Quarter bridge temperature compensation for force / torque sensors

Abstract

In force / torque sensors using strain gages, a hardware temperature compensation procedure using trim resistors and a single unstressed strain gage substantially eliminates thermal drift of multiple load sensing strain gages as a function of temperature. The strain gages are connected in multiple parallel stages in a quarter bridge configuration. The unstressed strain gage in the quarter bridge configuration is connected in parallel. In the compensation procedure that substantially eliminates thermal drift of the load sensing strain gages over a predetermined temperature range, trim resistors are added in parallel on one or more of the unstressed strain gage and the load sensing strain gages.

Description

TECHNICAL FIELD

[0001] The present invention relates generally to robotic force / torque sensors, and in particular to a system and method for temperature compensation of multiple load-sensing strain gauges configured in a Wheatstone quarter-bridge circuit with stress-free strain gauges. BACKGROUND

[0002] Robots are an integral part of product manufacturing, testing, assembly, and packaging; assisting and remote surgery; space exploration; operating in hazardous environments; and many other applications. Many robots and robotic applications require quantification of forces exerted or experienced, such as material removal (grinding, sanding, etc.), part assembly, remote excavation, or other manipulation of the environment, etc.

[0003] Industrial robots typically include a general-purpose actuator or "arm" that is programmed to move in space and operate on workpieces along multiple degrees of freedom. Various different tools (also known as end effectors) can be attached to the robot arm to perform different tasks.

[0004] In applications where the degree of force exerted by the robot on a workpiece must be monitored and controlled and / or the force experienced by the robot is fed back to control the motion of the robot ("force control" operation), a force / torque (F / T) sensor is placed between the robot arm and the tool.

[0005] One conventional type of F / T sensor uses strain gauges to measure the deformation of small beams connecting two mechanical components, one (directly or indirectly) coupled to the robot arm and the other (indirectly or directly) coupled to the robot tool. A compact example of such an F / T sensor is described in U.S. Patent No. 10,422,707 (hereinafter the '707 patent), assigned to the assignee of the present application, and incorporated by reference herein in its entirety.

[0006] Figure 1 Reproduced from the '707 patent, a central "hub", referred to in the art as a Tool Adapter Plate (TAP), is coupled to the tool. Another body, referred to in the art as a Mounting Adapter Plate (MAP), is coupled to the robot arm, arranged annularly around and spaced apart from the TAP. The MAP and TAP are connected by a plurality of relatively thin (and thus mechanically deformable) beams arranged radially around the TAP, which in one configuration resemble the spokes of a wheel. Relative force or torque between the objects coupled to the TAP and MAP, respectively, attempts to move the MAP relative to the TAP, causing at least some of the beams to deform or flex slightly.

[0007] Strain gauges are typically attached to multiple surfaces of at least some beams to detect this deformation. A beam undergoing mechanical deformation will elongate slightly along one side and compress along the opposite side. Multiple resistance strain gauges rigidly mounted to multiple sides of the beam experience the corresponding elongation or compression, and the resistance of the strain gauges is proportional to their length. Therefore, the resistance changes of the strain gauges can be detected and quantified, for example, through some configuration of a Wheatstone bridge circuit, and the signals from multiple strain gauges are combined to resolve the force and torque acting on the F / T sensor.

[0008] 707 patent describes and Figure 1 The image shows multiple strain gauges mounted only on one side of a deformable beam. This arrangement allows for a very compact sensor body and easy placement of the strain gauges during manufacturing, while still being able to resolve all forces and torques. The strain gauges are connected in a quarter-bridge circuit, as shown... Figure 2 As shown, this figure is reproduced from patent 707. Figure 3 .

[0009] The primary source of error in robot force / torque sensors is inaccuracy due to thermal drift. Sources of thermal drift include changes in ambient temperature, ambient temperature gradients, and self-heating. For silicon strain gauges, the change in output voltage of the strain gauge circuit caused by temperature variations can be several times greater than the change in output voltage caused by sensed stress. In fact, silicon strain gauges can be considered better as temperature sensors than as stress sensors. Connecting multiple strain gauges in a half-bridge topology can compensate for temperature effects, provided the strain gauges are well-matched and precisely positioned relative to each other. Furthermore, in addition to affecting the strain gauge output, temperature variations in robot force / torque sensors can also cause mechanical stress due to uneven expansion / compression of structural components, which the sensor can interpret as applied loads or forces.

[0010] One approach to compensating for thermal drift is to capture the effect of temperature changes on the strain gauge and eliminate them mathematically. Figure 3 It is a 707 patent. Figure 8 This illustrates a stress-free temperature-compensated strain gauge mounted on a stress-free mechanical component of an F / T sensor. Because this strain gauge does not undergo mechanical deformation to change its resistance, any change in resistance is solely due to temperature variations within the F / T sensor body. Mathematically, the signal from the stress-free strain gauge is removed from the signal from the load-sensing strain gauge to compensate for thermal effects.

[0011] International Patent Publication WO 2018 / 200668 (hereinafter referred to as 668 PCT Publication) describes subtracting the voltage output of a quarter-bridge circuit of a stress-free strain gauge from the voltage output of a quarter-bridge circuit of each load-sensing strain gauge, which is assigned to the assignee of this application and is incorporated herein by reference in its entirety. Figure 4is 668 PCT published Figure 2 which shows the signal from the strain gauge SG0 and SG1 is subtracted from the signal of the unstressed strain gauge SGU.

[0012] However, due to differences in the temperature coefficients of the individual strain gauges, even when the output of the unstressed strain gauge circuit is subtracted, the load sensing strain gauge circuit exhibits variability in output as a function of temperature. That is, at least some of the load sensing strain gauges exhibit different resistance changes with respect to temperature as compared to the unstressed strain gauge. This is referred to herein as the effective temperature coefficient with respect to the unstressed strain gauge. These effective temperature coefficients introduce thermal drift as a function of temperature change and result in errors in the measurement of force and torque.

[0013] The Background section of this document is provided for placing embodiments of the present application in technological and operational context, to help the skilled person understand its scope and utility. The methods described in the Background section can be employed without necessarily being previously contemplated or employed. Any statements as to the method of the application in this document are not to be taken as an admission that the application was previously practised or is not novel, unless expressly stated to be so. SUMMARY

[0014] In order to provide a basic understanding of the present disclosure, the following sets forth a simplified outline of the present disclosure. This outline is not an extensive overview of the present disclosure, and is not intended to identify key / important elements of the application embodiments or to delineate the scope of the application. The sole purpose of this outline is to give a simplified form of some of the concepts disclosed herein, as a prelude to the more detailed description that is set forth later.

[0015] According to one or more embodiments described and claimed herein, a hardware temperature compensation procedure using trim resistors substantially eliminates any effective temperature coefficient of the load sensing strain gauges with respect to the unstressed strain gauges. The strain gauges are connected in multiple parallel stages in a quarter bridge configuration. The unstressed strain gauges of the quarter bridge configuration are connected in parallel. In the compensation procedure that substantially eliminates any effective temperature coefficient of the load sensing strain gauges with respect to the unstressed strain gauges over a predetermined temperature range, trim resistors are added to one or more of the unstressed strain gauges and the load sensing strain gauges.

[0016] One embodiment relates to a temperature compensation circuit for a force / torque sensor. The circuit includes a first plurality of first stages connected in parallel. Each first stage is configured to output a respective voltage at a node between connecting a fixed resistor and a load-sensing strain gauge in series between a positive supply voltage and a negative supply voltage. The circuit also includes a second stage connected in parallel with the first stages. The second stage is configured to output a voltage at a node between connecting a fixed resistor and an unstressed strain gauge in series between the positive supply voltage and the negative supply voltage. The voltage of the second stage is subtracted from the voltage of each first stage. The circuit also includes an unstressed trim resistor connected in parallel with the unstressed strain gauge. The value of the unstressed trim resistor is selected to substantially cancel an effective temperature coefficient of all load-sensing strain gauges relative to the unstressed strain gauge of an opposite polarity to the polarity of the supply connected to the fixed resistor over a predetermined temperature range.

[0017] Another embodiment relates to a temperature compensation circuit for a force / torque sensor. The circuit includes a first plurality of first stages connected in parallel. Each first stage is configured to output a respective voltage at a node between connecting a fixed resistor and a load-sensing strain gauge in series between a positive supply voltage and a negative supply voltage. The circuit also includes a second stage connected in parallel with the first stages. The second stage is configured to output a voltage at a node between connecting a fixed resistor and an unstressed strain gauge in series between the positive supply voltage and the negative supply voltage. The voltage of the second stage is subtracted from the voltage of each first stage. The circuit also includes a second plurality of load trim resistors. Each load trim resistor is connected in parallel with a load-sensing strain gauge. The value of each load trim resistor is selected to substantially cancel an effective temperature coefficient of the associated load-sensing strain gauge relative to the unstressed strain gauge of a polarity of the supply connected to the fixed resistor over a predetermined temperature range.

[0018] Yet another embodiment relates to a method of compensating for thermal drift of a plurality of load sensing strain gauges in a force / torque sensor. A circuit includes a plurality of first stages connected in parallel. Each first stage is configured to output a respective voltage at a node between a positive supply voltage and a negative supply voltage in series connection of a fixed resistor and a load sensing strain gauge. The circuit further includes one second stage connected in parallel with the first stages. The second stage is configured to output a voltage at a node between the positive supply voltage and the negative supply voltage in series connection of a fixed resistor and a non-stressed strain gauge. The output voltages of the second stage and each first stage are measured, and the output voltage of the second stage is subtracted from the output voltage of each first stage over a predetermined temperature range. In response to at least one load sensing strain gauge exhibiting an effective temperature coefficient of polarity opposite to a polarity of a supply connected to the fixed resistor relative to the non-stressed strain gauge, a non-stressed trim resistor is added in parallel across the non-stressed strain gauge. The value of the non-stressed trim resistor is selected to substantially eliminate the effective temperature coefficient of polarity of all load sensing strain gauges opposite to the polarity of the supply connected to the fixed resistor relative to the non-stressed strain gauge over the predetermined temperature range. For each load sensing strain gauge, in response to the load sensing strain gauge exhibiting an effective temperature coefficient of polarity of a supply connected to the fixed resistor relative to the non-stressed strain gauge, a load sensing trim resistor is added in parallel across the load sensing strain gauge. The value of the load sensing trim resistor is selected to substantially eliminate the effective temperature coefficient of polarity of the load sensing strain gauge of the supply connected to the fixed resistor relative to the non-stressed strain gauge over the predetermined temperature range. BRIEF DESCRIPTION OF DRAWINGS

[0019] The present application will be described more fully hereinafter with reference to the accompanying drawings, in which embodiments of the application are shown. This application, however, can be susceptible to various modifications, variations and substitutions, and specific embodiments are shown by way of example in the drawings and / or detailed description. It should be noted that any reference to prior art throughout the specification and claims, is not to be construed as an admission that such prior art is widely known or forms part of the common general knowledge in the art.

[0020] Figure 1 is a cross-sectional view of a prior art force / torque sensor showing load sensing strain gauges.

[0021] Figure 2 is a diagram of strain gauges in a prior art quarter bridge circuit.

[0022] Figure 3 is a perspective cross-sectional view of a prior art force / torque sensor showing non-stressed strain gauges.

[0023] Figure 4 is a circuit diagram of a prior art showing a non-stressed strain gauge signal subtracted from two load sensing strain gauge signals.

[0024] Figure 5 is a circuit schematic diagram showing a quarter bridge circuit of six load-sensing strain gages and one unstressed strain gage.

[0025] Figure 6 is a plot of measured strain gage response as a function of temperature before compensation.

[0026] Figure 7 is a plot of measured strain gage response as a function of temperature with the response partially compensated to remove a negative effective temperature coefficient.

[0027] Figure 8 is a plot of measured strain gage response as a function of temperature with the response fully compensated to remove a positive effective temperature coefficient.

[0028] Figure 9 is a plot of measured strain gage response as a function of temperature after a number of iterations of the compensation procedure, Figure 8 is a plot of measured strain gage response as a function of temperature after a number of iterations of the compensation procedure,

[0029] Figure 10 is a flowchart of a method of compensating for thermal drift in a plurality of load-sensing strain gages in a force / torque sensor.

[0030] Figure 11 is a flowchart of a trim resistor determination procedure.

[0031] Figure 12A is a plot of measured strain gage response as a function of temperature before temperature compensation.

[0032] Figure 12B is a plot of measured strain gage response as a function of temperature after temperature compensation. DETAILED DESCRIPTION

[0033] For purposes of simplicity and illustration, the present application is primarily described by reference to exemplary embodiments thereof. In the following description numerous specific details are set forth in order to provide a thorough understanding of the present application. It will be apparent, however, to one ordinarily skilled in the art that the present application can be practiced without limitation to these specific details. In this description, well-known methods and structures have not been described in detail so as not to unnecessarily obscure the present application.

[0034] Embodiments of the present application are described below in the context of the compact force / torque (F / T) sensor described in the above-incorporated U.S. Patent No. 10,422,707 (“the ‘707 patent”). However, the present application is not limited to this application, and those skilled in the art will readily recognize that embodiments can be advantageously applied to a wide variety of F / T sensors. The background of the compact F / T sensor described herein is set forth in the ‘707 patent as follows: Figures 1 to 3 reproduced from the ‘707 patent as follows: Figure 1 , respectively.Figure 3 and Figure 8 , Figure 4 Reproduced from 668 PCT publication Figure 2 For clarity of explanation, element numbers from 707 patent are used here.

[0035] Figure 1 A plan view showing one embodiment of a F / T sensor 10. A TAP 12 is connected to a MAP 14 via three deformable beams 16a, 16b, 16c. In the embodiment shown, each beam 16 is connected directly to the TAP 12 and via a thin crimp 17 to the MAP 14, which facilitates deformation of the beams 16 under mechanical load. The TAP 12 is configured to be connected to a first object, such as a robot tool, via a through-hole 30. The MAP 14 is configured to be connected to a second object, such as a robot arm, via a plurality of mounting holes 32. The TAP 12 and MAP 14 are connected only by the beams 16 and crimps 17.

[0036] Load-sensing strain gauges 1-6 are affixed to the upper surface of each beam 16a-c. As used herein, the term "load-sensing strain gauge" refers to a strain gauge mounted to a component of the F / T sensor (e.g., deformable beam 16) that experiences mechanical stress when the F / T sensor measures an applied force and / or torque. In other embodiments, load-sensing strain gauges can be mounted to multiple surfaces of a deformable beam 16, e.g., on opposite surfaces thereof (e.g., opposite sides; top and bottom). The F / T sensor 10 also includes processing circuitry (not shown) operable to receive electrical signals from each load-sensing strain gauge 1-6 and process the signals to resolve the magnitude and direction of forces and torques applied between the MAP 14 and TAP 12. Such processing circuitry can include, for example, a microprocessor coupled to memory operable to store program code and sensor data.

[0037] Figure 2 Two load-sensing strain gauges wired in a quarter-bridge configuration are shown. As described in the 707 patent, this circuit configuration is sufficient to distinguish whether the portion of the surface of the beam 16 to which each load-sensing strain gauge is mounted is in tension or compression from changes in electrical resistance R LSn of the two load-sensing strain gauges. Comparison of information for all six load-sensing strain gauges is sufficient to determine forces Fx, Fy, Fz and torques Tx, Ty, Tz.

[0038] Figure 3A stress-free strain gauge 38 is shown. The stress-free strain gauge 38 is mounted to an extension 37 formed by creating a void 35 in the TAP 12. The extension 37 does not experience deformation due to any forces or torques applied to or experienced by the F / T sensor 10. Thus, any resistance change of the stress-free strain gauge 35 is due to temperature change. As used herein, the term “stress-free strain gauge” refers to a strain gauge that is mounted to a member of the F / T sensor that experiences substantially no mechanical stress when the F / T sensor measures applied forces and / or torques, otherwise the strain gauge is substantially similar to a load-sensing strain gauge.

[0039] Figure 4 Two load-sensing strain gauges SG0 and SG1 in a quarter bridge circuit configuration are shown. The load-sensing strain gauge signals are partially compensated for thermal drift by subtracting the signal of a stress-free strain gauge SGU (also connected in the quarter bridge circuit). While subtracting the stress-free strain gauge SGU signal improves the performance of the F / T sensor when temperature changes, the resistance change of each load-sensing strain gauge SG0, SG1 relative to temperature does not exactly match the resistance change of the stress-free strain gauge SGU. Thus, some of the load-sensing strain gauges SG0, SG1 exhibit an effective temperature coefficient relative to the stress-free strain gauge when the F / T sensor experiences a temperature change. This effective temperature coefficient detrimentally affects the force and torque measurements when the temperature of the F / T sensor changes.

[0040] Figure 5 A circuit for measuring the resistance change of load-sensing strain gauges and possibly trim resistors in an F / T sensor to compensate for thermal drift is shown. A plurality of first stages are connected in parallel, where each first stage is configured to output a respective voltage at a node between a fixed resistor and a load-sensing strain gauge connected in series between a positive and negative power supply voltage. The resistance of the ith load-sensing strain gauge is denoted as R LSi (i.e., R LS0 , R LS1 ,... R LS(n-1) ). In the circuit shown, n = 6, although this is not a limitation of the invention. While the variable R LSi only correctly refers to the resistance of the ith load-sensing strain gauge, for ease of discussion below, the strain gauge itself is sometimes denoted as R LSi (and the load-sensing strain gauges collectively as R LS ). Each of the n load-sensing strain gauges R LS0 , R LS1 , R LS(n-1) is connected in series to a respective fixed resistor R F0 , R F1 ,..., R F(n-1) . Note that in Figure 5In some circuit configurations, the fixed resistors are connected to a positive voltage source; in other embodiments, the fixed resistors can be connected to a negative voltage source. In practice, all fixed resistors R F may have the same value, such as 1 kΩ. In one embodiment, the value of the fixed resistors R F may be varied to keep the voltage measurements within a preferred range, as discussed in more detail herein.

[0041] The first stage is connected in parallel between the applied supply voltages V+ and V-. A voltage measurement is made at the midpoint between the load-sensing strain gauge R LSi and the fixed resistor R Fn forms a quarter-bridge circuit.

[0042] The second stage is connected in parallel with the first stage. The second stage is configured to output a voltage at the node between the fixed resistor R FU and the unstressed strain gauge R U connected in series between the positive and negative supply voltages. The fixed resistor R FU is connected to the positive supply voltage. In practice, the fixed resistor R FU may have the same value as the fixed resistor R F in the first stage. The unstressed strain gauge R U is substantially similar to the load-sensing strain gauge R LS , with the only difference being that it is mounted to a component of the F / T sensor that will not experience mechanical stress under the applied force or torque.

[0043] The output of each of the strain gauges R LSi , R U is a voltage measured at the midpoint of the respective stage. As shown in Figure 4 , the second stage output voltage is subtracted from the first stage output voltage. From this, using the known values of the fixed resistors R Fn , R FU , the resistance change of the load-sensing strain gauge R LSn can be calculated, and thus the applied force and torque causing the mechanical stress can be calculated. The compensation procedure described herein substantially eliminates any effective temperature coefficient of each load-sensing strain gauge R LSi relative to the unstressed strain gauge R U to substantially eliminate thermal drift from the force / torque measurements.

[0044] As noted above, the fixed resistors R Fn , R FU are connected to a positive voltage source. This determines the direction of the resistance change of the load-sensing strain gauge R LSn and the unstressed strain gauge R UThe direction of the relative temperature drift between the strain gauges, i.e., positive or negative, is also referred to as "polarity" in this paper. This relative thermal drift, more precisely, is related to the load-sensing strain gauge R... Lsi With stress-free strain gauge R U The difference in the rate of change of resistance with increasing temperature is referred to in this paper as the load-sensing strain gauge R. LSn Compared to stress-free strain gauge R U The effective temperature coefficient. The direction or polarity of the effective temperature coefficient depends on the fixed resistor R. Fn R FU Is it connected to a positive or negative power supply? For clarity, this article will assume... Figure 5 The circuit configuration, in which the fixed resistor R Fn R FU Connect to a positive voltage source.

[0045] Figure 5 The diagram shows the relationship between each load-sensing strain gauge R. LS Possible fine-tuning resistors R connected in parallel LST and stress-free strain gauge R U Possible fine-tuning resistors R connected in parallel UT Fine-tuning resistor R LST R UT The dashed lines indicate that, in the temperature compensation procedure, each trimming resistor may or may not be added to any given strain gauge. Load-sensing strain gauge R LSn Any trimming resistor on the circuit is referred to in this paper as a load-sensing trimming resistor R. LST And without stress strain gauge R U Any trimming resistor on the resistor is referred to in this document as a stress-free trimming resistor R. UT (load sensing fine-tuning resistor R) LST It does not sense any load and has no stress-adjustable resistor R UT Whether or not mechanical stress is experienced is irrelevant; the terminology used for trimmer resistors is for reference only. Any trimmer resistor R LST R UT The values ​​are determined during the iterative process of temperature change, and can be simulated in some embodiments.

[0046] According to an embodiment of the present invention, the load sensing strain gauge R LS Temperature compensation is applied to substantially eliminate the strain gauge R relative to the stress-free strain gauge within a predetermined temperature range. U The effective temperature coefficient. For Figure 5 The circuit configuration will include the fine-tuning resistor R. UT Add to stress-free strain gauge R U Increase all load-sensing strain gauges RLS Compared to stress-free strain gauge R U The effective temperature coefficient. Furthermore, for this circuit configuration, a trimming resistor is added to the load-sensing strain gauge R. LSi The load-sensing strain gauge R was reduced LSi Compared to stress-free strain gauge R U The effective temperature coefficient.

[0047] Initially, without applying any mechanical load, the output of each first and second stage is measured, and when the sensor is heated within a predetermined temperature range, the output voltage of the second stage is subtracted from the output voltage of each first stage. This is done across all load-sensing strain gauges R... LS The change in resistance value and the stress-free strain gauge R U In cases where the resistance change is unlikely to be precisely matched, compensation is unnecessary. In most practical situations, within a predetermined temperature range, at least one load-sensing strain gauge R... LSi It will exhibit a negative effective temperature coefficient relative to a stress-free strain gauge.

[0048] Figure 6 Showing each load-sensing strain gauge R LSi The response as temperature increases. As used in this paper, the load-sensing strain gauge R... LSn The response is the difference between the output voltages of the first and second stages, normalized by the excitation voltage (the difference between V+ and V-). This response is measured as temperature increases. For example, R LS0 The curve represents the values ​​in mV / V when the F / T sensor body is heated within a predetermined temperature range. The time series plot. Due to each fixed resistor R Fi The value of is known, therefore the strain gauge R LS The resistance value relative to R U The resistance value can be easily calculated based on the measured voltage and Ohm's law. In another embodiment, the resistance R of each load-sensing strain gauge can be directly measured at multiple points within a predetermined temperature range. LSi The resistance. In Figure 6 In the example, R LS1 It has the largest negative effective temperature coefficient, meaning that as the temperature increases, the corresponding voltage drop of the first stage output is the greatest relative to the second stage. Note that because R... LS0 and R LS1 The curves intersect, so Figures 6 to 9 R is drawn with a dashed line in the middle. LS0 The curve is to avoid confusion.

[0049] Then determine the stress-free fine-tuning resistor R. UT The value of R, which makes the strain gauge R relative to the stress-free strain gaugeU It exhibits the maximum negative effective temperature coefficient (in Figure 5 In general, a circuit is represented as R. LSmax-neg Load sensing strain gauge R LS1 The response to temperature flattens out. In one embodiment, a stress-free trimming resistor R is determined. UT The values ​​include performing a binary search on the resistance value, using a stress-free trimmer resistor R. UT The new value replaces the current value, and for different R... UT In each iteration of the value, the output of the first stage is measured or simulated within a predetermined temperature range until the maximum load-sensing strain gauge R is found to be within the predetermined temperature range. LSmax-neg ( Figure 6 R in the middle LS1 The effective temperature coefficient of R is essentially zero. UT value.

[0050] Figure 7 The determined stress-free trimming resistor R is shown. UT Connected to stress-free strain gauge R U Subsequent load sensing strain gauge R LS Response. The load-sensing strain gauge R previously exhibited the largest negative effective temperature coefficient. LS1 The response is now essentially constant over the predetermined temperature range. However, the stress-free trimmer resistor R... UT The addition of this increases the value of all other load-sensing strain gauges R LS Compared to stress-free strain gauge R U The effective temperature coefficient. In other words, compared to the second stage, the remaining first stage now outputs a voltage that increases more significantly with increasing temperature.

[0051] Next, relative to the stress-free strain gauge R U Each load-sensing strain gauge R exhibits a positive effective temperature coefficient LS A load sensing trimmer resistor R is connected in parallel above. LST Determine the load sensing trimmer R for each first stage. LSTi The value of makes the load sensing strain gauge R LSi The response is essentially flat within the predetermined temperature range. Here, a binary search can also be applied to replace the load-sensing trimmer resistor R. LSTi The new value is determined, and the load-sensing strain gauge R is measured or simulated at each iteration. LSi Resistance within a predetermined temperature range.

[0052] Figure 8 The load-sensing strain gauge R is shown after temperature compensation following one iteration of the compensation procedure. LSresponse. All load-sense strain gauges R LS exhibit a substantially constant response over a predetermined temperature range (note that in the present example, R LS0 and R LS5 exhibit overlapping curves).

[0053] Figure 9 show the response of the temperature-compensated load-sense strain gauges R LS after multiple iterations of the compensation procedure. The response is more flat, that is, all load-sense strain gauges R U exhibit a substantially zero effective temperature coefficient with respect to the unstressed strain gauge R LS . In addition, the response is more centered around zero mV / V.

[0054] The skilled person will note that the above temperature compensation procedure is related to the quarter-bridge circuit configuration of Figure 5 . In particular, in this circuit configuration, the fixed resistors R Fn , R FU are connected to a positive voltage source. If the polarity of the supply voltage is reversed, the fixed resistors R Fn , R FU are connected to a negative voltage source, then the opposite bridge output change occurs. That is, adding an unstressed trimmer resistor R U to the unstressed strain gauge R UT will decrease the effective temperature coefficient of the load-sense strain gauge R LSn with respect to the unstressed strain gauge R U . In addition, adding a load-sense trimmer resistor R LSi to the load-sense strain gauge R LSTi will increase the effective temperature coefficient of this load-sense strain gauge R LSi with respect to the unstressed strain gauge R U .

[0055] Therefore, in general, the value of the unstressed trimmer resistor R UT is chosen such that the effective temperature coefficient of all load-sense strain gauges R LSn with respect to the unstressed strain gauge R U of opposite polarity to the polarity of the supply voltage connected to the fixed resistors R Fn , R FU is substantially eliminated over a predetermined temperature range. In addition, in general, the value of each load trimmer resistor R LSTi is chosen such that the effective temperature coefficient of the associated load-sense strain gauge R LSi with respect to the unstressed strain gauge R U connected to the fixed resistors R Fn , R FUThe effective temperature coefficient of the polarity of the power supply voltage.

[0056] Figure 10 It is a compensation for multiple load sensing strain gauges R in the F / T sensor LS A flowchart of the steps in the thermal drift method 100. The method operates in a circuit comprising multiple first stages connected in parallel. Each first stage i is configured to... Fi and load sensing strain gauge R LSi The circuit outputs its respective voltage at the node between the positive and negative power supply voltages, connected in series. The circuit includes a second stage connected in parallel with the first stage. The second stage is configured to connect a fixed resistor R... FU and stress-free strain gauge R U The output voltage is connected in series at the node between the positive and negative power supply voltages.

[0057] Measure the output voltage of the second stage and each of the first stages, and subtract the output voltage of the second stage from the output voltage of each of the first stages within a predetermined temperature range (box 102).

[0058] The need for a fine-tuning resistor for each strain gauge and the value of the fine-tuning resistor are determined in the fine-tuning resistor determination procedure (box 200) described below. In one embodiment, procedure 200 is executed in a computer simulation of the circuit. In one embodiment, the circuit is modeled using Python and its operation over a predetermined temperature range is simulated; however, those skilled in the art can implement the modeling and simulation using any of a variety of programming languages ​​and / or circuit simulation programs. Alternatively, procedure 200 can be executed using discrete resistors or potentiometers, iteratively thermally cycling the circuit as needed to converge to the fine-tuning resistor value, as described more fully herein.

[0059] Once all the required trimmer resistors and their values ​​(if needed) have been determined according to the trimmer resistor determination procedure (box 200), the determined trimmer resistors R are... UT R LSTn Installed in the circuit (box 104). The circuit is then cyclically run within a predetermined temperature range, and the load-sensing strain gauge R is measured. LS The responses are used to verify that they remain essentially constant (Box 106).

[0060] Figure 11 The steps of one embodiment of the fine-tuning resistor determination procedure 200 are shown. Figure 11 The following description of the trimmer resistor determination procedure 200 describes the most general case, and is not limited to it. Figure 5 Circuit configuration (where the fixed resistor R) Fn R FUConnect to a positive power supply voltage. If in method 100 ( Figure 10 During the measurement process of frame 104, at least one load-sensing strain gauge R LS Compared to stress-free strain gauge R U An effective temperature coefficient (box 202) exhibiting the opposite polarity to the supply voltage connected to the fixed resistor is then present in the stress-free strain gauge R. U Add a stress-free fine-tuning resistor R in parallel. UT Stress-free trimming resistor R UT The value was selected such that virtually all load-sensing strain gauge R was eliminated within a predetermined temperature range. LS Compared to stress-free strain gauge R U The effective temperature coefficient (box 204) is opposite in polarity to the supply voltage connected to the fixed resistor.

[0061] In one embodiment, by first identifying the strain gauge R relative to the stress-free strain gauge U A load-sensing strain gauge R exhibiting a maximum effective temperature coefficient with a polarity opposite to that of the supply voltage connected to a fixed resistor. LSmax To perform stress-free fine-tuning resistor R UT Determination of the value of R. In the case of a stress-free strain gauge R... U Add a first stress-free fine-tuning resistor R in parallel. UT1 And the identified load-sensing strain gauge R is measured again within a predetermined temperature range. LSmax The response of the load sensing strain gauge R. LSmax If the effective temperature coefficient still exhibits a polarity opposite to that of the power supply voltage connected to the fixed resistor, then different stress-free trimmer resistors R can be used. UT2 Replace the current stress-free trimmer resistor. In one embodiment, j consecutive stress-free trimmer resistors R are selected by binary search. UTj The value of is determined iteratively by repeatedly measuring the response within a predetermined temperature range and replacing different stress-free trimming resistors R. UTj The steps continue until the identified load-sensing strain gauge R... LSmax It exhibits a practically zero effective temperature coefficient (or a small temperature coefficient of polarity of the supply voltage connected to the fixed resistor). The final stress-free trimmer resistor R is stored. UT The value of , and all load-sensing strain gauges R LS The response during the last measurement iteration.

[0062] Then, the fine-tuning resistor determination procedure 200 determines whether a load-sensing fine-tuning resistor R needs to be added to the first stage of the circuit. LST If so, determine their values. Stress-free trimmer resistor R UTSelected to make the identified load-sensing strain gauge R LSmax The effective temperature coefficient is zero. Thus, most (if not all) residual load sensing strain gauges R LS The effective temperature coefficient is now likely to be non-zero, and is the polarity of the power supply connected to the fixed resistor.

[0063] If during the last measurement iteration in box 204, there is still a load-sensing strain gauge R LSi Compared to stress-free strain gauge R U The effective temperature coefficient (box 206) that exhibits the polarity of the power supply voltage connected to the fixed resistor is then expressed in the load-sensing strain gauge R. LSi Add a load sensing trimmer resistor R in parallel. LSTi Load sensing trimmer R LSTi The value is selected such that the associated load-sensing strain gauge R is substantially eliminated within a predetermined temperature range. LSi Compared to stress-free strain gauge R U Effective temperature coefficient (box 208).

[0064] Similar to a stress-free trimmer resistor R UT In one embodiment, the determination of the load sensing fine-tuning resistor R is performed during the iterative measurement / replacement process. LSTi The value of R is determined. For the strain gauge R relative to the stress-free strain gauge... U Each i-th load-sensing strain gauge R exhibits an effective temperature coefficient that reflects the polarity of the power supply voltage connected to a fixed resistor. LSi In the associated load-sensing strain gauge R Lsi Add a first load sensing trimmer resistor R in parallel. LST1 Measure at least the load-sensing strain gauge R again within the predetermined temperature range. LSi The response of the load sensing strain gauge R. LSi If it still exhibits an effective temperature coefficient, then use different load sensing trimmer resistors R. LST2 Replace the load sensing strain gauge R Lsi The current load sensing trimmer resistor is used. In one embodiment, j consecutive load sensing trimmer resistors R are selected by binary search. LSTj The value of is determined iteratively by repeatedly measuring the response and replacing different load-sensing trimmer resistors R within a predetermined temperature range. UTj The steps continue until the i-th load sensing strain gauge R LSi Compared to stress-free strain gauge R U The effective temperature coefficient is essentially zero. This process is then repeated for the next first stage until all n load-sensing strain gauges exhibiting effective temperature coefficients that reflect the polarity of the power supply connected to the fixed resistor are compensated.

[0065] Alternatively, all of the load-sensing trim resistors R LSTn may be changed simultaneously, and then the strain gauge responses are measured in a thermal cycle. All of the final load-sensing trim resistor R LST values are stored. The trim resistors with these values are installed in the circuit (block 104) and the temperature stability is verified (block 106), as discussed above with respect to the method 100. Figure 10

[0066] In some cases, the addition of the trim resistors R LST , R UT may cause the output to exceed the desired voltage range. In one embodiment, in such cases, the values of one or more of the fixed resistors R F , R FU are changed to bring the output voltage within the desired range. In such cases, any trim resistors R LSTn , R UT are removed, the measurement cycle is performed, measuring the responses of all of the load-sensing strain gauges R LS over a predetermined temperature range, and then the trim resistor determination procedure 200 is restarted at block 202, using the updated fixed resistor values R F , R FU .

[0067] Returning Figure 5 , after performing the compensation procedure (e.g., the method 100), there are several possible configurations of trim resistors R LSTn , R UT depending on the responses measured during the compensation procedure. As noted above, in the unlikely case that the first measurements show that all of the load-sensing strain gauges R LS exhibit zero effective temperature coefficients relative to the unstressed strain gauge R U , no trim resistors R LST or R UT are needed.

[0068] Another possibility is that all of the effective temperature coefficients of one or more of the load-sensing strain gauges R LS are zero or positive. In this case, no unstressed trim resistors R UT are added to the unstressed strain gauge R U , but load-sensing trim resistors R LS are added to one or more of the load-sensing strain gauges R LST .

[0069] In the general case, the load-sensing strain gauges R LSmax ​Initially exhibiting the largest negative effective temperature coefficient, this suggests the need for a stress-free trimmer resistor R. UT Add to stress-free strain gauge R U So that the strain gauge R LSmax The response is essentially constant within a predetermined temperature range (i.e., relative to the stress-free strain gauge R). U This could cause most or all other load-sensing strain gauges to have their effective temperature coefficient reduced to zero. LS The presence of a positive effective temperature coefficient suggests the addition of a load-sensing trimmer resistor R. LST Therefore, up to (n-1) first stages may include load-sensing trimmer resistors R. LST One exception is the load-sensing strain gauge R. LSmax Initially, it was controlled by a stress-free trimmer resistor R. UT Zeroing out. However, even in this case, as a practical problem, because the stress-free trimming resistor R... UT The available values ​​are discrete, so the load-sensing strain gauge R can be selected. LSmax Having a slightly positive effective temperature coefficient instead of zero also suggests adding a load-sensing trimmer resistor R. LST .

[0070] Figure 6 The data plotted are measurements taken from representative F / T sensors. Using Figure 5 The model of the circuit simulates the changes in the circuit (i.e., the adjustment of the resistor R). LSTn R UT (Additions and adjustments to values). Therefore... Figures 7 to 9 The strain gauge response plotted in the image reflects the simulated data. In contrast, Figure 12A and 12B The strain gauge response of the F / T sensor, measured before and after thermal compensation method 100, is depicted, with the added and adjusted trimmer resistors being actual hardware, and the strain gauge response measured within a predetermined temperature range. These figures illustrate the significant improvement in thermal drift achieved through embodiments of the invention.

[0071] Compared with the prior art, the embodiments of the present invention have many advantages. By substantially eliminating multiple load-sensing strain gauges R within a predetermined temperature range... LSn Relative to a stress-free strain gauge R UThe effective temperature coefficient of the strain gage is improved, and other thermal drift mitigation measures become moot. Temperature compensation of strain gages according to embodiments of the invention results in a specific value of a discrete trim resistor that matches a specific strain gage installed in a F / T sensor quarter bridge measurement circuit. Compensation can be performed at F / T sensor manufacturing, and once the trim resistor is installed in the circuit, thermal drift is permanently eliminated in hardware over the life of the F / T sensor.

[0072] In general, all terms used herein are to be interpreted according to their ordinary meaning in the technical field of the application, unless explicitly defined or implied by the context of their use. All references to a list of elements are to be interpreted as open-ended, meaning that the list of elements does not preclude additional, unrecited elements being present. Reference to or description of an embodiment of the present application or comprising an embodiment of the present application means that the embodiment incorporates features, elements and / or components of the present application; it does not, however, mean that every embodiment of the present application must include the described or referenced feature, element or component. Likewise, reference to or description of an embodiment of the present application or comprising an embodiment of the present application means that the embodiment incorporates features, elements and / or components of the present application; it does not, however, mean that every embodiment of the present application must include the described or referenced feature, element or component. Any method disclosed herein does not necessarily have to be performed in the exact order disclosed, unless explicitly stated or implied by the context of the steps. Any feature of any embodiment disclosed herein can be applied to any other embodiment, where appropriate. Likewise, any advantage of any embodiment can be applied to any other embodiment, and vice versa. Other objects, features and advantages of the included embodiments will be apparent from the following description. As used herein, the term "configured to" means established, organized, adapted or arranged to operate in a particular way; the term is synonymous with "designed to". As used herein, the term "substantially" means nearly or essentially, but not necessarily completely; the term encompasses and accounts for mechanical or component value tolerances, measurement errors, random variations and similar sources of inaccuracy.

[0073] Of course, the application can be carried out in other ways than those specifically set forth herein without departing from essential characteristics of the application. The present embodiments are to be considered in all respects as illustrative and not restrictive, and all changes coming within the meaning and equivalency range of the appended claims are intended to be embraced therein.

Claims

1. A temperature compensation circuit for a force / torque sensor, comprising: a first plurality of first stages connected in parallel, each first stage configured to output a respective voltage at a node between connecting a fixed resistor and a load-sensing strain gauge in series between a positive supply voltage and a negative supply voltage; one second stage connected in parallel with the first stages, the second stage configured to output a voltage at a node between connecting a fixed resistor and a non-stressed strain gauge in series between the positive supply voltage and the negative supply voltage; wherein the voltage of the second stage is subtracted from the voltage of each first stage; and one non-stressed trim resistor connected in parallel with the non-stressed strain gauge; wherein the value of the non-stressed trim resistor is selected to substantially cancel the effective temperature coefficient of all load-sensing strain gauges relative to the non-stressed strain gauge of an opposite polarity to the polarity of the supply voltage connected to the fixed resistor over a predetermined temperature range.

2. The temperature compensation circuit of claim 1, further comprising: a second plurality of load trim resistors, each load trim resistor connected in parallel with a load-sensing strain gauge; wherein the value of each load trim resistor is selected to substantially cancel the effective temperature coefficient of the associated load-sensing strain gauge relative to the non-stressed strain gauge of the polarity of the supply voltage connected to the fixed resistor over the predetermined temperature range.

3. The temperature compensation circuit of claim 2, wherein the second plurality is one less than the first plurality, such that one load-sensing strain gauge has no load trim resistor connected thereto.

4. The temperature compensation circuit of claim 3, wherein the load-sensing strain gauge without a load trim resistor is the load-sensing strain gauge exhibiting the greatest effective temperature coefficient of the polarity of the supply voltage connected to the fixed resistor relative to the non-stressed strain gauge when no trim resistor is connected to the non-stressed strain gauge.

5. The temperature compensation circuit of claim 1, wherein the non-stressed strain gauge is mounted to a member of the force / torque sensor that experiences substantially no mechanical strain when the force / torque sensor measures an applied force and / or torque.

6. The temperature compensation circuit of claim 1, wherein the load-sensing strain gauges are mounted in pairs to the same side of different deformable beams that experience mechanical deformation caused by the force and / or torque measured by the force / torque sensor.

7. The temperature compensation circuit of claim 6, wherein the first plurality is six, and wherein three deformable beams have load-sensing strain gauges mounted thereto.

8. A temperature compensation circuit for a force / torque sensor, comprising: a first plurality of first stages connected in parallel, each first stage configured to output a respective voltage at a node between connecting a fixed resistor and a load-sensing strain gauge in series between a positive supply voltage and a negative supply voltage; one second stage connected in parallel with the first stages, the second stage configured to output a voltage at a node between connecting a fixed resistor and a non-stressed strain gauge in series between the positive supply voltage and the negative supply voltage; wherein the voltage of the second stage is subtracted from the voltage of each first stage; and a second plurality of load trim resistors, each load trim resistor connected in parallel with one load sense strain gauge; wherein the value of each load trim resistor is selected to substantially cancel the effective temperature coefficient of the associated load sense strain gauge relative to the polarity of the supply voltage connected to the fixed resistor of the unstressed strain gauge over a predetermined temperature range.

9. The temperature compensation circuit of claim 8, further comprising: one unstressed trim resistor connected in parallel with the unstressed strain gauge; wherein the value of the unstressed trim resistor is selected to substantially cancel the effective temperature coefficient of all load sense strain gauges relative to the polarity opposite to the polarity of the supply voltage connected to the fixed resistor of the unstressed strain gauge over a predetermined temperature range.

10. The temperature compensation circuit of claim 9, wherein the second plurality is one less than the first plurality, such that one load sense strain gauge is not connected with a load trim resistor.

11. The temperature compensation circuit of claim 10, wherein the load sense strain gauge without a load trim resistor is the load sense strain gauge exhibiting the largest negative effective temperature coefficient relative to the unstressed strain gauge prior to connection of the load trim resistor.

12. The temperature compensation circuit of claim 8, wherein the unstressed strain gauge is mounted to a member of the force / torque sensor that does not substantially experience mechanical strain when the force / torque sensor measures an applied force and / or torque.

13. The temperature compensation circuit of claim 8, wherein the load sense strain gauges are mounted to the same side of different deformable beams that experience mechanical deformation caused by the force and / or torque measured by the force / torque sensor.

14. The temperature compensation circuit of claim 13, wherein the first plurality is six, and wherein load sense strain gauges are mounted on three deformable beams.

15. A method of compensating for thermal drift of a plurality of load sense strain gauges in a force / torque sensor in a compensation circuit, the circuit comprising a plurality of first stages connected in parallel, each first stage configured to output a respective voltage at a node between a positive supply voltage and a negative supply voltage in series connection of a fixed resistor and a load sense strain gauge, and one second stage connected in parallel with the first stages, the second stage configured to output a voltage at the node between the positive supply voltage and the negative supply voltage in series connection of a fixed resistor and an unstressed strain gauge, the method comprising: measuring the output voltage of the second stage and each first stage over a predetermined temperature range, and subtracting the output voltage of the second stage from the output voltage of each first stage; in response to at least one load-sensing strain gauge exhibiting an effective temperature coefficient of polarity opposite to the polarity of the supply voltage connected to the fixed resistor relative to the unstressed strain gauge, adding an unstressed trim resistor in parallel across the unstressed strain gauge, wherein the value of the unstressed trim resistor is selected to substantially eliminate all load-sensing strain gauges exhibiting the effective temperature coefficient of polarity opposite to the polarity of the supply voltage connected to the fixed resistor relative to the unstressed strain gauge within the predetermined temperature range; and for each load-sensing strain gauge, in response to the load-sensing strain gauge exhibiting an effective temperature coefficient of polarity of the supply voltage connected to the fixed resistor relative to the unstressed strain gauge, adding a load-sensing trim resistor in parallel across the load-sensing strain gauge, wherein the value of the load-sensing trim resistor is selected to substantially eliminate the load-sensing strain gauge exhibiting the effective temperature coefficient of polarity of the supply voltage connected to the fixed resistor relative to the unstressed strain gauge within the predetermined temperature range.

16. The method of claim 15, wherein adding an unstressed trim resistor comprises: identifying a load-sensing strain gauge exhibiting a maximum effective temperature coefficient of polarity opposite to the polarity of the supply voltage connected to the fixed resistor relative to the unstressed strain gauge; adding a first unstressed trim resistor in parallel across the unstressed strain gauge; measuring a response of the identified load-sensing strain gauge within the predetermined temperature range; and iteratively performing the following steps: in response to the identified load-sensing strain gauge exhibiting an effective temperature coefficient of polarity opposite to the polarity of the supply voltage connected to the fixed resistor relative to the unstressed strain gauge, replacing the current unstressed trim resistor with a different unstressed trim resistor; and measuring a response of the identified load-sensing strain gauge within the predetermined temperature range; until the identified load-sensing strain gauge exhibits an effective temperature coefficient of substantially zero relative to the unstressed strain gauge within the predetermined temperature range.

17. The method of claim 16, wherein iteratively measuring a response of the identified load-sensing strain gauge within the predetermined temperature range comprises simulating a model of the circuit within the predetermined temperature range and recording a simulated response of the identified load-sensing strain gauge.

18. The method of claim 15, wherein adding a load-sensing trim resistor comprises, for each load-sensing strain gauge identified as exhibiting an effective temperature coefficient of polarity of the supply voltage connected to the fixed resistor relative to the unstressed strain gauge: adding a first load-sensing trim resistor in parallel across the load-sensing strain gauge; measuring a response of the load-sensing strain gauge within the predetermined temperature range; and iteratively performing the following steps: in response to the load-sensing strain gauge exhibiting an effective temperature coefficient of polarity of the supply voltage connected to the fixed resistor, replacing the current load-sensing trim resistor with a different load-sensing trim resistor; and measuring a response of the load-sensing strain gauge over the predetermined temperature range; until the load-sensing strain gauge exhibits a substantially zero effective temperature coefficient over the predetermined temperature range relative to the unstressed strain gauge.

19. The method of claim 18, wherein iteratively measuring a response of the load-sensing strain gauge over the predetermined temperature range comprises simulating a model of the circuit over the predetermined temperature range and recording a simulated response of the load-sensing strain gauge.

20. The method of claim 15, wherein the unstressed strain gauge is mounted to a member of the force / torque sensor that does not substantially experience mechanical strain when the force / torque sensor measures an applied force and / or torque.

21. The method of claim 15, wherein the force / torque sensor comprises six load-sensing strain gauges, two of which are mounted to each of three deformable beams that experience mechanical deformation caused by an applied force and / or torque.

22. The method of claim 21, wherein the two load-sensing strain gauges on each deformable beam are mounted on the same side of the deformable beam.

Images