Sensors - June 2004 - A Novel Force Transducer and Load Cell Design Using “Leaky” Electromagnetic Waves

A Novel Force Transducer
and Load Cell Design
Using “Leaky” Electromagnetic Waves

A new technology lets you create high-performance load cells from virtually any material.

Force transducers, which include load cells, accelerometers, and pressure sensors, have been designed, manufactured, and commercialized using tried-and-true methods for more than 50 years. Even with the advent of newer materials, electronics, miniaturization, and better design analysis, transducers are still designed and created in much the same way as they were many years ago. The design methodology, focusing on tradeoffs among cost, manufacturing, reliability, and performance, has prevented these sensors from expanding beyond the most obvious markets. Because of the relatively large engineering effort required, manufacturers are reluctant to spend money to develop novel transducers for new applications. However, reconsidering the design methodology and using new technologies opens up many new application opportunities, both within the traditional force transducer markets and in new, (so far) unconsidered ones.

This article develops new load cell designs based on electromagnetic “leaky” waves. This technology lets us create high-performance load cells out of virtually any material, for a fraction of the cost of current designs. These can be mass produced without requiring tweaking, and can be fabricated, assembled, and tested without the need for manual labor. We start by explaining sensor operation and construction, along with manufacturing flow and system benefits, and we end by showing data for load cells constructed out of either aluminum or plastic (Delrin) and indicating the current state of commercial development.

Old-School Transducers
Force transducers fall into one of two design classes: stress-measurement transducers and deflection-measurement transducers. The transducer can measure stress in the material under load as a function of the applied load, or it can measure deflection of the mechanical design as a function of applied load. Both have their engineering limitations.

Stress-Measurement Transducers. The fundamental restriction on stress-measurement devices such as strain gauges and piezoresistive and piezoelectric transducers is the need for an intimate coupling between the sensing electronics and the mechanical structure. For example, the strain gauges in strain gauge load cells must be directly attached to the load cell body to detect stress. The tight coupling between mechanical and electrical elements means that you need engineering trade studies to produce a practical load cell design. Some of the tradeoffs considered include the following: the deflection of the load cell cannot be too great because it will overstress the strain gauge; the strain gauge material must match the temperature coefficient of the load cell body; and the adhesive attaching the strain gauge to the load cell must transfer stress optimally.

Deflection-Measurement Transducers. The second class of force transducers relies on deflection measurements to determine force. Here, the fundamental restriction is the complexity of implementing the technologies used to sense the deflection. The complexity can be in either the sensing technology itself or the manufacturing of the device. For example, using light to measure the deflection in a small-sized load cell requires either a large separation distance (time-of-flight measurement) or an inordinately small distance (Fabry-Perot or interferometry), not to mention that the load cell body must reflect an adequate amount of light. Acoustic technology is subject to severe reflections and noise because the wavelengths are too long; the entire load cell serves as an acoustic reflector, thereby muddying the return signal. Magnetic sensors do not seem to have the resolution required for the small (<0.005 in.) standard deflections and also require at least part of the load cell to be of a suitable magnetic material. Capacitive sensors are somewhat suitable, but require the load cell body to serve as one of the capacitance plates; this requires the load cell to be part of the sensor electronics—a potential assembly nightmare. Even though researchers have come up with solutions to some of these issues, the solutions require further compromises and the initial, simple load cell becomes complicated and, for high accuracy, expensive.

So, What’s the Solution?
We need a sensing technology that’s cheap to make and needs no adjustment, has robust resolution for the deflection range considered, has minimal temperature effects, is linear by nature to remove the need for conditioning electronics, and meets the performance standards of today’s more expensive high-accuracy load cells. Good luck finding the appropriate technology, right? Wrong! There’s a technology that satisfies all the requirements, but until now it hasn’t been applied to force transducer design—it’s the use of high-frequency electromagnetic (EM) waves from “leaky” structures. By using EM waves, we essentially decouple the mechanical and electrical designs. Because EM waves can propagate without the need for wires, there is no need for physical contact between the mechanical sensor body and the sensing electronics. This enables independent best-in-class design of both the electronics and the mechanical structure.

Sensor Design
In order to illustrate the use of EM waves from leaky structures in force transducer design, let’s continue to use our load cell example. A generic design for a dual-beam load cell transducer using our deflection measurement concept is shown in cross section in Figure 1.

The transducer is constructed of two independently designed, fabricated, and tested subassemblies. The first subassembly contains the sensor electronics, while the second is the mechanical load cell body. Both subassemblies are mated to each other to form the load cell transducer.

It is well known that as you increase the frequency of an electrical signal, the EM waves characteristic of the signal leak out of the nontransverse electromagnetic guiding structure into the surrounding environment [1]. The guiding structure can be a transmission line, open-ended waveguide, dielectric waveguide, coplanar waveguide, or microstrip. Normally you’d consider a leaky signal to be neither desirable nor useful, but in force transducer design it’s like manna from heaven. This is because the leaky signal and the environment interact, with changes in the local electromagnetic environment affecting the leaky signal. For example, if a dielectric or metallic material is brought within range of the leaky EM wave, the characteristics of the wave traveling within the guiding structure are perturbed, causing the electrical signal to change. Specifically, the amplitude of the signal, its phase, or its frequency are affected by disturbing the leaky EM wave. This perturbation effect is highly localized; it’s limited to the volume of space around the signal guiding structure (see Figure 2).

Figure 2. The unperturbed signal is shown traveling along the guiding structure, in this case a simple wire (A). By bringing a perturbing object (either metallic, dielectric, or semiconductor) close to the wire (B), the “leaky” EM fields are compressed and the characteristics of the guiding structure and hence the signal traveling along the wire are changed. In contrast, when a perturbing object is outside the leaky EM fields (C), the perturbing object has no effect on the signal.

The benefit of using a controlled mechanical structure such as a load cell body to perturb a leaky wave is threefold. First, the amount of change in the signal is directly correlated to the position (i.e., deflection) of the perturbing object (i.e., the end of the load cell’s cantilever beam). Second, this perturbation is highly localized so that only the part of the load cell closest to the leaky structure influences the signal. For highly accurate deflection measurements, the geometry of both the leaky wave guiding structure and the perturbation structure can be optimized. Third, the perturbating structure is only a perturbation on the existing signal, not a radical influence. If the perturbation is removed, the electronics structure still functions as expected and can be tested independent of the load cell body. This is a significant benefit because when the load cell assembly is done, you need only a quick test of the transducer—the electronics have been validated before the load cell assembly process.

Building the Transducer
Using perturbation of a high-frequency leaky EM signal to sense displacement lets you construct the transducer body using virtually any elastic material, including metals, plastics, injection-molded materials, semiconductors, and even wood, brick, and bone. In this article, we’ll show results for load cells constructed of aluminum and of Delrin, a thermoplastic material. Using a range of materials in transducer construction lets you use a uniform design topology with the benefits shown in Figure 3.

Figure 3. Current technology vs. BHTechnology is used in the design of a family of load cell products.

Conventionally, designing a family of load cells involves a progressive decrease in the cross section of the deflecting members to cover the intended sensing range. This is illustrated in Equation (1), the canonical equation for the deflection of a cantilever beam with a point end load.

Y_MAX	=	maximum deflection of the beam
F	=	applied load
L	=	length of the beam
E	=	modulus of elasticity of the beam material
w	=	width of the beam
t	=	thickness of the beam

Conventional load cell design assumes that to maintain a constant amount of deflection while changing the range of the load (i.e., maximum load results in a specific maximum deflection regardless of the absolute load value), either the length or the thickness of the beam must change. Changing the beam’s thickness leads to reliability, creep, fatigue, and overload issues for the light-duty load cell design, while changing the length of the beam leads to changing the form factors and overall sizes of the load cell family.

In contrast, using the technology illustrated here, the configuration and dimensions of the load cell don’t have to change to accommodate a variety of load ranges. We don’t vary any of the physical dimensions to maintain a constant maximum deflection value for arbitrary maximum load—rather, we change the value of E by selecting different materials to accomplish our desired deflection.

For example, you might need to construct a 100 lb. load cell from stainless steel while using Teflon for a 1 lb. load cell. Both transducers will have identical dual-beam dimensions. Changing the material to provide the correct load range rather than the design of the beam reduces the design issues mentioned earlier. This is because thicker beams inherently provide higher reliability, robustness, and resistance to creep and fatigue, while a constant beam length ensures a constant load cell footprint.

Additionally, by preserving the load cell’s form factor and letting the material do the design work, the load cell family becomes more versatile and modular in application. You can provide an end user with a single electronics subassembly and many different load cell bodies of different characterized materials. The user can then mix and match the load cell body with the electronics in a simple plug-and-play transducer implementation to suit his or her needs. Since the form factor of the load cell doesn’t change, swapping load cell bodies within the application doesn’t become an issue, yet the end user can measure 100 lb. loads one day and 1 lb. loads the next without sacrificing performance and functionality.

Alternatively, to reduce the cost of the load cell, and if high accuracy is not necessarily a concern (e.g., electronic postage scales), the proposed technology lets you form the load cell out of injection-molded plastic rather than machined aluminum. You will appreciate the variations and tradeoffs among cost, accuracy, and modularity that this new technology enables.

Putting It to the Test
To illustrate the design, manufacturing, and application points presented previously, we constructed a number of load cells from either aluminum or Delrin. All the data presented here are raw—there’s no normalizing, level shifting, sampling, linearizing, or signal conditioning; we simply power up the load cells and present the data as recorded by a standard digital multimeter.

There are two manufacturing methodologies to assemble the load cells. In a two-subassembly approach, the electronics module is mated to the load cell body. In the other approach, the load cell is constructed as a single body and the electronics board is attached to the load cell. We opted for the more challenging two-subassembly approach to illustrate the ease and repeatability of even simple designs. Our mated assemblies are flush mounted and held together via machine screws. No keyways or self-aligning structures were used to facilitate mating; we used stainless steel shims to prevent rotation of the two subassemblies relative to each other. Had we used such self-aligning structures we could have achieved higher degrees of accuracy and repeatability. Photographs of the load cells, their components, and the mating process are shown in Figures 4, 5, and 6.

Figure 4. Here are the individual components of the load cell (the circuit board is not populated); note the two distinct subassemblies. The electronics require only a simple two-sided PCB with discrete components on the top side of the board and the backside serving as a ground plane.

Figure 5. This shows the fully assembled aluminum (A) and Delrin (B) load cells—both use the same mounting block and electronics, validating the modular construction concept.

Figure 6. Here’s a close-up of the mating process of the two subassemblies. The shims are used to prevent rotation of the mounting block relative to the load cell body during mating, and are removed after final tightening of the screws.

The electronics subassembly is a completely independent component of the overall load cell and can be fully populated and tested at either the PCB vendor or the load cell manufacturer, pending cost effectiveness. The electronics are simple enough that the populated PCB can be tested using Automated Testing Equipment (ATE).

Measured Results. Figures 7 and 8 show data from the aluminum and Delrin load cells, respectively. We arbitrarily picked 30 lb. as the full load for the aluminum load cell. Because Delrin has a modulus of elasticity ~30 × smaller than that of aluminum, the maximum load applied to the Delrin load cell was 1 lb. The aluminum load cell was subjected to temperature variation, and exhibits minimal temperature effects and virtually no hysteresis. As noted above, the data presented here are raw—no data massaging, amplification, or signal conditioning has been done. Figures 7 and 8 show the inherent linearity of the measurement technology.

Figure 7. Aluminum load cell from Figure 5A is measured over the temperature range from 5°C–36°C.

Figure 8. Delrin load cell of Figure 5B is measured at room temperature. The red curve is a simple linear curve fit.

Figure 9 compares price and performance of conventional dual-beam strain gauge load cells to BHTechnology’s load cell design.

Figure 9. Comparison between conventional dual-beam load cells and BHTechnology’s dual-beam load cell design. Note that our design is not yet optimized.
Parameter	Conventional Dual-Beam Aluminum Load Cell (Typical)	BHTechnology Dual-Beam Aluminum Load Cell (Raw Data)*	Units
Excitation	10	5	V
Sensitivity	2	125	mV/V F.S.
Total Error	0.02 compensated 10°C–40°C	0.05 not compensated 5°C–25°C	% F.S.
Bridge Resistance	350	N/A
Current Consumption	29	25	mA
Deflection at Capacity	0.015	0.003	in.
Load Rating	15	15	kg
Safe Overload	150	500	% F.S.
Production Cost**	$100	<$25	N/A
Nonoptimized performance *Estimates at quantity of 100

Note that our current design is nonoptimized and noncompensated—with optimization and compensation, performance will be greatly enhanced with minimal impact to production cost, as all optimization and compensation are absorbed in simple electronics additions to the circuit design.

Summary
BHTechnology has developed a novel approach to force transducer design by using leaky EM waves. This allows us to design transducers with conventional high performance at a fraction of the conventional cost. Our design methodology enables the realization of new applications as well as the maximization of profits in conventional applications. Manufacturing methodologies also enable rapid transducer development, modular design, and automated assembly and testing of the finished product without the need for manual labor. We currently have opportunities for commercialization and transfer of our technology to the force transducer marketplace. Please note that this article contains information that is either patented or is patent pending.

Reference
1. Ramo, S., et al., Fields and Waves in Communications Electronics, J. Wiley & Sons, New York, 1984, Ch. 4-10.