A strain gauge instrument works by allowing a machined element to twist under the load being measured and reading that twist electrically. Bonded foil gauges track the deformation as a resistance change, a bridge converts it to a voltage, and calibration converts the voltage to foot-pounds.
What makes this one different is that it never has to stop. Ball bearing support and coined silver slip rings let the drive turn without limit, so the sensor keeps reading through revolution after revolution instead of being restricted to whatever arc a cable allows.
That sounds like a convenience. It is actually a change in what can be measured, and it is worth being precise about why.
A tightening torque is not a value. It is a curve. When a fastener is run down and tightened, torque climbs through several distinct phases, each governed by different physics, and the number quoted in a specification is a single point near the end of that climb. Published tightening practice divides the cycle into four zones: a rundown zone before the fastener contacts its bearing surface, where any prevailing torque from thread-locking features, misalignment or debris appears; an alignment and snugging zone where the joint is drawn into a stable clamped condition; an elastic clamping range where the slope of torque against angle is essentially constant and tension becomes proportional to the angle turned; and a post-yield zone beyond an inflection point where something — bolt, threads or joint — begins to give.
A sensor that cannot rotate freely sees the end of that. This one sees all of it. The shape of the curve, not the final figure, is what identifies what actually happened in the joint.
And the shape carries information the final figure destroys. This is the argument for instrumented fastening in one sentence: two joints can arrive at exactly the same target torque by completely different routes, and only one of them is any good. Published guidance on tightening curves is direct about the consequence — cross-threading, bad seating, a bolt stretched past its elastic limit and even a socket beginning to wear can all pass as good on a torque controller, while the deviations that reveal them are plainly visible in the trace. A good process is repeatable, and every fastening's signature should look very nearly identical to the last one.
Which is why this configuration exists. It puts a laboratory-grade measurement into the tool chain itself, between the drive and the socket, where the fastening is actually happening. What comes out is not a pass or a fail. It is a record of how the joint got there, and that is the difference between knowing a bolt reached forty foot-pounds and knowing whether it should have.
The RSS Series covers ranges from 10 through 250 ft-lbs and is used for bolt and nut fastening, automotive manufacturing, aerospace and aviation, construction, maintenance and repair, quality control, calibration services, research and development, robotics and automation, and fastening cycle analysis.
The measuring element is conventional. The things worth knowing before one is specified are all about how it sits in a fastening system.
One body, five capacities. This is unusual and it is not advertised anywhere. Every model in the range shares identical outside dimensions — the same overall length, the same widths, the same mounting geometry. Only the drive size changes as capacity rises. A fixture, a bracket, a holder or a robot end-effector designed around one RSS accepts any of the others without modification, which means a facility can standardise its tooling once and then change capacity as a drop-in. Anyone who has had to redesign a fixture because the next sensor up was a different shape will recognise how much that is worth.
The drive size is tied to the capacity. A quarter inch square at the bottom of the range, three eighths through the middle, half an inch at the top. That is mechanically sensible — a drive has to survive the torque it carries — but it has a practical consequence. Choosing a model means choosing a socket set, so a customer whose fasteners span a wide torque range will need more than one sensor regardless of measuring range, because a single drive size will not physically suit both ends of their work.
Five hundred revolutions per minute is a real ceiling. Rotation is unlimited in extent but not in speed, and 500 RPM is a deliberate figure for a sensor with bearings and sliding contacts. It is ample for controlled instrumented fastening, for hand and powered tools used at measured pace, for tool testing and for calibration work. It is not enough to sit permanently in line with a production nutrunner spinning a fastener down at full speed. If your process runs fast, the sensor belongs in the audit and validation loop rather than in the tool that does the work — and that is usually where you want the measurement anyway.
The specifications are cut from the reaction family, not from our other rotating sensor. Rated output is 2 mV/V from a 350 ohm bridge at 10 VDC, safe overload is 150 percent of rated output, zero balance is within 1.0 percent, and repeatability is 0.1 percent of rated output against nonlinearity and hysteresis of 0.25 percent. That is the ordering you want to see — repeatability better than linearity — and it means that while any single reading carries the sensor's absolute uncertainty, one curve compared against another is a far tighter comparison than the headline figures suggest. For signature work, where you are looking for a trace to deviate from a known good one, repeatability is the specification that matters and this one is good.
Temperature is close to a non-issue. Zero drifts by 0.001 percent of rated output per degree Fahrenheit, which is the lowest figure of any product in our catalogue, with output drifting at 0.005 percent of load per degree. Internal temperature compensation is built into the four-conductor wiring. A sensor that lives on a shop floor, gets carried between a heated building and an unheated bay, or sits next to a process that warms up during a shift will hold its zero through all of it.
Choosing an RSS is mostly a question of what you intend to learn, because the same sensor supports two quite different programmes and they want different capacities.
Decide first whether you are checking tools or studying joints. Checking tools means applying known torques to verify that a wrench, a driver or a nutrunner delivers what it claims, which points you at the capacities that bracket your tool crib. Studying joints means capturing what actually happens in your product's fasteners, which points you at the capacities that bracket your fastener specifications. Those two lists overlap less often than people expect, and it is worth writing both down before choosing.
Size to the peak of the curve, not to the specification. The target torque is one point on a rising trace, and several ordinary things sit above it — overshoot on a powered tool, an operator's follow-through, a stiff joint that reaches target abruptly, a fastener that binds. Leave room above your nominal figure so the interesting part of the curve is not clipped at full scale, and keep the worst case inside the 150 percent safe overload.
Account for prevailing torque separately. If your fasteners use nylon inserts, deformed threads or any other locking feature, torque appears in the rundown zone before the fastener has touched anything, and it is not doing any clamping. Published practice treats it explicitly: the elastic origin is located where the prevailing torque level meets the backward projection of the elastic slope. A target figure that does not say whether prevailing torque is included is ambiguous, and the RSS will show you exactly how much of it there is.
Remember what torque does and does not tell you. Torque is not clamp load; it is the effort needed to produce clamp load, and most of it is spent overcoming friction rather than stretching the bolt. Published analysis makes the point that the torque required to yield a bolt depends on the thread friction coefficient as much as on the material, so a change in lubrication, plating or surface condition moves the torque needed for identical clamp load. This is precisely why the angle information alongside torque is so valuable, and why a curve beats a number.
Think about what goes between the sensor and the fastener. Extensions, universal joints, swivels and wobble adaptors all add friction, backlash and compliance between the point of measurement and the joint, and all of them show up in the trace. Keep the assembly short and direct where you can, and where you cannot, keep it identical from one measurement to the next so at least the comparison stays honest.
Then choose the instrumentation, and choose it for capture. Curve work asks more of the readout than a pass/fail check does. An amplifier signal conditioner module feeding a data acquisition system is what gets you the whole trace at a sample rate fast enough to resolve the transitions between zones; a digital display with peak capture and setpoints covers verification work where a single number per fastener is genuinely all you need. Add Cal-Teds plug and play if your traces have to be traceable: an IEEE 1451.4 memory travels with the sensor, so the scale factor behind a recorded curve is never in doubt months later when somebody queries it.
Tell us the fastener sizes, the target torques, whether locking features are involved, how fast the tool turns and whether you want traces or numbers, and we will help you pick the model. Lead times on the popular capacities are short because we keep them on the shelf, and academic buyers should mention it when they enquire.
RSS Series Torque Sensor Applications.
The Transducer Techniques RSS Series rotating socket wrench torque sensors measure bolt and nut wrenching torques in ranges from 10 through 250 ft-lbs, are bi-directional for both tightening and break-away measurement, and support unlimited shaft rotation for monitoring the complete fastening cycle.
- Bolt and Nut Fastening: Manufacturing environments use these sensors to ensure that bolts and nuts are tightened to the specified torque values, preventing over- or under-tightening failures.
- Automotive Manufacturing: Critical assembly areas employ RSS sensors for engine parts, suspension systems and braking systems, where accurate torque measurements are essential for vehicle safety and performance.
- Aerospace and Aviation: These instruments are vital in the assembly of aircraft components, including fuselage sections, wings and engine components, where precision ensures reliability.
- Construction: Structural assembly applications use the sensors for monitoring and controlling the torque applied to various structural components to ensure stability.
- Maintenance and Repair: Technicians apply correct fastening specifications, preventing damage or loosening caused by improper torque application.
- Quality Control: Verification departments confirm that fasteners meet the required torque specifications, maintaining product quality and safety standards.
- Calibration Services: Laboratories employ RSS sensors as reference standards for calibrating torque wrenches, testers and other torque measurement devices.
- Research and Development: Engineers study fastener behavior and optimize fastening methodologies using these measurement tools.
- Robotics and Automation: Integrated into robotic systems to monitor and regulate torque applied by robotic arms and end-effectors for consistent operation.
- Fastening Cycle Analysis: The unlimited rotation capability supports comprehensive monitoring of the entire fastening cycle, enabling process optimization.
RSS Series sensors use bonded foil strain gauges of the highest quality along with coined silver slip rings for data transmission, and are calibrated clockwise and counterclockwise on every model.
Frequently Asked Questions
What torque range and drive sizes does the RSS Series cover?
Five models from 10 through 250 ft-lbs, with the square drive tied to the capacity: a quarter inch on the two smallest, three eighths in the middle, half an inch at the top. Because the drive is fixed by the model, choosing a capacity also chooses a socket set. If your fastener range is wide enough to need both a quarter inch and a half inch drive, you are looking at more than one sensor whatever the torque figures say, since no single unit will physically fit both ends of the work.
What does unlimited rotation let me measure that a fixed sensor cannot?
The whole fastening, rather than its result. Ball bearing support and slip rings mean the drive can turn indefinitely in either direction, so the sensor stays connected and reading from the first thread engagement through to final torque. That gives you a continuous trace rather than a value, and the trace is what tells you how the joint got where it is. A sensor restricted to a limited arc can tell you what the torque was at the end. It cannot tell you what the tightening looked like on the way there.
What are the phases of a fastening cycle?
Published tightening practice recognises four zones. First a rundown zone, before the fastener contacts its bearing surface, where any prevailing torque appears — from thread-locking features, misalignment, debris or tolerance problems. Then an alignment and snugging zone, where the parts are drawn together and coatings and thread flanks deform slightly until the joint is stably clamped. Then the elastic clamping range, where the slope of torque against angle is essentially constant and tension is proportional to the angle turned from the elastic origin. Finally, past an inflection point, a post-yield zone where the bolt, the threads or the joint itself starts to give. Each zone looks different on a trace, and knowing which one a reading came from is often more useful than the reading.
What faults show in a curve but not in a final torque reading?
A striking number of them. Published guidance on tightening traces lists cross-threading, bad seating, a bolt stretched past its elastic limit and even a socket beginning to wear as conditions that can all pass as acceptable on a torque controller while being obvious in the trace. The reason is simple: those faults change the route to the target, not the target. A cross-threaded fastener shows a torque rise where there should be free rundown. A missing or misplaced washer changes where seating occurs. A joint that has yielded shows a slope that flattens instead of climbing. All of them can finish at exactly the specified number.
What is prevailing torque and how should I handle it?
It is the torque needed to turn a fastener before it has begun to clamp anything — produced deliberately by nylon inserts and deformed threads, or accidentally by misalignment, damage or debris. It appears in the rundown zone and it does no clamping whatever. That matters because it is included in any torque a tool reports, so a target figure that does not state whether prevailing torque is added to it or included in it is genuinely ambiguous and two shops will interpret it differently. Published practice handles it by locating the elastic origin where the prevailing torque level meets the backward projection of the elastic slope. An RSS shows you how much prevailing torque your fasteners actually carry, which is usually the first thing worth measuring.
How accurate is the RSS Series?
Nonlinearity and hysteresis are each 0.25 percent of rated output, repeatability is 0.1 percent, and zero balance is within 1.0 percent. The ordering is the right way round, with repeatability better than linearity, and that matters more than the headline number for the work this sensor does. Comparing one trace against another is a repeatability question rather than an absolute accuracy question, so signature analysis benefits from the tightest figure on the sheet. Put it in context too: hand and powered torque tools are typically specified in whole percentage points, so the sensor has ample margin over the things it is used to check.
What does the 500 RPM limit rule out?
Rotation is unlimited in how far it goes but capped in how fast, because there are bearings and sliding contacts involved. Five hundred revolutions per minute suits hand tools, controlled powered tools, tool verification, calibration benches and instrumented fastening at a measured pace. It does not suit sitting permanently in line with a production nutrunner running a fastener down at full speed, and it rules out impact tools entirely — those deliver hammer blows rather than continuous torque, which is both outside the speed rating and a poor input for a strain gauge sensor. If your process is fast, put the RSS in the audit and validation loop rather than inside the tool.
Do all five models share the same body?
Yes, and it is one of the more useful things about the series. Every dimension on the outside is identical across the range — only the drive size changes. Anything built around one of them, whether a bench fixture, a holding bracket, a guard or a robot end-effector, will accept any other model without alteration. That lets a facility design its tooling once and then move up or down in capacity as a straight swap, which is unusual and is worth knowing before somebody designs five different fixtures.
How does temperature affect an RSS reading?
Very little. Zero drift is 0.001 percent of rated output per degree Fahrenheit, the lowest figure in our catalogue, and output drift is 0.005 percent of load per degree, with compensation built into the wiring. The compensated range comfortably covers ordinary industrial conditions and the safe range extends well past them. In practice that means a sensor carried out to a cold bay, or used beside a process that warms through a shift, holds its zero rather than needing constant re-zeroing — which is exactly what you want from something used for auditing, where re-zeroing between fasteners is not always practical.
Is the RSS calibrated in both directions?
Yes, clockwise and counterclockwise on every model as standard. That is not a formality for this product, because half of what it is used for happens in the loosening direction — break-away measurement, removal torque, checking that a joint can be serviced, and verifying that a locking feature still has grip left in it after a cycle. A sensor certified in one direction only would leave that work uncalibrated. Both directions are covered here without asking.
Questions From The Field
Every joint reaches target torque but parts still come loose in service.
Then torque is not the problem and torque was never going to reveal it. What holds a joint together is clamp load, and torque is only the effort spent producing it — most of which goes into friction rather than into stretching the bolt. Change the lubrication, the plating, the surface finish or the thread condition and the same torque produces a different clamp load. Capture the full traces from a batch and look at the slope of the elastic region rather than the endpoint: a joint reaching target with an unusually steep slope has spent its torque on friction, and there is less tension in that bolt than the number suggests. That is the failure you are describing, and it is visible in a curve and invisible in a value.
The first fastenings of each shift look different from the rest.
Usually real, and usually not the sensor. Cold lubricant is more viscous, so the friction component is higher and the trace is steeper until things warm up. Tools behave differently on their first cycles. And the operator has not settled into a rhythm yet. Establish whether the difference is in the rundown zone, which points at lubricant and thread condition, or in the elastic slope, which points at friction under load. The sensor itself is a poor candidate given its temperature figures, but if you want to eliminate it, capture the same joint at the start and end of a shift with the tool held constant.
The trace rises, flattens off, then rises again. What am I looking at?
Most often a seating event you did not expect — a coating crushing, a gasket compressing, a washer bedding, a part pulling into alignment. Torque climbs while something is being deformed, levels while it gives way, then climbs again once the joint has genuinely closed. Occasionally it is something less welcome, such as a thread stripping progressively or a component yielding. The distinguishing question is whether it happens in the same place on every fastener. Consistent and repeatable across a batch is the joint's normal behaviour and belongs in your reference trace. Appearing on some fasteners and not others is a fault worth chasing.
Can I put an RSS in line with a powered nutrunner or an impact wrench?
A controlled powered tool running within 500 RPM, yes, and that is a genuinely useful arrangement for validating what the tool actually delivers into a real joint. An impact wrench, no. Impacts work by delivering rapid hammer blows rather than steady torque, which is both outside the speed rating and a poor thing to put through a strain gauge element and a bearing. For impact tools the sensible approach is to audit the result afterwards rather than to instrument the tool, and we can advise on how to set that up.
Bench readings are fine but the sensor disagrees with the tool's own display in the cell.
Expect some disagreement and be suspicious only if it is large or inconsistent. The tool reports what it produces at its own output; the sensor reports what arrives after everything in between. Extensions, universal joints, swivels, adaptors and the socket itself all absorb a little and add friction, and the tool's own transducer sits in a different place in the chain. Look for a consistent offset, which is the assembly and is characterisable, versus a scattered difference, which usually means backlash or something moving in the drive train. Keep the intervening hardware identical between measurements and the comparison becomes meaningful.
Two identical joints reach the same final torque with different curves. Which one is right?
Neither is guaranteed right, and the fact that you can ask the question is the reason this sensor exists. A repeatable process produces traces that overlay almost exactly; two traces that differ mean two different things happened, regardless of where they finished. Look at where they separate. Divergence in the rundown zone points at thread condition, debris or a locking feature behaving inconsistently. Divergence in the snugging zone points at seating, a missing or doubled component, or surfaces not meeting as intended. Divergence in the elastic slope points at friction, which means lubrication or surface condition. The endpoint agreeing is a coincidence of two different journeys.
I need to measure a rotating fastening above 250 ft-lb. What are my options?
Two hundred and fifty foot-pounds is the top of this series, so above it the choices are to measure without continuous rotation, using a reaction-style socket sensor which covers a considerably higher range but does not follow the full cycle, or to talk to us about what is available for large fasteners specifically. Which route suits depends on whether you need the curve or the number. If you are auditing already-tightened joints the reaction approach is entirely appropriate; if you need the tightening signature at high torque, tell us the fastener and the application and we will go through it with you.
Does it matter what I put between the sensor and the fastener?
It matters more than most people expect. Every extension, universal joint, swivel and wobble adaptor adds friction, backlash and a little springiness between where the measurement happens and where the joint is, and all three affect the trace — friction shifts the level, backlash blurs the transitions, compliance softens the slope. Keep the assembly as short and as direct as the access allows. Where the geometry forces you to use adaptors, use the same ones every time and record what they were, so that even if the absolute figure carries some overhead, the comparison from one fastening to the next stays valid.