Strain gauge measurement works by letting a piece of metal be twisted or stretched by the load you care about, and reading that deflection electrically. Foil gauges follow that deflection as a change in their own resistance; a bridge circuit converts the imbalance into a few thousandths of a volt, and the calibration certificate supplies the factor that turns those millivolts into inch-pounds.
What separates one torque product from another is not that principle but where the sensing body is put. Every other torque sensor we make is a bystander — bolted to a fixture, a wall or a machine's foot, watching a process it takes no part in. A rotating sensor is not a bystander. It is a shaft in your machine.
It carries the load, and it turns. An RST goes in-line: the driving end couples to one side, the driven end to the other, and every inch-pound that reaches the load passes through the sensing element on its way. There is nothing to hold still and nothing to react against. A published description puts the relationship neatly — a rotary sensor is a reaction sensor that has been allowed to rotate as part of the running system.
Which raises the obvious problem: the gauges are spinning and the instrument is not. The signal has to cross from a rotating body to a stationary one, continuously, without a wire wrapping itself round the shaft. On the RST that crossing is made mechanically, by coined silver slip rings running against silver brushes. The rings turn with the shaft, the brushes stay put, and contact is maintained through every revolution for as long as the machine runs.
That single design decision explains most of this product's character. Sliding contact is simple, proven and unfussy about installation, and it works from a dead stop — there is no minimum speed and no start-up sequence. It is also the reason the accuracy figures here sit a step behind our reaction sensors, the reason there is a speed ceiling at all, and the reason this is the one sensor in our catalogue with parts that wear. None of those are defects. They are the price of a measurement no stationary sensor can make.
And what it buys is the real number. A reaction sensor tells you what a machine's mounting is holding. An in-line rotary sensor tells you what is actually arriving at the load, while the shaft turns, at speed, under the real conditions of the test — through gearboxes, past couplings, after losses. For engine and transmission work, dynamometer testing, drivetrain characterisation and machinery development, that is not a nicer way of getting the same answer. It is the only way of getting that answer at all.
The RST Series covers eleven full scale ranges from 10 through 30,000 inch pounds and is used for automotive testing, aerospace and aviation, manufacturing and assembly, research and development, material testing, calibration services, industrial machinery, energy and power generation, robotics and automation, and dynamic testing.
The bridge behaves as bridges do. What is worth understanding before an RST is specified is that the eleven models are not one product in eleven sizes. They sit on four separate bodies, and moving between them changes far more than the number on the label.
Capacity picks the body, and the body picks everything else. The A body covers the three lowest ranges, the B body the next four, the C body two, and the D body the two largest. Between the smallest A and the largest D, torsional stiffness rises by a factor of several thousand and weight rises from a few pounds to something you would not lift one-handed. Those are not variations on a theme; they are four different mechanical propositions that happen to share a specification sheet. Take your figures from the model, never from the series.
The speed ceiling does not simply fall as capacity rises. It is easy to assume bigger means slower, and mostly that holds — published guidance on rotating sensors attributes the pattern to weight and diameter, noting that as capacity goes up a sensor generally has to spin more slowly. But it is not a straight line on this series, and the middle of the range is not the slowest part of it. If speed matters in your test, read the rating for the exact model rather than inferring it from where the model sits in the list, and ask us if the answer surprises you.
Torsional stiffness is a specification you are inserting into your own drivetrain. This is the figure people skip and later wish they had not. A rotating sensor is a torsional spring placed between two masses that were previously connected more directly, and together they set a natural frequency. Get it right and the sensor faithfully follows the torque fluctuations you are trying to capture. Get it wrong and you have built a resonance into the driveline that was not there before — and the interesting thing about a torsional resonance is that it does not announce itself as a measurement error. It looks like real data.
The sensor's own mass is now rotating mass. Whatever the body weighs, your machine has to accelerate it. On a steady-state test that costs nothing. On rapid transients, start-up characterisation or anything where the drivetrain's inertia is part of what you are studying, a sensor added to the shaft has changed the system you are measuring. Worth accounting for deliberately rather than discovering afterwards.
Signal and wiring. Full scale output is 1.5 mV/V from a 350 ohm bridge at 10 VDC, on four colour-coded conductors. Insulation resistance is high — thousands of megohms bridge-to-ground and shield-to-ground — which matters more here than on a bench sensor, because this one lives in a test cell surrounded by drives, inverters and heavy switching, and a clean shield is what keeps that out of a millivolt signal.
Specifying a rotating sensor involves more of your machine than specifying a stationary one, because the sensor is joining the machine rather than watching it. Our application engineers would far rather see the drivetrain arrangement, the speed range and the worst-case torque before anything is ordered.
Start with the peak, and be pessimistic about it. The RST's safe overload is 120 percent of full scale. That is deliberately less headroom than our reaction sensors carry, and it is the single most important number on the page for anyone connecting one to a live drivetrain. A driveline delivers shocks that a fixture never does — direct-on-line starts, gear engagement, a load that snatches, a sudden stop, a jam. Published selection guidance for rotating sensors is blunt about this: peaks and spikes outside the safe range damage the sensor, so accommodate all of them when choosing. Size to the worst event your drivetrain can produce, not to its normal running torque.
Then check the speed, for that exact model. Running above the rating is not a gradual degradation; it is a mechanical limit on a spinning assembly. If your machine's top speed is close to a body's ceiling, that usually settles which model you are buying regardless of what the torque figure suggests.
Then decide what the couplings have to do. They are not hardware; they are part of the measurement. Their job is to pass torque faithfully while absorbing the misalignment that exists between three separately mounted machines. Published guidance distinguishes single-flex couplings, which allow angular misalignment only, from double-flex, which allow angular and radial — and recommends flexible disc types for their high torsional rigidity and zero backlash, with balancing available for high speed work. Backlash and softness in a coupling both end up in your data.
Ask honestly whether this is a test or an installation. Sliding contacts wear. Published guidance on slip ring sensors is direct — they are generally not recommended for continuous or very high speed duty, brush wear generates dust that degrades the signal, and periodic cleaning and brush replacement are part of owning one. Manufacturer guidance adds that brush dust mixed with oil can settle on the rings and cause electrical shorts. For a test cell, a development rig, a dynamometer or a calibration bench, all of that is routine and entirely manageable. For a sensor that has to run unattended on a production machine for years, tell us and we will point you at a different approach.
Confirm the direction of calibration before you order. If your test turns both ways — forward and reverse drive, motoring and braking, tightening and loosening — say so explicitly on the enquiry rather than assuming, and we will confirm exactly what your unit will be certified for.
Then choose the instrumentation. Most rigs want the torque channel logged next to speed, which means an amplifier signal conditioner module feeding your data acquisition or dynamometer controller. Where somebody is standing at the rig watching a number, or where reaching a limit has to trigger something, a digital display with setpoints does the job on its own. The Cal-Teds plug and play option puts the calibration data on an IEEE 1451.4 memory in the connector, which earns its keep in a facility running several sensors through the same instruments.
Send us the torque range, the top speed, the worst-case shock the drivetrain can deliver, the shaft ends you have to couple to and how many hours a week it will run, and we will help you settle on a model. The commonly used capacities are held in stock, and there are discounts for academic and research buyers.
RST Series Torque Sensor Applications.
The Transducer Techniques RST Series rotating torque sensors are available in eleven full scale ranges of 10 through 30,000 inch pounds, and are used wherever torque has to be measured on a shaft while it turns.
- Automotive Testing: Measures torque on vehicle engines, transmissions, and drivetrain systems for performance and fuel efficiency optimization.
- Aerospace and Aviation: Employed to measure aircraft engine torque and control surface forces for safety and performance verification.
- Manufacturing and Assembly: Quality control tool ensuring fasteners and components meet precise tightening specifications.
- Research and Development: Assesses rotating machinery performance, tests prototypes, and studies material behavior under torque loads.
- Material Testing: Characterizes mechanical properties of materials subjected to torsional forces.
- Calibration Services: Functions as reference standards for calibrating other torque measurement devices.
- Industrial Machinery: Monitors and controls torque in mixing, blending, and conveyor systems for consistent operations.
- Energy and Power Generation: Monitors torque in turbines, generators, and pumps for safe, efficient production.
- Robotics and Automation: Measures and controls robotic arm torque application for precise, safe movements.
- Dynamic Testing: Measures torque variations over time in dynamic systems and machinery studies.
RST Series sensors use bonded foil strain gauges of the highest quality along with coined silver slip ring and silver brushes for data transmission, and are supported by our full range of amplifiers, displays and accessories.
Frequently Asked Questions
What torque range does the RST Series cover, and how is it organised?
Eleven full scale ranges from 10 through 30,000 inch pounds, arranged on four distinct bodies. The A body carries the three smallest ranges, the B body the next four, the C body two and the D body the two largest, and the model number tells you which — the letter after RST is the body. That letter matters as much as the number after it, because speed rating, torsional stiffness and weight all belong to the body rather than to the series. Two models that look adjacent in a price list can be very different objects to install.
Why should I read the speed rating for my exact model rather than the series?
Because the ceiling is set by the physical assembly — its mass, its diameter and how it is balanced — rather than by its torque rating, and those do not track each other neatly across four different bodies. The general rule in rotating sensor selection is that higher capacity means lower permissible speed, driven by weight and diameter, but the RST does not follow that as a straight line and the fastest body is not the smallest one. Look up the number for the model you intend to buy. If your application sits near a ceiling, talk to us before ordering rather than after.
What does torsional stiffness change from one RST to another?
Across the full range it changes by a factor of several thousand, and it decides what your drivetrain does after the sensor is fitted. Stiffness plus the inertia either side of it sets a torsional natural frequency; if that frequency lands near something your machine already produces — a firing order, a gear mesh, a pump's blade passing rate — you have created a resonance by installing an instrument. A stiffer sensor pushes that frequency higher and follows fast torque changes more faithfully; a softer one is more forgiving of the drivetrain but slower to respond. If you are measuring transients or torsional vibration rather than steady running, this figure deserves more attention than the accuracy specification does.
How much does an RST weigh, and why does it matter here?
From a few pounds on the smallest body to around forty on the largest. On a stationary sensor weight is a handling question. On this one it is a measurement question, because that mass is now rotating mass your machine has to accelerate and decelerate. Steady-state testing is unaffected. Start-up characterisation, rapid transients, and any test where drivetrain inertia is part of what you are investigating are affected, and the honest position is that you are measuring a system that now includes the sensor. Weight also drives the shaft support and the speed rating, which is why the heaviest bodies are not the fastest.
How accurate is the RST Series?
Nonlinearity, hysteresis and repeatability are each specified at 0.25 percent of full scale, with calibration accuracy at the same figure and zero balance within 2.5 percent of full scale. Temperature effect on zero is low, at 0.0025 percent of full scale per degree Fahrenheit. Those numbers are a step behind our reaction sensors and that is inherent rather than accidental — sliding electrical contact adds a small varying resistance in series with a millivolt-level signal, and no rotating contact sensor escapes it. Put it in proportion: for engine, transmission and drivetrain work, where the process being measured varies by whole percentage points between runs, a quarter of one percent is comfortably below the variation you are trying to resolve.
Why is the safe overload only 120 percent?
Because a rotating element under load has less margin available to it than a solid stationary body does. It is the tightest overload figure in our torque range and it sits on the product most exposed to shock, which is why it deserves a deliberate answer rather than a glance. Treat 120 percent as a survival limit, not a working allowance, and size the sensor against the worst thing your drivetrain can do rather than what it normally does. There is a separate, much higher figure quoted for electrical failure; that is the point of destruction, not a second safe zone.
What output does an RST give and what will drive it?
1.5 mV/V at full scale from a nominal 350 ohm bridge on 10 VDC excitation, over four colour-coded conductors. That is a slightly lower sensitivity than our reaction sensors give, so front-end gain should be sized for it rather than assumed. Insulation resistance is specified in thousands of megohms both bridge-to-ground and shield-to-ground, which is worth having in a test cell full of variable speed drives — ground the shield properly at the instrument end and keep the run away from motor leads, because electrical noise is a far more common cause of a disappointing rotary measurement than the sensor is.
How does temperature affect an RST reading?
Zero shifts by 0.0025 percent of full scale per degree Fahrenheit, which is a good figure and better than several of our stationary products manage. The practical caution is that a rotating sensor lives in a warmer and less stable place than a bench instrument — a test cell heats up as the machine runs, and the sensor sits in the middle of it, next to the source. Let the rig reach a stable temperature before taking data you intend to trust, and re-zero when it has. If your test cell runs hot or cold enough that you are checking the sensor's temperature limits against it, send us the numbers and we will confirm what is appropriate.
What does an RST need around it to be installed?
Two couplings, proper shaft alignment, adequate support for its weight, and space to get at it. The couplings are the part people underestimate: they have to transmit torque without adding backlash or compliance while absorbing the misalignment that always exists between separately mounted machines. Published guidance recommends flexible disc types for their high torsional rigidity and zero backlash, notes that single-flex accommodates angular misalignment only where double-flex handles angular and radial, and points out that they can be balanced for high speed work. Get the alignment right mechanically first; couplings are there to accommodate what is left, not to rescue a poorly aligned installation.
Is the RST suitable for permanent installation on production plant?
It is best suited to test cells, development rigs, dynamometers and calibration work — places where a scheduled maintenance visit is normal and someone competent is around. Slip ring sensors are generally not recommended for continuous or very high speed running, because the brushes wear, the wear produces dust, and the dust degrades the signal until the rings are cleaned and the brushes replaced. In a test facility that is an ordinary service item. On a machine expected to run unattended for years it is a liability, and there are better approaches for that duty. Tell us which situation you are in and we will be straight with you about whether this is the right sensor.
Questions From The Field
My torque sits right on the boundary between two bodies. Which way should I go?
Let the rest of the application decide, because the torque figure alone is telling you the two are equivalent and everything else says they are not. Go up if your drivetrain can deliver shocks, if you want overload margin, or if you need the stiffness for transient work. Go down if speed is tight, if the added rotating mass would disturb the test, or if space and shaft support are constrained. Write down your top speed, your worst-case peak and what you are trying to resolve, send those three things over, and we will tell you which side of the boundary your application actually belongs on.
The drivetrain has developed a vibration that was not there before the sensor went in.
You have most likely moved a torsional natural frequency into your running range. The sensor added a spring and two lumps of mass to a driveline that previously had neither in that position, and the combination resonates somewhere — if that somewhere is a speed you run at, the driveline will tell you loudly. Check whether the vibration appears at a particular speed rather than across the range, which is the signature. Remedies are to change the stiffness by moving to a different body, to change the inertia with different couplings or adapters, or to shift the running speed. Send us the arrangement and the speed it happens at and we will help work out which lever to pull.
My machine runs faster than any RST body is rated for. What are my options?
Do not run it over the rating and hope — that limit is mechanical, and the failure mode on a spinning assembly is not a gradual loss of accuracy. Realistic options are to gear down and measure on the slower side of the reduction while accounting for the ratio and the losses, to measure the reaction at the machine's mounting instead if the test is steady state, or to move to a sensor built for the speed. Which one is right depends entirely on what you are trying to learn, so tell us the test rather than just the speed.
My peak torque is inside the rating but sensors keep coming back damaged.
Then the peak you know about is not the peak the sensor is seeing. Drivetrains produce transients far shorter and larger than anything a meter or a logger will show at ordinary sample rates — gear engagement, a load snatching, a direct-on-line start, an emergency stop, backlash taking up under power. Any of those can pass 120 percent of full scale in milliseconds while your recorded trace looks entirely reasonable. Look also for a bending or axial load being fed in through misaligned couplings, which loads the sensor in a direction it was never sized for. Capture at a high rate, look at the start and stop events rather than the steady running, and check the alignment properly.
The zero shifts every time I stop and restart the machine.
Check the mechanical arrangement before suspecting the sensor. A drivetrain that has been stopped under load can leave residual wind-up trapped in the couplings and the shafts, which reads as a genuine offset because it is a genuine torque. Temperature is the other common cause — a rig that has been running is warmer than one that has just started, and the zero moves with it. Establish a routine: bring the machine to a stable temperature, break the driveline free of any residual load, then zero. If a shift persists with the couplings disconnected and the rig at a steady temperature, that is worth reporting to us.
Can I bolt an RST straight to my shaft flanges and skip the couplings?
Please do not. Two shafts on separately mounted machines are never perfectly aligned, and a rigid connection converts that misalignment into a bending moment and a side load applied directly to the sensing element and its bearings, every revolution, for the whole test. Off-axis forces of that kind are recognised as a major error source on rotating torque sensors and they will damage the sensor as well as spoil the data. The couplings exist precisely to absorb what your alignment cannot, and they are much cheaper than the sensor.
Does an RST wear out, and what should I plan for?
The sensing element does not, but the sliding contacts do — that is the trade for being able to measure a turning shaft. Plan for periodic inspection and cleaning of the rings, brush replacement as an ordinary consumable, and recalibration on your normal cycle. How often depends almost entirely on running hours and speed, so a sensor used for a few hours of testing a week is a different proposition from one turning continuously. Keep a note of hours run alongside the calibration records, and it stops being a surprise. If you would like an estimate for your own duty cycle, tell us the speed and the hours and we will give you a realistic view.
My test runs in both directions. Is that a problem?
It is not a problem, but it is a thing to confirm rather than assume. Bidirectional data on a rotating sensor is entirely normal — forward and reverse drive, motoring and braking, tightening and backing off. Say so on your enquiry so we can be explicit about what your unit will be certified for and in which directions, and so the certificate you receive matches the test you intend to run. It is a five minute conversation before the order and an awkward one afterwards.