Hart's Exact Straight-line Inversor is a six-bar planar linkage that converts rotary motion at one pivot into mathematically exact straight-line motion at a tracing point. Unlike the eight-bar Peaucellier-Lipkin cell, Hart's design achieves the same exact straight line with only six links arranged as an antiparallelogram with two attached connecting bars. Engineers use it where an approximate straight-line linkage like Watt's or Chebyshev's introduces unacceptable lateral deviation. Real builds reach better than 5 µm straightness over a 50 mm stroke, which is why it shows up in precision metrology fixtures and optical alignment rigs.
Hart's Exact Straight-line Inversor Interactive Calculator
Vary bar length, pair mismatch, and stroke to estimate straightness error and see the Hart inversor geometry respond.
Equation Used
The article states that a 100 mm Hart linkage needs equal long-bar and short-bar pairs matched to about 0.02 mm, and that a 0.1 mm pair drift can create about 30-50 um lateral deviation over a 40 mm stroke. This calculator scales that worked-example error range with mismatch and stroke.
- Uses the article worked-example relationship near 100 mm bar length, 0.1 mm mismatch, and 40 mm stroke.
- Error is scaled linearly with pair mismatch and stroke for teaching and early tolerance estimation.
- Applies to matched long-bar or short-bar pair error, assuming other geometry ratios are correct.
Inside the Hart's Exact Straight-line Inversor
The mechanism starts from a kinematic fact published by Harry Hart in 1875: if you take a crossed four-bar antiparallelogram with bars a-b-a-b and pick three collinear points on it — one on each long bar and one on the coupler — those three points stay collinear as the linkage moves, and one of them traces an exact straight line when the other two are constrained to circular arcs. That's the whole trick. You fix one of those collinear points to ground through a rotating crank, you fix the second to ground through another crank of the right length, and the third point — the tracing point — has no choice but to draw a straight line perpendicular to the line joining the two fixed pivots.
The geometry is unforgiving. The antiparallelogram needs the two long bars equal to within roughly 0.02 mm on a 100 mm linkage, and the two short bars must match each other to the same tolerance. If you let those pairs drift apart by even 0.1 mm, the tracing point starts drawing a shallow arc instead of a line, and you'll see lateral deviation of 30-50 µm over a 40 mm stroke — enough to ruin any application that chose this linkage over a Watt or Chebyshev approximate straight-line mechanism in the first place. The collinearity ratio along the bars also has to be exact: if the tracing point sits at fraction k along one bar, the other two collinear points must sit at the same fraction k on their respective bars. Get that ratio wrong and the straight line becomes a lemniscate-like curve.
Failures in real builds almost always come from three places. Pivot clearance — anything above an H7/g6 fit on the pin joints lets the antiparallelogram cross over into a parallelogram configuration at dead centre, and the linkage locks. Bar flex — long thin bars under side load bow enough to spoil the geometry, so you want bars stiff enough that mid-span deflection stays below 5 µm at peak load. And out-of-plane wobble — if the six bars don't sit in parallel planes within 0.05 mm, you introduce a sinusoidal wobble in the tracing point that no amount of bar-length trimming will fix.
Key Components
- Antiparallelogram long bars (×2): The two equal-length crossed bars that form the heart of the inversor. Length match must hold within 0.02 mm on a 100 mm bar. These cross over each other in operation, which is why you need bars in offset parallel planes rather than coplanar.
- Antiparallelogram short bars (×2): The two equal-length bars that close the antiparallelogram between the long-bar ends. Their length sets the aspect ratio of the linkage and therefore the stroke length. Match them to each other within 0.02 mm.
- Fixed-pivot crank A: Grounds one of the three collinear points and constrains it to a circular arc of fixed radius. The crank length and ground-pivot position are what force the straight-line solution rather than a generic curve.
- Fixed-pivot crank B: Grounds the second collinear point. Crank A and crank B together define the line that the tracing point traces — the line is perpendicular to the segment joining the two ground pivots and passes through their midpoint.
- Tracing point: The third collinear point on the linkage. Sits at the same fractional position along its bar as the other two collinear points sit along theirs. Holds straightness to within 5 µm over 50 mm in a well-built rig.
- Pin joints (×6): Six revolute joints, ideally H7/g6 fit or better. Joint slop is the single biggest enemy of straightness — 25 µm of radial clearance per joint typically gives 60-80 µm of straightness error at the tracing point.
Real-World Applications of the Hart's Exact Straight-line Inversor
Hart's Inversor lives in places where you need true rotary-to-linear conversion without prismatic guides — no rails, no bushings, no sliding surfaces to wear or contaminate. That makes it useful in vacuum chambers, cryogenic rigs, contamination-sensitive optics, and teaching collections where the exactness of the line is itself the point. It rarely shows up in heavy machinery because the load capacity is limited by the slender bar geometry, but for low-force precision tasks it competes well against approximate straight-line linkages and even against linear bearings.
- Precision metrology: Comparator straightness reference on the National Physical Laboratory's heritage length-bar comparator, used as a teaching artefact to demonstrate guideway-free straight motion.
- Vacuum equipment: Sample-transfer arm in Oxford Instruments' low-temperature STM rigs where sliding bushings would shed particulates and trap gas.
- Educational kinematics: Reuleaux Kinematic Models collection at Cornell University, where the Hart Inversor sits alongside the Peaucellier cell as the classic six-bar straight-line teaching mechanism.
- Optical alignment: Linear-motion stages on autocollimator calibration rigs where prismatic-guide stiction would corrupt sub-arcsecond angular readings.
- Watchmaking and horology: Lever-escapement test fixtures at independent ateliers using Hart geometry to advance a probe in a true straight line without lubricant migration.
- MEMS prototyping: Macro-scale demonstrators of compliant straight-line flexures, where teams scale Hart geometry up to 200 mm to validate kinematics before etching a 2 mm silicon version.
The Formula Behind the Hart's Exact Straight-line Inversor
The useful number to compute is the straight-line stroke length S that the tracing point sweeps for a given input crank rotation. At the low end of the typical operating range — short bars and small crank angles — you get a tiny stroke but vanishingly small straightness error. At the high end, where the antiparallelogram swings close to its singular fold-flat position, stroke length grows but bar interference and joint loading climb fast. The sweet spot for most builds sits at around 60-70% of the geometric maximum stroke, where you get usable travel without the linkage approaching the configurations where the antiparallelogram threatens to invert into a parallelogram.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| S | Stroke length traced by the tracing point along the straight line | m | in |
| a | Length of the long antiparallelogram bars (matched pair) | m | in |
| b | Length of the short antiparallelogram bars (matched pair) | m | in |
| k | Collinearity fraction — position of the tracing point along its bar (0 < k < 1) | dimensionless | dimensionless |
| d | Distance between the two fixed ground pivots | m | in |
Worked Example: Hart's Exact Straight-line Inversor in a coffee-roasting equipment lab in trieste
A coffee-roasting equipment lab in Trieste is building a bench rig to drive a thermocouple probe in a true straight line into a sample chamber on a Probat Probatino test roaster. The probe must enter and retract along a 60 mm path with straightness better than 10 µm so the bead sits at the same isothermal layer on every insertion. The team picks Hart's Inversor over a linear bearing because the chamber atmosphere reaches 230 °C and they want no sliding surfaces. They size the long bars at a = 120 mm, the short bars at b = 80 mm, the collinearity fraction k = 0.5, and the ground-pivot spacing at d = 80 mm.
Given
- a = 120 mm
- b = 80 mm
- k = 0.5 dimensionless
- d = 80 mm
Solution
Step 1 — compute the geometric core (a2 − b2) at the nominal bar lengths:
Step 2 — apply the stroke formula at nominal geometry to get the achievable straight-line travel:
That is the geometric maximum. The team will only use about 60% of it — call it 60 mm of usable stroke — to keep the linkage well away from the fold-flat singularity where the antiparallelogram threatens to cross into a parallelogram.
Step 3 — at the low end of the typical range, shrink the long bars to a = 100 mm (a common downsizing for benchtop kits):
45 mm of geometric stroke means roughly 27 mm usable — barely enough to clear the chamber wall. You feel this in the build as a short, stiff motion that runs out of travel before the probe bead reaches the target layer.
Step 4 — at the high end, push the long bars to a = 140 mm:
Theoretically excellent, but in practice bars at a = 140 mm and b = 80 mm sit close to the configuration where the long bars cross at a shallow angle, and side loads on the tracing point bow the long bars enough to push straightness past 20 µm. Above roughly a/b = 1.7 you will see the straightness budget collapse on a workshop-grade build.
Result
Nominal geometric stroke is 100 mm, of which the team will use 60 mm of clean straight travel for the probe insertion. At the low end (a = 100 mm) the geometry yields only 45 mm of stroke — a stiff, short motion that runs out of travel before the bead reaches the target isothermal layer — while at the high end (a = 140 mm) the formula promises 165 mm but bar bowing and near-singular geometry push real straightness above 20 µm, blowing the 10 µm budget. If you measure the probe drawing a shallow arc instead of a line, suspect three things in this order: long-bar length mismatch above 0.05 mm (will produce 30-50 µm of arc curvature on this scale), out-of-plane bar wobble from non-coplanar mounting brackets, or a collinearity fraction k that drifted from 0.5 because the tracing-point hole was drilled off-centre on the coupler.
When to Use a Hart's Exact Straight-line Inversor and When Not To
Hart's Inversor competes against the Peaucellier-Lipkin cell on one side and against approximate straight-line linkages — Watt, Chebyshev, Roberts — on the other. The choice comes down to whether you need exact straightness, how much stroke you need, and how many links you can tolerate. Below is how those options stack up on the dimensions that actually matter on a build sheet.
| Property | Hart's Inversor | Peaucellier-Lipkin Cell | Watt's Linkage (approximate) |
|---|---|---|---|
| Straightness accuracy | Mathematically exact (≤5 µm in a 50 mm build) | Mathematically exact (≤5 µm in a 50 mm build) | Approximate, 50-200 µm deviation over similar stroke |
| Number of links | 6 bars | 8 bars | 3 bars |
| Number of pin joints | 6 revolute joints | 8 revolute joints (including 2 rhombus crossings) | 4 revolute joints |
| Typical usable stroke / longest bar ratio | 0.4-0.6 | 0.3-0.5 | 0.25-0.35 |
| Build complexity | Moderate — antiparallelogram needs offset planes | High — rhombus plus two equal links plus two equal cranks | Low — three bars, no special geometry |
| Sensitivity to bar-length error | High — 0.02 mm match required on long-bar pair | High — rhombus must hold 0.02 mm equality on all four sides | Low — modest tolerance, the curve is approximate anyway |
| Load capacity (typical workshop build) | Low — 5-20 N at tracing point | Low — 5-20 N at tracing point | Medium — 50-200 N at coupler point |
| Best application fit | Vacuum, optical, contamination-sensitive precision motion | Teaching demonstrators, historical replicas | Engine crossheads, suspension geometry, robust low-precision linear motion |
Frequently Asked Questions About Hart's Exact Straight-line Inversor
You are hitting the antiparallelogram's singular configuration — the geometry where the two long bars momentarily align with the two short bars and the linkage can flip from antiparallelogram into parallelogram. At that instant the kinematics are indeterminate and any joint slop above about 25 µm radial will let the wrong branch take over.
The fix is to bias the linkage. Add a light torsion spring at one of the antiparallelogram pivots, sized to maintain about 0.5 N·m of preload toward the antiparallelogram branch. This biases the geometry so it cannot cross into the parallelogram state at the singular angle. Alternatively, restrict the input crank's swing range so it never reaches the singular configuration in the first place — usually keeping it within ±50° of the symmetric position is enough.
Count the joints, not the bars. Peaucellier has 8 pin joints versus Hart's 6, and joint clearance is the dominant straightness error in any real build. Six joints at H7/g6 fit will out-perform eight joints at the same fit by roughly 25-30% on tracing-point straightness, simply because errors stack.
That said, Peaucellier is easier to lay out symmetrically and easier to teach, which is why it dominates university kinematics collections. Hart wins when you actually need to build the thing for a working precision rig — fewer joints, fewer parts to match-grind, and a more compact footprint for the same stroke.
This is almost always asymmetric joint loading combined with bar bow. On the forward stroke the tracing-point load pulls the long bars one way, on the return stroke it pushes them the other. If the bars are not stiff enough — say, less than 5 µm mid-span deflection at peak load — the geometry shifts slightly between the two directions and you get a hysteresis loop that looks like a straight line going out and a curve coming back.
Diagnose by reversing the load direction at the tracing point and re-measuring. If the curve flips to the opposite stroke, it's bar bow. The fix is stiffer bars, usually a doubling of the section modulus by going from 6 mm flat stock to 8 mm. Alternatively, run the linkage unloaded and apply the load through a counterweight that always pulls in the same direction.
k = 0.5 is the default and it is almost always correct. Putting the tracing point at the midpoint of its bar gives the largest symmetric stroke and the most forgiving error budget — bar-length mismatches average out symmetrically rather than amplifying.
You only deviate from k = 0.5 when you need to bias the stroke direction or extend reach in one direction at the cost of the other. At k = 0.3 or k = 0.7 you trade away symmetric error cancellation, which means a 0.05 mm bar-length error that would have produced 10 µm straightness deviation at k = 0.5 now produces 25-30 µm. On a precision rig, stick with 0.5 and machine the bar holes carefully.
The formula assumes ideal pin joints, perfectly rigid bars, perfectly equal long-bar pair, perfectly equal short-bar pair, and bars in coplanar (or correctly offset) planes. In a workshop build, joint clearance alone typically eats 15-25 µm of the budget. Bar-length mismatch above 0.03 mm adds another 10-20 µm. Out-of-plane wobble from sloppy bracket mounting adds 10-30 µm of sinusoidal error that looks like straightness deviation on a single-axis indicator.
To narrow it down, mount a dial indicator perpendicular to the expected line and run the input crank slowly through one full sweep. If the error trace is smooth and parabolic, it's bar-length mismatch. If it's sinusoidal with the period of the input crank, it's out-of-plane wobble. If it's noisy and discontinuous, it's joint clearance.
Not as a pin-jointed linkage. At sub-millimetre scale, pin clearances dominate the geometry and friction at the joints overwhelms any input torque you can deliver. What MEMS designers actually do is replace the six revolute joints with compliant flexure hinges and etch the whole thing as a single monolithic silicon part.
The kinematic equivalence is only approximate at that point — flexures introduce parasitic motion of a few percent — but the straightness can still hit sub-100 nm over a 100 µm stroke, which is good enough for fibre alignment and AFM probe positioning. The macro-scale Hart layout serves as the design template that engineers tune in finite-element software before committing to a mask set.
References & Further Reading
- Wikipedia contributors. Hart's inversor. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
