Push on a strain gauge sensor and it answers in millivolts. A metal element bends by an amount you would struggle to see, gauges riding on it shift in resistance, the bridge behind them reports the imbalance, and a calibration converts that into pounds or ounces.
Almost everything described by that sentence is sold as a finished instrument: you bolt it into a machine, connect it to a display, and read a number. A thin beam force sensor is not that. It is a component — a part you design into a product, alongside the circuit board, the housing and the firmware. The distinction is not marketing; it changes what the specifications mean and who is responsible for what.
The datasheet is written for a designer. Read it that way and the unusual entries start to make sense. A single combined error figure instead of separate nonlinearity, hysteresis and repeatability numbers, because a designer wants one bound to work with rather than three to add up. A zero balance quoted in millivolts per volt rather than as a percentage, because you are going to null it. A bridge resistance given with a wide tolerance, because your circuit has to accommodate the spread rather than assume a value.
Three things become yours rather than ours. On a finished instrument we control the mounting, the cable and the zero. On a component you control all three, and each one can undo the sensor if it is left to chance.
The zero is the first of them, and the numbers surprise people. A zero balance of a few tenths of a millivolt per volt on a sensor rated at two millivolts per volt is a meaningful fraction of full output present before any load is applied. On a finished instrument that would be a fault; on a component it is entirely normal, because the assumption is that your electronics will tare it out at power-up or at assembly. Design in that tare and the offset is a non-event. Ignore it and a large part of your measuring range disappears.
The second is where the load lands. A single bending beam is sensitive to the position of the load along its length — the same force applied nearer or further from the clamped end produces a different bending moment and therefore a different signal. Manufacturer guidance on bending beams is explicit that single-beam designs need special measures to hold the load application point constant. In a product, that measure is your mechanism.
The third is the wiring. A component ships with a short, fine lead intended to reach your board, not your factory wall. Anything you add to it becomes part of the measuring circuit and is your responsibility to get right.
Sensors of this kind belong in medical instruments and infusion pumps, home appliances and touch controls, robotic grippers and end effectors, automotive safety systems, consumer electronics and haptic devices, process control, and small-scale test and quality work.
The bridge works as it does anywhere. Five things about integrating a thin beam are worth setting out, because between them they account for most of the difference between a design that works first time and one that does not.
Signal, and why the bridge resistance is so high. Excitation goes in at 10 VDC and the output at full scale is 2 mV/V — about twenty millivolts, expressed per volt because it is a ratio of the supply. What is unusual is the bridge itself at 1,200 ohms rather than the 350 ohms typical of industrial load cells. A bridge of that resistance draws well under ten milliamps, which is a very different proposition for a battery-powered instrument, a handheld device or a board with a modest regulator. The wide tolerance on that figure is normal for a component and simply means your excitation and input stage should not assume an exact value.
Nulling the zero is a design step, not a nuisance. The zero balance specification is given as a voltage ratio rather than a percentage for a reason: on a two millivolt per volt sensor, a few tenths of a millivolt per volt is a substantial share of the full output, present with nothing on the beam. Every practical design deals with this the same way — a tare taken in firmware at power-up or at end-of-line assembly, or an offset trim in the front end. It costs a line of code and it is the difference between using the sensor's range and losing part of it.
The load application point is part of the calibration. Because bending moment is force multiplied by distance, a thin beam reports where the load is as well as how heavy it is. Move the contact point along the beam and the reading changes with identical force. This is the reason mounting kits exist for this product rather than being an upsell: the kit's job is to hold the load application point where the calibration assumes it is. Whatever you design must do the same thing, and it must keep doing it after a thousand cycles of whatever your mechanism does.
Deflection is real travel your mechanism has to allow. A thin beam is a spring by design, and these deflect from six thousandths of an inch at the lowest capacity to sixty thousandths at the highest — movement you can see, and considerably more than an industrial load cell gives. Your assembly has to permit that travel freely, without the beam fouling a housing or a wire at full load. It is also worth designing a hard stop just beyond it: safe overload is 150% of full scale, and a compliant beam in a product that gets dropped, leaned on or over-tightened will find that limit if nothing prevents it.
The lead is nine inches of very fine wire. Four conductors of 30 AWG in a shielded jacket, sized to reach a nearby board rather than a distant instrument. Two consequences follow. Any extension becomes part of the measuring circuit, and fine wire has appreciably more resistance per foot than the cable industrial load cells use, so a long run costs sensitivity in a way that a four-conductor connection cannot compensate for. And the joint you make is now the weakest mechanical point — provide strain relief so nothing pulls on the sensor's own leads.
Materials suit the force. The beams are 301 stainless steel, with beryllium copper used at the two lowest capacities. That substitution is deliberate: at a quarter of a pound, the elastic behaviour of the flexure material dominates the measurement, and copper beryllium alloys are long-established for low-force flexures because of how cleanly and repeatably they return to shape.
Choosing a component sensor is a design exercise rather than a purchase, and the decisions that matter most are the ones made before anything is ordered. Talk to our application engineers early — it is far cheaper than discovering a constraint after a housing has been tooled.
Pick capacity by force, not by budget. All eight ranges are the same price, so nothing pushes you toward over-specifying, and you should not. The accuracy figure is a percentage of full scale, so a 40 lb sensor used to measure ounces spends almost all its resolution on force you never apply. Choose the smallest range that comfortably covers your maximum, and remember that the maximum includes what happens when someone presses too hard.
Design the load path before the housing. Decide early how force will be introduced onto the beam and how that point will be held constant. Our mounting kits exist to do exactly that, and for many designs they are the quickest route to a repeatable result. If you are building the fixture into your own product, treat the contact point as a controlled dimension rather than something the assembly happens to produce.
Budget for the fixture, not just the sensor. This trips up first-time integrators. The sensor is inexpensive; making it read reliably takes a mount, a contact feature, a travel allowance, an overtravel stop and a tare routine. That is normal for a component, and it is worth pricing at the design stage rather than discovering it during prototyping.
Think about temperature where the product will actually live. The compensated range extends well below room temperature, which suits appliances, vehicles and cold environments that our industrial products are not compensated for. The trade is that the temperature coefficients are wider than a laboratory instrument's, so across a large swing the drift is worth calculating rather than assuming. Where the reading matters, tare at operating temperature rather than at whatever temperature the device happened to be switched on.
Plan the connection. Work out where the sensor's short lead terminates, how it is strain relieved, and whether the shield is grounded at one end. If the run to your electronics is long, tell us — there are better answers than simply splicing on more wire.
Then decide what reads it. Many designs digitise the bridge directly on their own board. Where you would rather not, an amplifier signal conditioner module gives a usable voltage or current output, and a digital display suits bench and test use where a person is reading the number. If units get swapped in and out during development, Cal-Teds plug and play keeps each sensor's calibration data on the sensor instead of in a spreadsheet somebody has to maintain.
Tell us the force, the space, how the load is applied and what the sensor is going into, and we will help you settle on a range and a mounting approach. Every range in the table is a stock item, and schools, universities and research programmes qualify for reduced pricing.
TBS Series Load Cell Applications.
The Transducer Techniques TBS Series thin beam force sensors, known for their sensitivity and compact design, are widely used in industries and applications that require precise small load measurements, in eight capacity ranges from 0.25 lb to 40 lb.
- Medical Instrumentation: TBS Series force sensors are employed in medical devices and instruments for applications such as force sensing in surgical instruments, infusion pumps, and medical testing equipment.
- Home Appliances: These force sensors find applications in home appliances, including washing machines, dishwashers, and appliances with touch-sensitive controls, ensuring precise load measurements and user interactions.
- Process Control: TBS Series force sensors are integrated into process control systems to monitor and control small forces and loads in manufacturing and industrial processes. They contribute to process efficiency and quality control.
- Robotics: In robotics, these force sensors are used to measure forces and loads in robotic grippers, end-effectors, and arms. This information helps robots interact with objects and perform tasks more accurately.
- Automotive: Automotive manufacturers and suppliers use TBS Series force sensors in various applications, including automotive safety systems, seatbelt tensioning, and airbag deployment, to ensure accurate load measurements and safety.
- Consumer Electronics: TBS Series sensors are used in consumer electronics devices, such as touchscreens and input devices, to provide precise force sensing capabilities for user interactions.
- Industrial Automation: In industrial automation and manufacturing, these sensors are integrated into machinery and equipment for force monitoring and control in applications like assembly, material handling, and testing.
- Quality Control: Industries with stringent quality control requirements, such as electronics manufacturing, utilize TBS Series force sensors for quality testing of small components and products.
- Aerospace: The aerospace industry uses these force sensors for various applications, including load measurement in small components and test setups for aircraft and spacecraft.
- Research and Development: Engineers and researchers use TBS Series force sensors during product development and testing to measure and analyze small forces and loads in various experimental setups.
- Haptic Feedback: TBS Series sensors can be integrated into devices that provide haptic feedback, enhancing the user experience in virtual reality applications and touch-sensitive devices.
The Transducer Techniques TBS Series thin beam force sensors offer exceptional stability and reliability, making them ideal for small load measurement applications across a wide range of industries. Their sensitivity and accuracy contribute to improved product quality, safety, and efficiency in diverse applications.
Frequently Asked Questions
What is a thin beam force sensor and how is it different from your other load cells?
It is a small cantilever beam with strain gauges on it: clamp one end, apply force near the other, and the bending produces a signal. The difference is not really the shape but the role. Our other products are finished instruments you install and read. This is a component you design into something — a medical device, an appliance, a robotic gripper — alongside your own board, housing and firmware. That means the mounting, the zeroing and the wiring are part of your design rather than part of the product, which is why the specifications are written the way they are.
What capacities does the TBS Series cover and what does it cost?
Eight ranges: 0.25, 0.50, 1, 2, 5, 10, 20 and 40 lb, all at $135. A flat price across the range means nothing encourages over-specifying, and you should choose the smallest range that comfortably covers your maximum force — the accuracy figure is a percentage of full scale, so a 40 lb sensor asked to resolve ounces gives away most of its usefulness. Mounting kits are available separately at $95, and for many designs they are the quickest route to a repeatable installation.
What does "combined error 0.25% of full scale" mean?
It is a single figure covering the effects that our industrial products list separately as nonlinearity, hysteresis and nonrepeatability. Component datasheets are usually written this way because a designer wants one bound to work within rather than three numbers to combine. The practical reading is that within the sensor's range, any given force should land within a quarter of one percent of full scale of where the calibration says it should — and because it is referenced to full scale, that band is a fixed number of pounds regardless of where in the range you are working.
Why is the zero balance given in mV/V rather than as a percentage?
Because you are expected to remove it, so the useful thing to know is how much signal it represents rather than how it compares with a rated output. The figure is a few tenths of a millivolt per volt on a sensor whose full-scale output is two millivolts per volt, so a noticeable share of your span can be sitting there before any force is applied. Every practical design handles this the same way: a tare in firmware at power-up or at final assembly, or an offset trim in the input stage. Budget for it at the design stage and it is a non-issue.
Why is the bridge resistance 1,200 ohms with such a wide tolerance?
The high resistance keeps current draw low — well under ten milliamps at rated excitation, against roughly thirty for a 350 ohm industrial bridge. In a battery-powered instrument, a handheld device or anything with a modest regulator, that difference matters. The wide tolerance is normal for a component and simply means your excitation source and input stage should be designed to accommodate a spread rather than assume an exact value. It does not affect the output, which is a ratio of the excitation regardless.
How much does a TBS deflect under load?
Considerably more than an industrial load cell, and by design: 0.006" at 0.25 lb, 0.009" at 0.50 lb, 0.011" at 1 lb, 0.015" at 2 lb, 0.025" at 5 lb, 0.031" at 10 lb, 0.048" at 20 lb and 0.060" at 40 lb. A thin beam is a spring, and that movement is what produces the signal. Your mechanism has to allow the beam to move freely through its full travel without fouling a housing, a wire or a fastener — and a hard stop just beyond full deflection is worth designing in, since safe overload is 150% of full scale.
What is the TBS made from?
301 stainless steel on most of the range, with beryllium copper used for the 0.25 and 0.50 lb sensors. The change of material at the bottom of the range is deliberate rather than incidental: at very small forces the elastic behaviour of the flexure itself dominates the measurement, and copper beryllium alloys are long established for low-force flexures because of how cleanly and repeatably they return to shape after being deflected.
What lead wire is fitted, and can I extend it?
Nine inches of shielded four-conductor 30 AWG PVC lead — sized to reach a nearby circuit board rather than a distant instrument. You can extend it, but treat the extension as part of the measuring circuit rather than as cabling. Fine wire has appreciably more resistance per foot than the cable industrial load cells use, so a long run costs sensitivity, and a four-conductor connection has no sense leads to compensate for it. If your electronics are more than a short distance away, tell us the run length and we will suggest a better arrangement than splicing.
What temperature range is the TBS compensated over?
20° to 120°F, which reaches considerably lower than the 60° to 160°F band our industrial load cells are compensated over — appropriate for appliances, vehicles, cold rooms and anything that lives outside a heated building. The trade is that the coefficients are wider, at 0.02% of full scale per °F on both zero and output, so across a large temperature swing the drift is worth calculating rather than assuming away. Where the reading matters, tare at operating temperature rather than at switch-on.
What are the TBS-MK-1 and TBS-MK-2 mounting kits for?
They clamp the beam properly and hold the load application point where the calibration expects it. That second job is the important one. A thin beam responds to bending moment, which is force multiplied by distance, so the same load applied a little further along the beam produces a different reading. A mounting kit fixes that geometry so the sensor gives you force rather than force-and-position. Both kits are $95 — tell us your capacity and how the load arrives and we will confirm which one suits.
Questions From The Field
My readings change depending on exactly where the force touches the beam.
That is the beam behaving correctly, and it is the single most common surprise on this product. A bending beam measures moment, not force — the same load applied nearer or further from the clamped end bends it by a different amount. Manufacturer guidance on bending beam designs states plainly that single-beam sensors need special measures to hold the load application point constant. The fix is mechanical: a defined contact feature, a pin or a boss that lands in the same place every time, or one of our mounting kits. It cannot be corrected in software, because the sensor genuinely cannot tell the two situations apart.
There is a large output with nothing on the sensor. Is it faulty?
Almost certainly not. The zero balance specification allows a few tenths of a millivolt per volt with no load, which on a two millivolt per volt sensor is a meaningful part of full output. On a finished instrument that would be a problem; on a component it is expected, because the design assumption is that your electronics remove it. Take a tare reading at power-up with the mechanism at rest, or trim the offset in the front end. If the offset is far larger than the specification allows, or it moves about, then look at the mounting — a beam clamped unevenly or pre-loaded by its own fixture will show up exactly this way.
How should the beam be clamped?
Firmly, flatly and against a rigid surface, with the clamped end fully supported and the rest of the beam free to bend. Three things go wrong in practice. A mounting face that is not flat pre-stresses the beam before any load arrives, which shows as a large or drifting zero. A clamp that is too short leaves part of the intended fixed length free to move, changing the effective geometry. And over-tightening a small fastener into a thin beam can distort it. If you are designing your own fixture rather than using a mounting kit, get the clamped length and the flatness right first — everything else follows from those.
Can I mount the sensor with adhesive or double-sided tape?
We would not recommend it for anything you intend to rely on. Adhesive layers creep under sustained load, which shows up as a reading that drifts slowly while the force is genuinely constant, and they soften with temperature, so the effect changes with the environment. Both problems are worse on a compliant beam than on a stiff industrial cell. Use a mechanical clamp against a flat, rigid surface. If a bonded mount is unavoidable in your product, tell us what you are considering and we will be honest about what it will cost you in stability.
The sensor reads correctly at first but drifts over a few minutes.
Work through three candidates in order. Temperature is the usual culprit — the coefficients on this series are 0.02% of full scale per °F, so a device warming up under its own power dissipation will move perceptibly; take the tare once the assembly has reached working temperature. Next, look for creep in the mounting: an adhesive layer, a plastic boss, or a fastener bedding into a soft housing will all relax under load. Third, check that nothing is pressing on the beam that should not be — a wire routed across it or a housing that touches at full deflection will produce exactly this pattern.
Can I use one of these in a battery-powered device?
Yes, and the 1,200 ohm bridge is part of why. At rated excitation it draws well under ten milliamps, roughly a third of what a 350 ohm industrial bridge takes, and many designs reduce that further by exciting at a lower voltage or pulsing the excitation and sampling between pulses. Output is a ratio of excitation, so running at a lower voltage costs you signal proportionally rather than breaking anything — it is a straightforward trade between power and resolution. Tell us your power budget and we can talk through where the sensible balance sits.
Should I calibrate each unit in my product, or trust the specification?
That depends on what your product claims. Working from the nominal sensitivity is entirely reasonable where you need a repeatable indication rather than a traceable measurement — a grip force, a presence check, a threshold. Where your device reports a number a user acts on, an end-of-line calibration against a known force is worth the production time, and it also captures your own mounting and electronics rather than the sensor alone. In either case, a tare is not optional; that is a separate step from calibration and every design needs it.
Can I check whether a sensor has been damaged?
Two quick checks catch most of it. Measure the bridge resistance across the excitation pair and compare it against the specification — a reading far outside the stated tolerance, or an open circuit, points at a broken gauge or a damaged lead. Then measure insulation resistance between the bridge circuit and the beam itself, which is specified at 1,000 megohms at 50 VDC when healthy; a much lower figure indicates moisture or contamination reaching the gauges. Beyond that, look for visible set in the beam — an overloaded thin beam frequently takes a permanent bend, and a zero offset that has grown and stayed grown is the giveaway.