Patent Application
20090141065
The present disclosure relates generally to synthetic jet actuators, and more particularly to methods and devices for controlling diaphragm displacement in synthetic jet actuators.
A variety of thermal management devices are known to the art, including conventional fan based systems, piezoelectric systems, and synthetic jet actuators. The latter type of system has emerged as a highly efficient and versatile solution where thermal management is required at the local level. Frequently, synthetic jet actuators are utilized in conjunction with a conventional fan based system. In such hybrid systems, the fan based system provides a global flow of fluid through the device being cooled, and the synthetic jet ejectors provide localized cooling for hot spots and also augment the global flow of fluid through the device by perturbing boundary layers.
Various examples of synthetic jet actuators are known to the art. Some examples include those disclosed in U.S. 20070141453 (Mahalingam et al.) entitled “Thermal Management of Batteries using Synthetic Jets”; U.S. 20070127210 (Mahalingam et al.), entitled “Thermal Management System for Distributed Heat Sources”; 20070119575 (Glezer et al.), entitled “Synthetic Jet Heat Pipe Thermal Management System”; 20070119573 (Mahalingam et al.), entitled “Synthetic Jet Ejector for the Thermal Management of PCI Cards”; 20070096118 (Mahalingam et al.), entitled “Synthetic Jet Cooling System for LED Module”; 20070081027 (Beltran et al.), entitled “Acoustic Resonator for Synthetic Jet Generation for Thermal Management”; and 20070023169 (Mahalingam et al.), entitled “Synthetic Jet Ejector for Augmentation of Pumped Liquid Loop Cooling and Enhancement of Pool and Flow Boiling”.
In one aspect, a method for calibrating a synthetic jet ejector is provided which comprises (a) taking a first measurement DCR0 of the dc resistance of the coil; (b) adjusting the actuator drive voltage Vd to achieve a desired maximum displacement dmax1 at a frequency f1; (c) measuring the input current Iin and input voltage Vin; (d) calculating the Back Electromotive Force BEMF, wherein BEMF=Vin−Iin*DCR; and (e) storing the calculated value of BEMF in a memory device associated with the synthetic jet actuator.
In another aspect, a method for calibrating a synthetic jet ejector is provided which comprises (a) providing a synthetic jet ejector equipped with a coil, wherein the coil causes a diaphragm to vibrate about a first axis which is perpendicular to a major surface of the diaphragm; (b) applying a periodic force such that the diaphragm is deflected from a resting position to a maximum displacement d0 along the first axis, wherein d0 is equal to the desired displacement of the diaphragm during operation of the synthetic jet ejector; and (c) measuring the BEMF voltage across the coil.
In a further aspect, a method for determining the BEMF in a coil of a synthetic jet ejector is provided which comprises (a) providing a synthetic jet ejector equipped with a first coil, wherein the first coil causes a diaphragm to vibrate about a first axis which is perpendicular to a major surface of the diaphragm; (b) providing a second coil; and (c) using the second coil to determine BEMF.
In still another aspect, a method for determining BEMF in a synthetic jet ejector having coupled first and second actuators is provided. The method comprises (a) deactivating the first actuator while operating the second actuator, thereby placing the synthetic jet ejector into a first operational state; (b) determining the Back EMF (BEMF1) of the first actuator while the synthetic jet ejector is in the first operational state; (c) deactivating the second actuator while operating the first actuator, thereby placing the synthetic jet ejector into a second operational state; and (d) determining the Back EMF (BEMF2) of the second actuator while the synthetic jet ejector is in the second operational state.
In a further aspect, a method is provided for determining DC resistance (DCR) in an actuator coil for a synthetic jet ejector while the ejector is operating. In accordance with the method, DCR is determined by from dynamic impedance measurements at one or more frequencies outside of the normal operating range. For example, DCR may be determined at 10 Hz, and preferably, from dynamic impedance measurements at both 10 Hz and 20 Hz. Even more preferably, DCR is determined in accordance with the equation DCR=Z(10 Hz)−(Z(10 Hz)−Z(20 Hz))/3.
In still another aspect, a method for monitoring resonance frequency in a synthetic jet ejector equipped with an actuator coil is provided. In accordance with the method, the phase of input impedance in the actuator coil is monitored. The resonance frequency is then determined by identifying the point at which the phase of the input impedance changes sign, and preferably, as the point at which the phase of the input impedance changes from positive to negative.
In a further aspect, a method for monitoring the phase relationship between two or more actuators in a multiple actuator system is provided. In accordance with the method, the phase of the calculated Back EMF signal of each actuator is monitored by recording the location of the negative-going zero-crossing of the waveform. The phase of each actuator drive signal is then modified such that the zero-crossings of all Back EMF signals occur simultaneously, thus matching the phase of all actuators within the system.
In yet another aspect, a synthetic jet ejector is provided which comprises (a) a diaphragm which undergoes displacements along an axis perpendicular to the surface of the diaphragm in response to a magnetic field; and (b) a sensor which senses the displacement of the diaphragm along the axis; wherein the diaphragm is driven by a magnetic field, and wherein the synthetic jet ejector is adapted to adjust the magnetic field in response to the sensed displacement of the diaphragm. In some embodiments, the sensor may comprise a capacitive plate, the magnetic field may be generated at least partially by a magnetic coil, and the plate may be capacitively coupled to the magnetic coil. In other embodiments, the sensor may be an optical sensor, and the synthetic jet ejector may further comprise a diode which is in optical communication with the sensor. In some such embodiments, the diode may be in optical communication with the sensor by way of an optical path, and the sensor may operate by sensing the degree to which the optical path is blocked. In other such embodiments, the diode may be in optical communication with the sensor by way of an optical path which includes a surface of the diaphragm, and wherein the sensor may operate by measuring the angle of incidence and the angle of reflection of radiation emitted by the diode which impinges on the diaphragm.
While the aforementioned synthetic jet actuators represent notable advances in the art, further improvements are still required in synthetic jet actuator technology. For example, many synthetic jet actuators are currently designed to operate with predetermined diaphragm displacements. These predetermined displacements typically do not take into account variations in environmental factors such as temperature, nor do they account for deviations in operational frequency or other such parameters which may occur as the device ages. Moreover, the predetermined displacements are typically based on the averages of various engineering parameters, and hence do not reflect deviations within manufacturing tolerances for a specific device.
Consequently, many synthetic jet actuators function at diaphragm displacements which are suboptimal in terms of the prevailing operational characteristics of the device at a given time and in terms of the efficiency at which the device can dissipate heat. Many of these devices also operate at diaphragm displacements which are suboptimal in terms of power consumption. In extreme cases, the actuator may be driven at diaphragm displacements which exceed maximum safe ranges, with the result that the diaphragm may come into contact with adjacent surfaces and may rupture.
There is thus a need in the art for synthetic jet actuators which overcome these issues. In particular, there is a need in the art for synthetic jet ejectors whose operating characteristics may be periodically or continually modified or optimized. These and other needs may be met by the devices and methodologies disclosed herein and hereinafter described.
It has now been found that the aforementioned needs may be met by controlling diaphragm displacement, preferably by periodically adjusting diaphragm velocity. This approach provides a simple and easy means by which the synthetic jet actuator may be recalibrated during subsequent uses (or during a particular use) so that the device will operate at optimum performance and/or at minimum energy consumption. This approach allows the operational parameters of the synthetic jet actuator to be modified to account for differences due to aging or the environment in which the device is operating in.
The methodologies disclosed herein may be better understood with respect to the factors controlling the operation of a synthetic jet actuator. In many embodiments, the input voltage (Vin) of a moving coil actuator is equal to the sum of the voltage drop (Vdcr) across the DC resistance of the coil and the Back Electromotive Force (BEMF). This relationship is expressed by EQUATION 1:
V
in
=V
dcr
+B
EMF (EQUATION 1)
Vdcr may itself be expressed as a function of the current input to the coil (Iin) and the voltage across the DC resistance of the actuator (Vdcr), as shown by EQUATION 2:
V
dcr
=I
in
*DCR (EQUATION 2)
Also, BEMF may be expressed as a function of the magnetic field in the coil region (B), the length of the coil (L), and the velocity of the coil. This relationship is expressed by EQUATION 3:
B
EMF
=B*L*v (EQUATION 3)
The actuator diaphragm displacement is related to velocity by the simple derivative shown in EQUATION 4:
v(t)=dx/dt (EQUATION 4)
If a synthetic jet actuator is driven with sinusoidal applied voltage, then the following relation holds:
V
in
=A*sin(wt) (EQUATION 5)
wherein:
A=peak input voltage;
w=radian frequency; and
t=time.
Hence, velocity may be derived by applying the derivative of EQUATION 4 to EQUATION 5:
V(t)=A*w*cos(wt) (EQUATION 6)
Ignoring phase, velocity may then be expressed as:
v=w*x (EQUATION 7)
It will thus be appreciated that displacement may be controlled by controlling velocity. Consequently, for sinusoidal inputs, velocity is a linear function of frequency. It follows that
B
EMF
=V
in
−I
in
*DCR=B*L*v (EQUATION 8)
In light of the foregoing, and with reference to
Referring now to
If it is desired to operate at another frequency, then the BEMFT may be adjusted (507) such that
B
EMFnew
=B
EMFT
*w
new
/w
factory (EQUATION 9)
where
BEMFnew=the BEMF target at the new frequency;
wnew=the new frequency;
wfactory=the frequency at which factory calibration was performed.
If it is desired to operate at a different displacement, then the BEMFT should be adjusted (509) such that
B
EMFnew
=B
EMFT
*x
new
/x
factory (EQUATION 10)
where
BEMFnew=the BEMF target at the new displacement;
xnew=the new displacement; and
xfactory=the displacement at which factory calibration was performed.
After the foregoing adjustments, the DCR may then be measured (511) continuously or intermittently and BEMF may be monitored to ensure that it stays at BEMFT. The drive voltage may be adjusted as necessary to keep BEMF=BEMFT.
The foregoing control algorithm may be implemented with a sinusoidal drive circuit, input voltage and current measurement, and a controller comprised of digital and/or analog circuits that computes BEMF and adjusts actuator drive voltage and frequency. It will also be appreciated that the control algorithm will also work for other periodic waveforms aside from sinusoidal waveforms, and may even work for arbitrary waveforms with minimal information about the waveform known, as long as the relationships between velocity and displacement are defined and are either known or approximated, and as long as the BEMF value is derived and treated properly.
Additionally, it may be desirable in some thermal management systems to operate the actuator at or near system resonance. This will ensure that the thermal management system operates at its point of maximum power efficiency. This may be accomplished, for example, in the same controller by finding the frequency of minimum power consumption for a given displacement. This frequency shifts with time and temperature. As the resonance is tracked, the displacement is held constant with the control algorithm as described above.
Various software programs may be used to implement the foregoing methodology. APPENDIX A contains a particular, non-limiting example of an algorithm that may be used for this purpose.
Some of the foregoing methodologies utilize BEMF to detect a quantity which is proportional to diaphragm displacement. In some embodiments of the methodologies described herein, BEMF may be detected through the use of a second coil which is wound around the coil former of the synthetic jet actuator. This second coil may be co-wound with the motor coil, disposed next to the motor coil, or placed on top of or around the motor coil. The voltage, present at the detection coil while the actuator is moving, is the pure BEMF signal which can then be processed for control purposes.
In some embodiments of this type, one can use two or more coils placed next to the motor coil to detect offsets in the motion of the coil or diaphragm. Alternatively, the BEMF signal acquired from the driving coil as described earlier can be combined with the detection coil signal to detect abnormalities in the motion of the coil or diaphragm, to detect offsets, or for other such purposes.
The embodiments described herein which utilized BEMF as a quantity which is proportional to diaphragm displacement typically require a calibration procedure, since BEMF typically varies from device to device. This calibration procedure may involve the use of a system having laser displacement sensors to measure diaphragm displacement and to adjust the BEMF target values accordingly to achieve the desired stroke length of the actuator. In some embodiments, the actuator may be shaken along its axis of motion such that the inertia of the diaphragm will cause it to deflect from its rest position so as to create the desired amplitude on the device to be calibrated. The actuator then acts as a generator and will produce the pure BEMF voltage on its leads. This voltage may then be measured and used as a reference.
In a variation of the foregoing methodology, rather than shaking the actuator along its axis of motion, air or another suitable fluid may be used to displace the diaphragm of the actuator. This may be accomplished, for example, by using an audio speaker or driven piston of appropriate size to generate a fluid pressure that varies over time sinusoidally and which is of the desired frequency, and creates the desired amplitude, on the diaphragm of the actuator to be calibrated. When the pressure wave is applied to the diaphragm, the actuator acts as a generator and will produce the pure BEMF voltage (VEMF) on its leads. This VEMF may then be measured and used as a reference. Depending on the calibration method utilized, this method would also allow the actuator to be calibrated to a certain air flow. Moreover, this method does not require optical access to the diaphragm for a laser measurement. In order to obtain feedback of the diaphragm displacement, fiber optics (or conventional optics with appropriate image acquisition systems) may be utilized to look at the coil or other moving parts (this may be accomplished by looking through the nozzles of the device). A gauge print may be provided on the coil or other moving parts of the device for this purpose.
As described herein, BEMF gives an indication of diaphragm displacement, and can be used in a control system to maintain specified displacement while the surround-diaphragm “spring constant” changes with temperature or age. In such applications, the control system typically reads the current BEMF, and then adjusts the drive to move back to the specified BEMF and displacement.
However, this procedure can become complicated when a synthetic jet ejector is driven by two or more actuators. This may occur, for example, when two or more actuators share a common air cavity, as may be the case, for example, in a dual actuator housing assembly in which the backsides of the actuators face each other, and in which the actuators drive air from the same cavity out of the jet ports. In such applications, BEMF may be determined by selectively switching off the drive to one of the actuators while continuing to drive the other actuator. The BEMF associated with the deactivated actuator may then be measured, and the procedure may be reversed to determine the BEMF associated with the other actuator.
As a specific example, the situation of a synthetic jet ejector equipped with first and second actuators may be considered. In this case, the drive to the first actuator may be switched off, while operation of the second actuator is maintained. The fluid in the common cavity housing the first and second actuators will couple the first and second actuators to each other. Consequently, the diaphragm of the first actuator will move, even though it is not being electrically driven. The motion of the drive coil of the first actuator through its B-field will generate BEMF1, which may be measured with circuitry which is simpler than that required to drive the actuator and measure BEMF values at the same time. In an analogous manner, the second actuator may be deactivated while the first actuator is driven, thus allowing BEMF2 to be determined.
The measured values of BEMF1 and BEMF2 can be related to actuator properties and drive corrections which may be applied as described above. After the periodic measurements are completed, the actuators are returned to normal operation, with both actuators being driven with corrected drive conditions.
When BEMF is used as a control parameter, it is important to get a good measurement on this variable. Previously, BEMF as determined by BEMF=V−I*DCR was strongly dependent on the DCR measurement performed on a resting actuator, with three implications: (1) small disturbances negatively affected the DC measurement; (2) the measurement was electronically challenging; and (3) the process was acoustically disturbing for the listener. These issues may be addressed with the following methodology.
DC resistance can be obtained by fitting the impedence curve as a function of frequency. In many applications, it is either impossible or impractical to do full frequency chirps in order to obtain the entire curve. However, it has been found that a good approximation of the DCR resistance may be obtained by using the 10 Hz impedance number. It can be shown that this number is less than 1% off of the target value.
This approach may be further improved by using two measurements, one at 10 Hz and one at 20 Hz, and then using a parabolic fit to extrapolate the 0 Hz (=DC) resistance:
DCR=Z(10 Hz)−(Z(110 Hz)−Z(20 Hz))/3 (EQUATION 18)
This approach is found to improve the error in resistance to a few mOhms. This improvement may also be used in manufacturing testing and in other tests that require knowledge of the DC resistance.
The choice of frequency or frequency pairs will be governed by the targeted precision of the measurement, acoustic considerations and equipment capabilities. For this method to be precise, it should only be used significantly below resonance, e.g., the Hz/20 Hz pair may be suitable for parts with resonance frequency at or above 50 Hz.
TABLE 1 shows results obtained in a resistance measurement comparison using a dynamic versus a handheld DVM method. These results are depicted graphically in
In order to utilize the dynamic measurement method for a DCR, it is necessary to superimpose a low frequency waveform onto the normal frequency voltage drive signal. If the low frequency wavelength is chosen to be an even multiple M of the normal frequency drive signal, this can be achieved by decreasing the drive signal by a constant value for M/2 normal frequency cycles, and then increasing the drive signal by the same constant value for M/2 cycles.
Measurement of the response to the low frequency waveform can be accomplished by sampling the voltage and current values multiple times during the normal frequency cycles and separately accumulating totals for the cycles when the low frequency waveform is high and low. By calculating the difference between these accumulates, and dividing voltage by current, a measurement of the low frequency signal can be extracted and used to calculate the DCR value:
DCR=(VSUMHigh−VSUMLOW)/(ISUMHigh−ISUMLOW) (EQUATION 19)
In order to maximize synthetic jet performance, it is important to operate the air-moving actuators at the maximum possible displacement and frequency. Power consumption of a synthetic jet operating at a given displacement is a strong function of frequency.
Methods are provided herein by which the resonant frequency may be found and tracked so that, as temperature and operating conditions change, the system can always be operated at the resonant frequency. This may be achieved by utilizing the rapid change of input impedance phase that occurs at the resonant frequency.
It is important as resonance is tracked to make appropriate adjustments to the BEMF target (BEMFT) which the displacement control-loop is using to maintain displacement. The target is proportional to frequency, so the target must be increased or decreased as the frequency is varied.
It is also important to implement voltage, current and power limiting. The power amplifier driving the cooler should not be operated beyond its limits. If this occurs, displacement control will be lost, and/or the amplifier and/or cooler may be damaged. Limiting can be implemented in the control software by reducing the drive voltage when limit conditions are detected. This will typically happen at lower temperatures (when the cooler resonance is higher in frequency, and when the actuator suspension is stiffer/more lossy).
With reference to
One difficulty encountered in measuring BEMF values is that synthetic jet actuators behave in a non-ideal manner. For example, it might be thought that the peak-to-peak displacement of the diaphragm as a function of frequency would be essentially linear and that, accordingly, the device could be calibrated for any displacement, thus allowing the BEMF to be scaled to achieve any other displacement. In practice, however, it has been found that the relationship between peak-to-peak displacement of the diaphragm and frequency is not linear. Without wishing to be bound by theory, this result is believed to arise, in part, from the non-ideality of the motor which drives the actuator. In particular, this result is believed to arise, in part, from the fact that the magnetic field associated with the moving coil of the motor does not behave in a linear fashion. Thus, as the displacement of the diaphragm increases, the magnetic force being applied to the coil does not increase in a linear fashion. Therefore, the BEMF falls off as the displacement continues to increase.
To compensate for this problem, a calibration algorithm is preferably utilized in the methodologies described herein which utilize a nonlinear (and preferably a polynomial) curve fitting technique. In such as approach, data is sampled at several points and is fitted with a polynomial curve. Typically, a second order polynomial is used for this purpose, although in some applications, higher order polynomials may be utilized.
Another problem encountered in the calibration process relates to the relationship between BEMF and voltage. In particular, for a particular BEMF target having an associated displacement, there may be limitations on the voltage (imposed by the electronics of the device) that may be utilized to achieve that target. For example, in a 5V system, voltage may be converted to a voltage differential so that, in theory, 10V can be used to achieve a BEMF target. However, due to losses at the switches of the device and other such factors, only 8V may be available to achieve the BEMF target. On the other hand, as temperature increases, the diaphragm softens, thus making it easier to drive it to larger displacements. Consequently, as temperature increases, displacement and BEMF is affected, thus giving rise to different curves. Consequently, the BEMF target can be achieved at a lower voltage.
While it would be desirable in many cases to calibrate to a large BEMF, this frequently cannot be done in practice. Consequently, in such cases, measurements are taken at other points, and the data is extrapolated to the point of interest. An appropriate curve fit (e.g., a polynomial curve fit) is utilized for this purpose which takes into account the non-ideality of the motor. Since the relationship between BEMF and displacement is typically not temperature dependent (or is only weakly temperature dependent), at a particular drive voltage, since displacement increases with temperature, BEMF also increases with temperature. Consequently, by determining the correct BEMF target, the correct displacement will be achieved. The voltage required to achieve that displacement will typically drop with temperature.
As discussed above, BEMF can be easily measured with electronic circuitry to control diaphragm velocity which, in turn, controls diaphragm displacement. BEMF is more typically associated with rotational motors, rather than the type of electromagnetic actuators employed in the present devices. This is because the actuators which are preferably utilized in the synthetic jet ejectors described herein are essentially audio speakers, and it is typically not necessary to control diaphragm displacement in audio speakers. By contrast, in a synthetic jet ejector, it is typically desirable to control displacement so as to achieve maximum air flow. In particular, it is desirable to move the diaphragm as close to the actuator housing as possible without actually hitting the housing. This may be achieved by using BEMF to control diaphragm displacement.
In some embodiments of the synthetic jet ejectors described herein, the actuators may be designed so that, even at maximum operating temperatures, the diaphragm does not contact the actuator housing. However, this approach is not preferred since it will typically mean that, at lower temperatures, maximum air flow will not be achieved.
It will be appreciated that, while the preferred methodologies disclosed herein utilize BEMF to determine the displacement of actuator diaphragms, other methods and devices may be utilized in the systems described herein to achieve a similar purpose. Some of these methods and devices are described below. In a typical example of these alternative approaches, some other means is used to determine diaphragm displacement, and voltage or other parameters are adjusted appropriately to maintain the maximum displacement at all times. While these alternative approaches will typically require a means for sensing the position or displacement of the diaphragm, the BEMF approach described above relies on features inherent in the system, and hence avoids the need for sensors or other such additional equipment.
In one possible alternative embodiment, optical sensors may be employed which may include laser diodes or photodiodes in combination with a photo sensor to sense position. In some cases, a protrusion may be placed on the diaphragm to facilitate measurements of the movements thereof. In a typical embodiment, the protrusion modulates the beam emitted by the diode such that the resulting signal generated at the sensor becomes lower and lower as more of the beam is blocked, thereby indicating the amount of the displacement. One advantage of this approach is that it automatically compensates for temperature, frequency and other such factors that may affect displacement.
Various capacitative methods could also be utilized to determine the displacement of the diaphragm. In one such approach, a plate may be placed above the diaphragm, which may or may not be in electrical communication with the coil of the actuator and/or a plate or magnet placed on the surface of the diaphragm. The difference in capacitance may then be sensed, which can be utilized to determine the extent of diaphragm displacement.
In other embodiments, pressure sensors may be utilized to determine the displacement of the diaphragm. Such sensors may operate by sensing the fluctuations in pressure within the actuator as the diaphragm moves towards, and away from, the housing.
In still other embodiments, ultrasonic methods may be used to determine the displacement of the diaphragm. These methods may include, for example, approaches similar to those utilized in ultrasonic imaging techniques, such as those based on the Doppler effect. Displacement may also be determined optically (e.g., through the use of lasers) using incident and reflected beams, and by measuring changes in the angles between the two beams.
The above description of the present invention is illustrative, and is not intended to be limiting. It will thus be appreciated that various additions, substitutions and modifications may be made to the above described embodiments without departing from the scope of the present invention. Accordingly, the scope of the present invention should be construed in reference to the appended claims.
Code is loaded into the board and a to-be-defined functional test is performed.
Set Need_to_Factory_Cal=True [in the non-volatile memory]
The actuator is assembled and mounted in a text fixture. The fixture can measure actuator displacement.
This application claims the benefit of priority from U.S. Provisional Application No. 61/002,237, filed Nov. 6, 2007, having the same title, and having the same inventors, and which is incorporated herein by reference in its entirety.
| Number | Date | Country | |
|---|---|---|---|
| 61002237 | Nov 2007 | US |
