve this is only a few degrees (about 1 to 10°), which can be achieved by the above-mentioned construction method and using a tapered mold instead of a constant diameter cylindrical mold.

[0155] Since the susceptor has a foil on the inner surface of the cylinder, the induction heating will only occur within the foil layer, which is in close contact with the outer layer of the consumable. Therefore, the paper that constitutes the outer layer of the thin sheet does not play a major role in the heat transfer process, which allows the paper to have a relatively wide range of thickness and be made of a variety of materials, although it should be able to be heated at temperatures above 200°C without damage. In some applications, if the paper is thicker, the efficiency of heat transfer between the foil and the consumable can be improved by preventing parasitic losses in high temperature environments.

[0156] 7. Correlate sensor temperature with resonant frequency shift

[0157] The above sections describe methods for measuring the ERF (resonant frequency) of a resonant circuit where the base values ​​of resistance, inductance, and capacitance vary due to changes in external parameters. Certain aspects of the present invention relate this variation to Figure 5 The temperature of the foil susceptor 16 is shown to be related to the temperature of the foil susceptor 16 when the component is replaced Figure 1 L t To form Figure 6 The circuit shown.

[0158] Assume that the wound foil is used as a single-turn secondary coil L of a magnetically coupled transformer. sec , L prim The current in L sec As the current flows around the circumference of the foil, the electrical energy is converted into heat and prim The temperature in the area of ​​magnetic coupling caused by the resulting oscillating magnetic field will rise. In this process, two effects occur:

[0159] i) The volume resistivity of the metal increases due to heat generation, which is defined as ρ( https: / / en.wikipedia.org / wiki / Electrical_resistivity_and_conductivity#Metals ), the resistance (ohmic resistance) of the foil will increase, resulting in increased losses and thus changing the energy delivered from the main circuit to the receptor.

[0160] ii) In addition, an increase in ρ increases the skin depth δ (see above). If the thickness of the foil > δ, this will have little effect on the overall system. On the contrary, if δ > foil thickness, the magnetic field generated by the coil will penetrate the foil, resulting in a weakened interaction between the two, which in turn changes the response of the frequency shift.

[0161] These two effects jointly affect the value of Zsec, making ERF(f res ) changes sufficiently to be measured by the method of the present invention. To illustrate this, a foil tube is integrated into the Figure 5 and Figure 6 The resonant circuit shown in the figure was placed in the chamber and temperature-controlled air was blown into it along its axis. Then, the ERF (f res ).

[0162] The core of the measuring circuit is a series LCR circuit consisting of a 22nF C0G ceramic capacitor and a 4.5 turns copper coil; the susceptor is a spiral coil of 7.5mm inner diameter made of the above-mentioned foil sheet, with a small overlap of a few millimeters along its length, and the foil is located on the inner surface, as Fig.13(b) and is coaxially mounted within the induction coil. Since the paper on the outer surface has a relatively high emissivity, this allows its temperature to be measured during the heating process using an infrared camera (IRC), in this case a FLIR ETS320 running the FLIR Tools PC software package. Advantageously, in this application, IRC provides a complementary non-contact / passive technology that does not alter the thermal load on the receptor in the first place, as would a wired temperature sensor such as a thermocouple or thermistor.

[0163] Fig.14 It shows that by scanning f d The value is measured by PDC for each f d Value V smpl , to obtain resonance comparison curves at 5 different temperatures (25°C, 50°C, 100°C, 150°C and 200°C). In these results, two features emerge from the data:

[0164] i) Peak frequency, which by definition corresponds to the ERF (f res ), which is 14KHz lower than room temperature at 200°C, corresponding to a range reduction of about 0.7%. Fig.14 The frontiers of the five data sets are shown clustered together on the right-hand side of the figure.

[0165] ii) Secondly, as the susceptor temperature increases, the peak height decreases, and the peak height decreases by 15% in the temperature range of 175°C. This decrease shows monotonic and gradual characteristics during the heating process of the sample.

[0166] Fig.15 Shows Fig.14 A more detailed profile of the 200 °C curve in Figure 1 is shown in Figure 2. Using these data, a quadratic polynomial least squares fit was performed and the fitted curve is represented by the superimposed dashed line with its coefficients shown at the bottom of the panel. Using the method outlined in Section 1, Equation 6, the ERF (f res ) is 2311KHz. Based on the comprehensive results, the quadratic fitting parameter analysis was performed on the five curves; Fig.16 Calculated values ​​of the estimated resonant frequency (ERF) of the system versus foil temperature are shown. The CD of the fitted quadratic polynomial regression was > 0.999 for the five data points generated, with an approximately linear shift corresponding to the ERF (f res ) shifts by about 0.013°K per Hz, indicating that over the temperature range, ERF(f res ) will decrease by about 800 Hz for every 10°K rise. In summary, the results show that within the temperature range of the materials, apparatus and methods detailed above, the temperature of a 12 μm foil can be measured by a heat generating system using the value of ERF.

[0167] Fig.17 It is illustrated how this parameter fitting method can be used to perform closed-loop control of the temperature profile in an induction heater. In this case, a cylindrical susceptor with a simple axial (non-helical) overlap and a diameter of 7.5 mm, consisting of the above foil sheets, is used with Figure 5 The temperature-frequency response of the assembly was then characterized by passing heated air through the susceptor axis at different temperatures. After allowing the system sufficient time to reach equilibrium, the external surface temperature was measured at multiple fixed temperatures using IRC. Fig.14 , 15 16 and the related description, a fit and parameterization was performed to derive the relationship between the IRC temperature and the frequency response of the induction heater / susceptor combination.

[0168] Then, the HHB is excited by a 100 Hz PWM waveform to induce heating of the sensor, which in turn is controlled by adjusting the power of the heater through the PID-C method. Fig.12 As shown in the figure, since the leading edge of each pulse in the PWM signal contains the res ), so the temperature is also monitored at a frequency of 100 Hz.

[0169] Fig.17 Three sets of traces are shown:

[0170] -ERF(f res ) behavior in the heating cycle (dotted line);

[0171] - calculated susceptor temperature (dashed line);

[0172] - Temperature measured independently using IRC (solid line).

[0173] By using ERF(f res ) is converted into foil temperature to increase and control the temperature and stabilize the following temperatures at five increasing levels and maintain for 15 seconds each: 50°C, 75°C, 100°C, 125°C and 150°C.

[0174] from Fig.17 It can be seen that PID-C maintains ERF (f res ) and from this it is inferred that the calculated susceptor temperature is stable. Therefore, in order to be able to measure the performance of the closed-loop temperature control loop, the output variance of the IRC is analyzed as an independent (external) variable. For the five temperature plateaus, at 7.5 seconds, 22.5 seconds, 37.5 seconds, 52.5 seconds and 67.5 seconds, the following values ​​are obtained:

[0175] Calculated susceptor temperature (°C)Measured IRC temperature (°C)

[0176]

[0177] The IRC readings indicate that the method according to an embodiment of the present invention is able to stabilize the susceptor temperature to within <±1°C relative and within ±3°C absolute, which is sufficient for generating and controlling the inhalable aerosol at the operating temperature of the device (>150°C).

[0178] 9. Correlate sensor temperature with resistance change

[0179] from Fig.14 It can be seen from the data that as the foil temperature increases, the peak amplitude decreases. Fig.18 This paper explains the origin of the peak voltage near resonance in a simple series resonant circuit and how it is related to the resistance of the circuit (see https: / / industrial.panasonic.com / ww / ss / technical / n3 ). In ERF(f res ), the reactive elements of the capacitor and inductor cancel each other, leaving only the inherent ohmic resistance of the circuit as the impedance of the circuit. On either side of resonance, the residual reactive elements begin to increase, so that the peak amplitude of the resonance curve is dominated by R t In the transition region where the reactive and ohmic impedances are equal in magnitude to Rt, the impedance starts to follow the behavior of the reactive impedance, resulting in a flat terrace-like characteristic.

[0180] from Fig.18 R can be defined t f res The impedance measurement at Z t (=R t ) increases, the peak amplitude shifts to lower values. Fig.15 The equation shown in res ) is substituted into the value, and the corresponding resonant circuit admittance Z is obtained t Since the induction heating device is formed as a transformer, with a primary drive circuit (p) and a secondary foil susceptor (s) placed in or around the consumable, it can be seen from Appendix A below that Z t =Z p +Z s , where Z s is the impedance of the foil. In addition, in ERF(f res ) s There is no reactive element, equivalent to the ohmic resistance of the foil. Therefore, as the susceptor heats up, the resistance of the foil increases due to the temperature coefficient of aluminum, ρ (= 0.43% ° K -1 ) increases, increasing Z s , thereby reducing I t(and V smpl ).

[0181] Back to Fig.14 , the resonance curves at higher temperatures show a gradual decrease in their peak voltages, and, using the above described positioning ERF (f res ) method, ERF(f res ) and then substitute Fig.15 A polynomial of the form shown in .

[0182] Fig.19 Shows Fig.14 The relationship between the peak amplitude and temperature of the data is shown in Figure 2. Fig.19 The equation shown in was fitted with a polynomial, CD>0.99, allowing the susceptor temperature to be determined by measuring the peak amplitude of the resonance curve, which can be used as the measured variable in the closed-loop control algorithm, such as Fig.17 shown.

[0183] From another perspective, since this method changes the resistance by about 15% over the temperature range, Fig.16 The frequency data plotted in varies by about 1%, so this parameter may be able to more finely determine the susceptor temperature due to its greater variation over this defined temperature range. Also, since it relies on locating the peak of a relatively flat curve, simpler peak finding methods than the LSA algorithm can be used, such as taking the voltage data on the table and using a simple sorting algorithm to locate the peak and use that as the peak amplitude. More complex algorithms can also be used, such as averaging the top value and multiple adjacent points to prevent a single highest point in the data set due to noise.

[0184] As mentioned above, the frequency is determined by V sw The magnitude of Fig.14 The voltage step shown in Figure 1 is different because it uses V smpl The value is directly related to V sw Therefore, the cell voltage should be measured simultaneously with each PDC reading and the latter normalized by the former to produce a dimensionless constant where the PDC output is expressed as a fraction of the cell voltage.

[0185] As described above, the frequency shift measurement of this method becomes more difficult as the foil becomes thicker for a variety of reasons. In contrast, the resistance-based method according to an embodiment of the present invention has advantages for aluminum foil (which has a thickness of multiple times the skin depth) and most common metals such as nickel, steel (including stainless steel), etc. (thickness> 10 times the skin depth at 2MHz).

[0186] 10. Advantages of Practical Induction Heating System Methodology

[0187] Using the methods of various aspects of the present invention to detect and quantify the shift in ERF and / or response amplitude, the inductive heater device will be able to simultaneously heat the susceptor and detect the temperature reached through the heating circuit, thereby eliminating the need for external sensors such as thermocouples or thermistors. This brings many advantages, as installing a separate sensor on the susceptor has the following disadvantages:

[0188] - The physical size and heat capacity of the sensor may disturb the reading temperature because the foil it is bonded to is only a few microns thick;

[0189] - Bonding a standalone sensor to a 10μm foil is a difficult process because the foil tends to tear if any mechanical load is applied through the sensor leads;

[0190] -If welding is used, care must be taken to avoid weakening and / or oxidation of the metal susceptor, which may damage the susceptor material;

[0191] - If glue is used, the glue should not add thermal mass and / or thermally hinder heat transfer from the susceptor to the sensor;

[0192] - By definition, the sensor needs to be placed where the magnetic field generated by the coil is the strongest. Therefore, during its service life, the sensor will experience rapid thermal shocks exceeding 200°C thousands of times, which will tend to degrade its performance and may cause the sensor or its thermal connection to fail, making the device inoperable;

[0193] - From the latter point of view, the electronics of the sensor will be subject to large currents induced by the strong magnetic field, which may cause problems for the sensor or the electronics used to read the sensor.

[0194] The advantage of using a frequency / amplitude temperature tuning system is that the induction system always operates near the actual resonant frequency (ARF) in all use cases; from the device being idle waiting to be activated, to intense heating near operating temperature. Without the optimization system of the various aspects of the present invention, if the device is tuned to the ERF (f res ),like Figure 2 (a), it will maximize and be consistent with V sw Synchronize and optimize R t Once activated, as the system heats up to its operating temperature of approximately 200°C, Fig.14 The data shown show that ERF (f res ) will move enough so that f d No longer overlap, produce Figure 2 The characteristic shown in (c), namely f d >f res , making I t Hysteresis and V swThe amplitude is reduced, resulting in less heat dissipated in the susceptor and more heat dissipated in the coil and electronics that make up the main drive circuit. By using the system of various aspects of the present invention, the actual resonant frequency (ARF) is estimated in real time and f is adjusted accordingly. d , meaning the system always operates at close to optimal energy transfer efficiency, extending the overall life of the battery and device.

[0195] This is in contrast to the case where the ERF (f) of a resonant circuit is directly measured using an external signal. res ), then using a digital counter to accumulate pulse response circuits over a fixed period of time, or using an interval timer to measure the gaps between individual pulses, the method of the present invention can measure about 20 DC values, each of which is generated in response to an external programmable clock signal and is independent of the induction heater electronics or resonant circuit. These data points are then mathematically analyzed to generate an estimate of the resonant frequency of the system. Therefore, the method of the present invention embodiment is different from many previous methods.

[0196] The method according to the invention cannot directly measure the actual resonant frequency (ARF), but can only estimate it for the following reasons:

[0197] - In essence, the least squares approximation method cannot produce an exact value of the ARF from a set of data, as its name suggests. For example, this technique will take 10 data points that will be used to generate the value that best represents the ERF (f res ) will almost always be different from any of the 10 readings collected.

[0198] -Secondly, the measurement process heats the receptor slightly as data points are collected, so when the 10th point is measured, the receptor will be slightly warmer than when the 1st point was measured. This effect can slightly skew the data fed into the LSA algorithm, causing systematic perturbations in the estimate. This effect can be minimized by using as few pulses as possible on the fastest possible time base to generate the final voltage level on the PDC.

[0199] 11. Use temperature sensing systems in alternative heater layouts

[0200] The above concerns the sensing and control of temperature using a helical primary coil and a hollow foil susceptor in a coaxial induction heating system. Since the method is derived from the general properties of induction heating systems, it can be understood that the method is applicable to many other induction heater geometries, some examples are as follows:

[0201] Planar Excitation Coil / Flat Plate Susceptor: The construction and operation of household induction cookers are well documented in the technical literature. In this application, the coil is configured as a flat (flat) multi-turn spiral separated from the flat base of the cooker by a sheet of glass or ceramic. When the two are in close proximity, the magnetic field is well coupled into the cooker base, facilitating efficient energy transfer. Due to the rapid growth in the use of induction cookers, there have been many attempts to design a non-contact temperature measurement method that utilizes the changes in the impedance of the inductive heating circuit driven by the changes in the impedance of the susceptor during heating and cooling. For example, Franco et al. ( https: / / doi.org / 10.1109 / JSEN.2011.2167226 ) discusses a method of using a second induction coil in addition to the heating coil to "probe" the impedance of a planar cookware base, with the measured frequency characteristics then used as a proxy for base temperature. Since embodiments of the present invention detect such changes by periodically reusing the induction heating coil as an excitation probe to measure the system frequency, no dedicated sensing coil is required.

[0202] Since the system of Franco et al. measures the same parameter (system frequency) as a proxy for susceptor temperature, albeit using different equipment, the method of the present invention is also applicable to systems in which a flat (planar) susceptor is excited (e.g., heated) by a flat (planar) induction coil. In these cases, the susceptor can vary from thin metal foil to thick metal plate, ranging from magnetically inert aluminum to materials containing ferromagnetic elements and their alloys.

[0203] Compatible susceptor form: In order for induction heating to be feasible, the conductive metal body must intercept the limited magnetic flux generated by the induction heating coil. In addition, in order to achieve efficient coupling / energy transfer between the two, the susceptor should be able to intercept the magnetic flux generated by the I prim Since the oscillating magnetic field can only penetrate the outer surface of the conductive solid to a depth determined by the skin depth δ, any interaction with materials at depths > δ is negligible. Therefore, if Figure 5 The tubular foil susceptor shown is made of aluminum foil with a thickness of 60 μm, and the system is operated at 2 MHz I prim excitation, nearly all of the energy will be absorbed by the susceptor. If the foil susceptor is replaced with a solid aluminum rod of the same diameter, the amount of energy absorbed will be roughly the same, but the rod will heat up more slowly due to the added thermal mass.

[0204] In a preferred embodiment, the structure of a feasible susceptor should meet the following criteria:

[0205] A: It has electrical conductivity, which is consistent with the conductivity range of metals and semiconductors;

[0206] B: intercepts most of the available magnetic flux;

[0207] C: Has the lowest possible heat storage capacity to maximize the rate of temperature rise.

[0208] Thus, the susceptor may be of any cross-sectional shape, such as a rod, tube, cone, wedge, sphere, cube, longitudinal plate, or a modified version of any of the above.

[0209] Furthermore, the method of determining temperature by using the resonant frequency of the system as a measurement proxy can be applied to these types of susceptors as long as the electrical criteria (A and B) described above are met.

Claims

1. A method comprising: (a) providing an LCR circuit including an inductor element and a capacitor; (b) applying a voltage to the LCR circuit, wherein the applied voltage induces a response between a capacitor and an inductor element of the LCR circuit; (c) generating an output signal of the response; (d) selecting a value representative of an estimated resonant frequency (ERF) based on the output signal; The method further comprises: (e) setting the frequency of the applied voltage to the estimated resonant frequency in step (d); (f) Repeat steps (b) to (e).

2. The method of claim 1, wherein the output signal comprises a voltage signal, the method further comprising determining a maximum value of the voltage signal, wherein: The estimated resonant frequency is selected corresponding to the determined maximum value.

3. A method according to claim 1 or 2, wherein the step of determining a maximum value comprises fitting a polynomial function to the output signal data.

4. The method of claim 3 further comprising minimizing the order of the polynomial function using a coefficient of determination. The method of claim 3 , wherein determining the maximum value comprises using a least squares approximation method.

6. The method of claim 1 , wherein in a second iteration, the steps are repeated to set the frequency to a value higher than the estimated resonant frequency in the first iteration of repeating steps (a) to (e) in claim 1 , and in a third iteration, the steps are repeated to set the frequency to a value lower than the estimated resonant frequency in the first iteration.

7. The method according to any of the preceding claims, wherein the estimated resonant frequency is selected using a peak detection circuit PDC, the PDC comprising a PDC capacitor, the method further comprising the steps of: charging the PDC capacitor during a sampling time; measuring a plurality of voltages in the output signal and calculating an average value of the plurality of voltages during the sampling time; discharging the PDC capacitor; The frequency of the applied voltage is set to the estimated resonant frequency.

8. The method of any preceding claim, further comprising inductively heating a susceptor using the inductive element to aerosolize a substance in a heating mode.

9. A method according to any preceding claim, wherein the output signal is used to provide a temperature measurement of the susceptor.

10. The method of claim 9, wherein providing a temperature measurement comprises determining a resistance of the susceptor.

11. The method of claim 1, further comprising determining a maximum value of the estimated resonant frequency, and if the maximum value of the estimated resonant frequency is within a predetermined range, setting the frequency of the applied pulse to the maximum value.

12. The method according to claim 11, further comprising repeating the steps of claim 11 until the set frequency is within the predetermined range.

13. A method according to any preceding claim, further comprising determining a change in the estimated resonant frequency.

14. A method according to any preceding claim, wherein selecting the estimated resonant frequency comprises determining a change in circuit resistance, the method further comprising determining a susceptor temperature.

15. A method for controlling the temperature of a susceptor, comprising the method according to any one of the preceding claims 1 to 14.

16. A system comprising a voltage generator for applying a voltage to an LCR circuit, the LCR circuit comprising an inductive element and a capacitor, wherein the applied voltage comprises a response between the capacitor and the inductive element of the LCR circuit, the system further comprising an output circuit for generating an output signal of the response; and a processor configured to: (g) selecting a value representative of an estimated resonant frequency based on the output signal; (h) setting the frequency of the applied voltage to the estimated resonant frequency ERF based on step (g); and (i) Repeat steps (g) to (h).

17. The system of claim 16, wherein the voltage generator comprises a switching device that generates pulses by switching between a positive voltage source and a negative voltage source.

18. The system of claim 17, wherein the switching device comprises an H-bridge. The system of claim 18 , wherein the H-bridge is a half H-bridge (HHB).

20. The system of any one of claims 17 to 19, wherein the voltage generator comprises a resistor and a diode.

21. The system of any one of claims 17 to 20, wherein the inductive element comprises a flat ribbon conductor.

22. The system of any one of claims 17 to 20, further comprising a susceptor capable of being heated by the inductive element.

23. The system of claim 22, wherein the susceptor is a spiral winding.

24. The system of claim 22 or 23, wherein the susceptor forms a cone.

25. The system of any one of claims 22 to 24, further comprising an electrical connection surrounding the susceptor.

26. The system of claim 25, wherein the electrical connection is formed by a capacitor.

27. The system according to any one of claims 16 to 26, further comprising a peak detection circuit (PDC) for determining a voltage maximum from a plurality of voltages in the output signal.

28. The system of claim 27, wherein the PDC comprises a unity gain operational amplifier OA.

29. The system of claim 27 or 28, wherein the PDC comprises a switch to make impedance switchable between ˜0 and ∞Ω.

30. The system of claim 27 or 28, wherein the PDC comprises a ceramic capacitor.

31. The system of claim 30, wherein the ceramic capacitor is made of COG / NPO material.

32. The system of claims 27 to 31 , wherein the PDC comprises a bipolar field effect transistor.

33. The system of any one of claims 16 to 32, further comprising a current sensor for measuring a current flowing through the inductive element.

34. The system of any one of claims 16 to 32, further comprising a control module for determining the system performance based on the output signal.

35. Computer readable instructions which, when executed by a computing device, cause the computing device to perform the method according to any one of the preceding claims 1 to 15.

36. An aerosol providing system for generating an aerosol from an aerosolizable substance, the aerosol providing system comprising a system according to any one of the preceding claims 16 to 34, wherein the aerosol providing system is configured to perform an operation in response to an output signal received from an output circuit.

Images