Ask what a load cell does and the short answer is arithmetic on a deflection: gauges bonded to a machined element register the few ten-thousandths of an inch that load produces, a Wheatstone bridge expresses that as millivolts, and a certificate written against reference standards expresses the millivolts as pounds.
Almost all of that description assumes force pushes one way. Most measurement does: a press compresses, a rope pulls, and zero sits at one end of the working range like the bottom of a scale. A meaningful number of applications are not like that at all. An actuator that drives and retracts, a fatigue test that alternates, a linkage that carries load in both directions, a mechanism that pushes on one part of its cycle and pulls on the next — in all of these the force reverses, and the reading has to follow it across the change of sign.
That is what “through zero” means, and it moves zero from the edge of the measurement into the middle of it.
Zero stops being a starting point and becomes a place you keep visiting. On a compression-only sensor you set zero once and work upward from it; a small offset shifts every reading by the same trivial amount. When your load swings from pull to push, you pass through zero on every cycle, and a zero that has moved does not shift your readings uniformly — it makes tension read differently from compression. That asymmetry is far more troublesome than a plain offset, and it is why zero balance and zero stability deserve more attention on a bidirectional job than on any other kind.
The crossing exposes the load path. This is the part that catches people out, and it is usually not the sensor's doing. Any clearance in the assembly — thread slack, a pin in a slightly oversized hole, a joint that only bears in one direction, a fitting that was merely tight — unloads and reloads as the force reverses. The sensor faithfully reports a short interval where very little is being transmitted, and a curve that should pass smoothly through zero develops a flat spot or a step instead. A bidirectional installation has to be built without slack, in a way a one-directional one never needs to be.
Hysteresis is charged to you repeatedly. On a single loading and unloading the difference between the up curve and the down curve costs you once. On a reversing test you traverse that loop continuously, so the same specification represents an error you meet on every cycle rather than once per run.
Symmetry is the property to look for. What makes a sensor good through zero is that it behaves the same in both directions — comparable sensitivity pulling and pushing, so the response continues smoothly across the crossing instead of changing character at it. That is what good tension and compression compliance describes, and it is a design property rather than something an installation can add.
Sensors built this way suit material testing in both directions, reversing actuator and linkage measurement, fatigue and durability work, assembly line force verification, aerospace and automotive structural validation, and any rig where the force genuinely changes sign.
The measuring principle is standard. Three aspects of this design are worth explaining properly, because two of them are unusual and the third decides whether the part fits at all.
Signal. The bridge takes 10 VDC of excitation and returns 2 mV/V at rated load — about twenty millivolts at full load, at any capacity in the range, expressed as a ratio of the supply rather than as a fixed voltage. A conditioner and the calibration data on file turn that into engineering units, on the assumption that the excitation in service matches the excitation at calibration.
One sensing element can carry more than one bridge. This is the option most people overlook, and it is worth understanding rather than skipping. A second complete Wheatstone bridge, with its own set of strain gauges, can be installed on the same element — giving two electrically independent measurements of the same force. Manufacturers of dual bridge sensors describe two reasons to want that. The first is redundancy: if one channel fails, the other keeps the system running until the fault can be dealt with, which matters where an interruption is expensive or the sensor is difficult to reach. The second is integrity checking: two outputs that should agree can be continuously compared, so a developing fault announces itself as a disagreement rather than as quietly wrong data. Redundant measurement channels of this kind are the construction that functional safety architectures are built on, and this class of sensor is used that way in aerospace and automotive structural testing and monitoring.
A third bridge answers a question two cannot. With two channels, a disagreement tells you something is wrong but not which one to believe, so the safe response is usually to stop. With three, the two that agree identify the one that does not, which is what allows a system to keep operating through a single channel failure rather than shutting down. That is the whole reason a triple bridge configuration exists, and it is why it belongs on installations that must not be interrupted for a fault that has not actually affected the measurement.
Height is the design constraint everything else answers to. An ultra low profile body puts the sensing structure into a fraction of the height a conventional design of the same capacity needs — under an inch and a half at 500 lb, under three inches at 20,000 lb. That matters when a sensor has to go into a fixture, a press, a test frame or a machine that was designed without one, where the available gap decides feasibility and no amount of accuracy compensates for a part that does not fit. Deflection and natural ringing frequency are both published for each model, running from 2,100 Hz at the smallest capacity to 10,000 Hz at the largest, so take both from the row you intend to order rather than from the series.
Specifying a low profile bidirectional sensor comes down to four questions, and the answers to two of them are frequently already fixed by the machine before anyone starts looking at datasheets. Talk it through with our application engineers rather than working from the table alone.
Measure the gap first. If a sensor has to fit an existing fixture, the available height rules everything else. Establish it before capacity, because it eliminates options faster than any other criterion and there is no point specifying a part that cannot be installed. Take the height from the specific model, since it grows with capacity across the range.
Establish whether your force changes sign. If it does, say so at the enquiry — it affects the calibration you need, how the fixture should be built and what you should expect at the crossing. If it does not, a bidirectional sensor is still a perfectly good choice, but you should not pay for capability the job does not use.
Then decide whether one measurement is enough. This is the question the specification table cannot answer for you. Ask what happens if this sensor fails silently — whether anyone would know, how long the bad data would flow, and what it would cost to find out late. Where the honest answer is uncomfortable, a second bridge gives you a channel to check against; a third gives you the ability to keep running when two agree and one does not. Where the sensor is on a bench and someone is watching it, a single bridge is the right and cheaper answer.
Match the certificate to the direction that matters. Compression calibration is included, and tension calibration is an available option. On a genuinely bidirectional application it is worth discussing with us which certification your results actually have to rest on, because that is a question about your reporting requirements as much as about the hardware.
Sort out the connection when you order, not afterwards. A twist-lock connector is available in place of the standard cable, which is worth having wherever the sensor will be removed for recalibration or swapped between fixtures. If you choose it, the mating assembly is a separate part and needs adding to the same order.
Then choose the instrumentation to match the channel count. A single bridge needs one amplifier signal conditioner module or one digital display. A dual or triple bridge needs conditioning for each channel, and if the point of the extra bridges is redundancy, those channels should be genuinely independent rather than sharing a single point of failure downstream — otherwise you have paid for redundancy and given it away again. Cal-Teds plug and play suits fixtures where sensors are interchanged and calibration data should travel with the part.
Tell us the height you have, the forces in both directions and what depends on the measurement, and we will help you settle on a configuration. Standard capacities are held for quick despatch, and there is discounted pricing for teaching and research.
LPO Series Load Cell Applications.
The Transducer Techniques LPO Series ultra low profile load cells, offering tension and compression measurement in a compact form, serve diverse industrial and research purposes.
- Material Testing: LPO Series load cells are used in laboratories for the evaluation of tensile and compressive properties across various material types.
- Quality Control: These load cells verify force and load compliance in aerospace and automotive manufacturing.
- Product Testing: LPO Series load cells are used for load testing, performance evaluation, and durability assessment.
- Industrial Automation: These load cells provide force monitoring and control within manufacturing and assembly systems.
- Assembly Line Testing: LPO Series load cells verify components and assemblies against design specifications.
- Research and Development: Engineers use these load cells for prototyping evaluation and structural integrity assessment.
- Custom Machinery: LPO Series load cells are integrated into specialized testing and manufacturing equipment.
- Through-Zero Applications: These load cells suit measurement scenarios involving simultaneous tension and compression forces.
- Aerospace Testing: LPO Series load cells are used for structural validation of aircraft components.
- Educational Laboratories: Educational institutions use these load cells for student instruction in force measurement principles and mechanics.
The LPO Series offers a compact yet reliable solution for precise force measurement in a wide range of industrial and scientific applications, contributing to improved product quality, safety, and operational efficiency.
Frequently Asked Questions
What does "ultra low profile" mean in practice on the LPO?
It means height has been minimised deliberately, and the numbers are the argument: 1.450" tall at 500 lb, 1.625" at 1,000 lb, 1.700" at 3,000 lb, 1.950" at 5,000 lb and 2.950" at 20,000 lb. A conventional tension and compression design of the same capacity is typically half again as tall or more. Where a sensor has to be introduced into a press, a fixture or a machine that was not designed with one in mind, the available gap usually decides the purchase before any other specification is consulted.
What capacities does the LPO Series cover and what do they cost?
Seven models: LPO-500 and LPO-1K at $825, LPO-2K at $890, LPO-3K at $985, LPO-5K at $1,115, LPO-10K at $1,430 and LPO-20K at $1,750. Unlike some of our series the price rises steadily with capacity, so there is a real cost to specifying higher than you need — and since accuracy figures are percentages of rated output, over-specifying costs you resolution as well as money.
What is a through-zero application?
One where the force reverses direction during normal operation, so the reading crosses from tension into compression or back again rather than staying on one side. Reversing actuators, alternating fatigue tests, linkages that push on part of a cycle and pull on the rest are all through-zero work. It matters because zero stops being the bottom of your scale and becomes the middle of your working range — which puts the emphasis on symmetrical behaviour either side of it and on a fixture built without slack for the load to take up as the direction changes.
What is the Dual Bridge option and why would I want it?
Option-DB, at $575, fits a second complete Wheatstone bridge with its own strain gauges to the same sensing element, giving two electrically independent measurements of the same force. There are two reasons to specify it. Redundancy: if one channel fails, the other keeps the measurement alive until the fault can be dealt with — valuable where the sensor is buried in a machine or an interruption is expensive. Integrity checking: two outputs that should agree can be compared continuously, so a developing fault shows up as a disagreement instead of as data that is quietly wrong. Each bridge needs its own conditioning.
What does the Triple Bridge option add over the Dual Bridge?
Option-TB, at $775, fits three independent bridges rather than two, and the difference is diagnostic rather than merely additional. With two channels a disagreement tells you something is wrong but not which reading to trust, so the safe response is generally to stop. With three, the two in agreement identify the odd one out, which means a system can continue operating through a single channel failure instead of shutting down. If you are building toward a redundancy requirement rather than simply wanting a spare, this is the configuration to discuss with us.
How accurate is the LPO Series?
Nonlinearity and hysteresis are each 0.1% of rated output, nonrepeatability 0.05%, and zero balance 1.0%. Those figures hold across the whole range and are considerably tighter than our thru-hole and tension link series. On bidirectional work, weigh the hysteresis figure carefully: a reversing test traverses the loading and unloading curve on every cycle rather than once per run, so it represents a recurring error rather than a single one.
What is the natural frequency and how much does the LPO deflect?
Both are published per model. Ringing frequency runs 2,100 Hz at 500 lb, 2,850 Hz at 1,000 lb, 3,150 Hz at 2,000 lb, 4,500 Hz at 3,000 lb, 6,250 Hz at 5,000 lb and 10,000 Hz at both 10,000 and 20,000 lb. Deflection is 0.003" on the 500 through 2,000 lb models and 0.005" from 3,000 lb upward. Higher capacities are stiffer structures, which is why they ring faster. For fatigue and cyclic work compare your test frequency against the figure for your specific model before ordering.
What is the LPO made from?
Heat-treated 17-4 PH stainless steel across every capacity — a precipitation-hardening grade chosen for strength and dimensional stability, which is what allows a short body to carry these loads without excessive deflection. Consistent material throughout the range also means the behaviour you characterise on one capacity carries over sensibly if you later move to another.
What threads and dimensions should I design to?
Thread is 1/2-20 UNF on the 500 through 3,000 lb models, 3/4-16 UNF on the 5,000 and 10,000 lb, and 1-14 UNS on the 20,000 lb. Overall length runs from 3.000" to 4.562" and width from 1.000" to 3.000", with height as listed above. Because the body is rectangular rather than square, take length, width and height separately from the row for your model rather than assuming a single dimension describes it.
Is the LPO calibrated for tension or compression?
Compression calibration is included as standard and tension calibration is available as the TC-CAL option at $125. The sensing element reads in both directions regardless; the certificate is what covers one or the other. On a genuinely through-zero application it is worth a short conversation about which direction your results have to be defensible in, because that depends on your reporting requirements rather than on the hardware.
Questions From The Field
My readings show a flat spot as the load passes through zero. Is the sensor at fault?
Almost certainly not — what you are seeing is usually slack in the load path, and the sensor is reporting it accurately. As the force reverses, any clearance in the assembly has to be taken up before load transfers again: thread slack, a pin in an oversized hole, a joint that only bears in one direction, a fitting that was tight rather than positively secured. During that take-up very little is transmitted, and the curve flattens. Work through the assembly joint by joint looking for anything that can move, and eliminate the free play rather than adjusting the sensor.
Tension readings and compression readings do not seem consistent with each other.
Check the zero first. On a bidirectional measurement an offset at zero does not shift everything equally — it makes one direction read high while the other reads low, which presents exactly as the inconsistency you describe. Establish the unloaded reading with the fixture genuinely at rest and at working temperature, and compare it against the 1.0% zero balance figure. If the zero is sound, the next candidate is the fixture: an assembly that is stiff in compression and compliant in tension, or vice versa, will not behave symmetrically no matter how good the sensor is.
Can I fit an LPO into a machine that was not designed to take a load cell?
That is much of the reason the series exists, and the low profile is what makes retrofits feasible. Two things to check before committing. Measure the available height accurately, allowing for any adapter or spacer, and take the figure from the specific model since it grows with capacity. Then consider what happens to the load path once the sensor is in it — the sensor has to carry the whole force, its mounting faces need to be flat and properly supported, and anything that introduces a bending moment will be reported as force. Send us the arrangement and we will look at it with you.
Do the two bridges on a dual bridge cell read exactly the same?
They read the same force, but they are separate measurement circuits with their own gauges and their own calibration, so expect a small difference between them rather than identical numbers. That is normal and it is also the point — you establish the usual level of agreement when the system is commissioned, and thereafter a departure from that baseline is your fault indication. Record the relationship between the channels while everything is known to be healthy; without a baseline, a comparison tells you very little.
Can I use the second bridge for something other than redundancy?
Yes, and it is a common and sensible use. Two independent outputs let you feed two systems that should not share a connection — a machine controller and a separate data acquisition setup, or a process control loop and an independent monitoring channel. Each gets a clean signal, and neither can disturb the other by being reconfigured, disconnected or powered down. If that is your intention rather than fault tolerance, tell us, because it affects how the channels should be conditioned.
Do I need separate instrumentation for each bridge?
Yes. Each bridge is an independent sensor circuit and needs its own excitation and its own conditioning — a dual bridge cell requires two channels, a triple bridge three. It is worth thinking about how independent those channels really are: if both feed a single instrument on a single supply, a failure of that instrument takes out both readings and the redundancy you paid for has quietly disappeared. Where fault tolerance is the objective, keep the chains separate all the way through.
Should I order the twist-lock connector or the standard cable?
Option-PTB at $75 gives you a twist-lock connector, and the question to ask is whether the sensor will ever come out. Fixtures that are reconfigured between jobs, sensors that go away for recalibration, and installations where the cable route would have to be dismantled all argue for it. A permanent installation on accessible equipment does not need it, and an integral cable has fewer parts to work loose. If you do specify the connector, add the mating assembly to the same order — it is a separate part and there is nothing to plug in without it.
How should I mount an LPO so the readings are trustworthy?
Flat, supported and square. The mounting faces need to be genuinely flat and stiff enough not to dish under load, because a surface that deforms puts bending into the sensing element and the gauges report bending as force. Fasteners should be evenly tightened so the body is not pre-stressed by its own mounting. On a bidirectional installation, everything in the load path should be positively secured rather than merely tight, since load reversal will find any joint that can move. Then zero the system once it is assembled and at working temperature, not before.