Hart's W-frame is a six-bar planar linkage that converts rotary input into exact straight-line motion of a tracing point, without sliding contact. Built correctly, it can hold straightness to within a few microns over a stroke of 30-60 mm, limited mainly by pivot clearance. Harry Hart published it in 1877 as a simpler alternative to the Peaucellier-Lipkin cell, and it still appears in flexure-free comparators, optical alignment rigs, and teaching kinematics models at institutions like the Cornell KMODDL collection.
Hart's W-frame Interactive Calculator
Vary bar length, matching tolerance, pivot clearance, and linkage amplification to see the allowable bar mismatch and predicted straightness wander.
Equation Used
This calculator follows the article examples: the allowable link mismatch is the nominal long-bar length times the matching tolerance, and input pivot clearance is multiplied by the linkage ratio to estimate lateral straightness wander at the tracing point.
- Hart geometry is otherwise exact and link ratios are correctly set.
- Straightness error is dominated by input pivot clearance amplified by the linkage ratio.
- Bar matching tolerance is applied to the nominal long-bar length.
Operating Principle of the Hart's W-frame
Hart's W-frame uses six rigid links pinned together in a configuration that looks like a flattened W when laid out on the bench. Two long links cross each other and two shorter links complete the frame, with a fixed pivot at one end and the tracing point at a specific ratio along one of the bars. When you rotate the input crank, the tracing point traces a mathematically exact straight line — not an approximation, not a near-line like a Watt linkage, but a true Euclidean straight path. That property comes from the antiparallelogram geometry buried inside the W: opposite bars stay equal in length throughout the motion, and the cross-ratio of the pivot points stays constant.
The geometry only works if the link ratios are exact. The two long crossing bars must be identical to within a fraction of a percent, and the two shorter bars likewise. If you build it with one long bar 0.5% longer than its partner, the tracing point no longer follows a straight line — it traces a shallow arc with a sagitta you can measure with a dial indicator. This is the most common failure mode in classroom builds. The second is pivot slop: each pin joint adds clearance, and clearance at the input crank pivot multiplies into lateral wander at the tracing point by roughly the link-length ratio. A 0.05 mm clearance at the input pin can show up as 0.2 mm of straightness error at the output, which kills any precision application.
Why build it this way at all? Because exact straight-line linkages let you guide a probe, mirror, or tool in a true line without rails, slides, or flexures. No sliding contact means no stick-slip, no lubrication, no wear-induced drift. For short strokes in precision instruments, that's worth the extra link count compared to a slider-crank.
Key Components
- Long crossing bars (2): These are the two equal-length links that cross each other in the middle of the W. They must match in length to within ±0.05% for a precision build — at 200 mm nominal that's a 0.1 mm tolerance window. Any mismatch shows directly as straightness error at the tracing point.
- Short outer bars (2): Two shorter equal-length links that close the antiparallelogram. They sit at the outer ends of the W and must also match each other to the same tolerance class as the long bars.
- Fixed ground pivot: Anchors the linkage to the frame. Pivot clearance here is critical — use a precision dowel pin in a reamed bore, target H7/g6 fit (around 6 µm diametral clearance on a 6 mm pin) to keep tracing-point wander under 10 µm.
- Input crank: Drives the linkage through a small angular range. You don't need full rotation — a typical Hart W-frame uses 30-60° of crank input to sweep the tracing point through its full straight-line stroke.
- Tracing point: Located at a specific ratio along one of the long bars (typically the same ratio that defines the link-length proportions). This is where you mount the probe, mirror, or tool. Position tolerance here is around ±0.1 mm for a 50 mm stroke at 10 µm straightness.
- Pin joints: Five revolute pairs total. Each one needs low clearance and low friction. Bushed bronze in a reamed steel bore works for teaching models; jewelled or ball-race pivots are needed for sub-micron work.
Real-World Applications of the Hart's W-frame
Hart's W-frame shows up wherever a designer needs exact straight-line motion over a short stroke without the cost or compliance of a slide rail or flexure. The stroke is typically 20-80 mm and the speeds are slow — this is a precision-positioning mechanism, not a high-throughput one. You see it most in optics, metrology, and educational kinematics, and occasionally in medical or horological instruments where contamination from sliding lubricant is unacceptable.
- Education and kinematics teaching: The Cornell KMODDL (Kinematic Models for Design Digital Library) collection holds physical Hart W-frame models used to teach exact straight-line motion alongside Peaucellier and Hart inversor cells.
- Precision optics: Tabletop alignment rigs at university optics labs use W-frame linkages to translate a folding mirror in a true line over 40 mm without introducing yaw, where a dovetail slide would add stick-slip.
- Metrology: Comparator probe carriers in workshop-grade flatness checkers use Hart W-frame guidance for the probe tip on strokes up to 60 mm, eliminating the lateral runout you get from a bushed slide.
- Horology and instrument making: Custom watchmaker's depthing tools and pivot-polishing fixtures occasionally use W-frame guidance where a clean, lubricant-free linear motion is needed near a movement under repair.
- Historical demonstration: The Science Museum in London and the Deutsches Museum in Munich both display straight-line linkage models in the kinematics section, with Hart's work cited alongside Peaucellier-Lipkin as a foundational 19th-century solution.
- Research instrumentation: Bench-built piezo-driven sample stages for low-load microscopy occasionally use a W-frame as the coarse-positioning stage above the piezo, giving 50 mm of clean travel before fine adjustment.
The Formula Behind the Hart's W-frame
The straightness error at the tracing point depends almost entirely on link-length matching and pivot clearance. The formula below estimates the lateral deviation Δy of the tracing point from the ideal straight line when the long bars differ in length by ΔL. At the low end of a typical build — ΔL/L around 0.01% on a tight precision build — straightness error stays under a micron. At the nominal hobby-grade build with ΔL/L around 0.1%, you're looking at tens of microns of error, which is fine for demonstration but useless for metrology. Push to ΔL/L of 1% on a sloppy laser-cut acrylic build and the tracing point traces a visible arc. The sweet spot for a useful workshop comparator sits around ΔL/L = 0.05% with reamed bronze bushings.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Δy | Lateral straightness error at the tracing point | m | in |
| ΔL | Length mismatch between the two long crossing bars | m | in |
| L | Nominal length of the long bars | m | in |
| s | Stroke of the tracing point | m | in |
| k | Geometry factor (typically 0.4-0.6 for a standard W-frame layout) | dimensionless | dimensionless |
Worked Example: Hart's W-frame in a tabletop wafer-edge inspection stage
A semiconductor metrology startup in Eindhoven is prototyping a tabletop wafer-edge inspection stage. They want a Hart W-frame to translate a 5 mW fibre-coupled illuminator across a 50 mm stroke at the wafer edge, holding straightness to under 10 µm so the spot stays focused on the bevel. Long bars are 200 mm, short bars are 80 mm, and the geometry factor for their layout works out to k = 0.5. They need to know what link-matching tolerance to specify when machining the bars.
Given
- L = 200 mm
- s = 50 mm
- k = 0.5 dimensionless
- Δytarget = 10 µm
Solution
Step 1 — rearrange the formula to solve for the required ΔL given the target straightness:
Step 2 — plug in nominal values for the 10 µm straightness target:
So the two long bars must match within 80 µm, which is ΔL/L = 0.04%. This is achievable on a manual mill with careful indicator work but tight enough that you should grind the bars as a matched pair from the same blank.
Step 3 — at the low end of build quality, ΔL/L = 0.5% (a quick laser-cut acrylic prototype with no matching pass):
That's 12× over budget. The illuminator spot would walk visibly off the wafer bevel — you'd see the illumination intensity flicker as the spot drifts in and out of focus. Useless for inspection.
Step 4 — at the high end of build quality, ΔL/L = 0.005% (jig-bored bars, lapped pivot bores, jewelled pin joints):
Now you're at sub-micron straightness, comparable to a low-end air bearing slide but with no air supply, no flexure compliance, and no sliding contact. This is the regime where Hart's W-frame earns its place in a precision instrument.
Result
The Eindhoven team needs to hold the long-bar length match to ±80 µm to meet their 10 µm straightness target — a tolerance of ΔL/L = 0. 04%, which means matched-pair grinding rather than independent machining. At the low-quality end (0.5% mismatch) straightness blows out to 125 µm and the illuminator spot wanders off the wafer bevel; at the high-quality end (0.005% mismatch) you get sub-micron straightness rivalling an air bearing. If the measured straightness comes in worse than predicted, look at three things first: short-bar length mismatch (often forgotten because builders focus only on the long bars), out-of-plane twist in the link plates from inadequate stiffness in the link cross-section, or fixed-pivot baseplate flex if the ground bracket is thinner than about 8 mm steel for this stroke.
Choosing the Hart's W-frame: Pros and Cons
Hart's W-frame is one of several exact and approximate straight-line linkages. The choice between them comes down to how much straightness you need, how much stroke, and how many parts you're willing to fit and tune. Here's how the W-frame stacks up against the two most common alternatives a precision designer actually considers.
| Property | Hart's W-frame | Peaucellier-Lipkin cell | Watt linkage |
|---|---|---|---|
| Straightness type | Exact (mathematical) | Exact (mathematical) | Approximate (3rd-order) |
| Achievable straightness over 50 mm stroke | 1-10 µm with precision build | 1-10 µm with precision build | 50-200 µm typical |
| Number of links | 6 | 8 | 4 |
| Number of pin joints | 5 | 7 | 4 |
| Build complexity | Moderate | High | Low |
| Sensitivity to link-length tolerance | High — 0.05% mismatch shows | Very high — multiple matched pairs | Lower — geometry self-corrects somewhat |
| Typical stroke range | 20-80 mm | 20-100 mm | 50-300 mm |
| Best application fit | Compact precision instruments | Demonstration, optics | Engine linkages, suspension |
Frequently Asked Questions About Hart's W-frame
Almost always the short outer bars. Builders fixate on matching the long crossing bars and ignore the short ones, but the antiparallelogram condition requires both pairs to be equal. If your short bars differ by 0.2% and your long bars match to 0.02%, the short-bar mismatch dominates the error budget.
Quick check: swap the two short bars left-for-right and run the linkage again. If the arc reverses direction, the short-bar mismatch is your problem. If the arc stays the same, look at out-of-plane twist in the link plates instead.
For a typical W-frame with long bars at 4× the short-bar length, you get the full straight-line stroke from roughly 40-50° of input crank rotation. Push beyond about 60° and the tracing point starts to deviate from the straight line because you're leaving the geometry's valid range.
Design rule: pick your stroke first, then size the long bars so the required crank angle stays under 50°. Going to longer bars buys you angular margin at the cost of overall package size.
Depends on stroke and stiffness needs. Flexures beat the W-frame on repeatability and have zero backlash, but they store energy — you fight a spring force that grows with displacement, and at 30 mm stroke you need a long flexure stack that takes up real estate. The W-frame is essentially neutral in force throughout its stroke and packages compactly.
Rule of thumb: under 5 mm stroke, flexure wins on every metric. 5-20 mm is a real toss-up that depends on load. Above 20 mm the W-frame's compactness and constant-force behaviour usually win unless you need nanometre repeatability.
That's pivot clearance showing up as dead-band motion. Each pin joint has some diametral clearance, and gravity or applied load lets the bars settle to one side of the clearance. Tap the linkage and watch the tracing point jump — that jump is the sum of all the joint clearances projected onto the output.
Fix: ream all bores together using a single reamer pass, fit dowel pins for an H7/g6 fit (around 6 µm clearance on a 6 mm pin), and pre-load the linkage with a light spring against the input crank to take up the clearance in one consistent direction.
Yes, and for many precision applications it's the cleaner solution. Drive the input link directly with a small Linear Actuator pushing tangentially on the crank arm — you trade rotary input for direct linear input, and you avoid backlash from a rotary-to-linear conversion stage.
The kinematics don't care how you supply motion to the input link, only that the link rotates through the required angular range. A 25 mm stroke actuator pushing on a 40 mm crank arm gives you about 36° of input rotation, which is enough for a typical W-frame stroke. Just make sure the actuator mounting pivot is itself low-clearance, or you import slop right into the input.
This is a quasi-static mechanism. You can run it up to a few hertz of cycle rate before inertial loads at the cross-bar centres start to deflect the bars and degrade straightness. For a 200 mm long-bar build in steel, expect clean behaviour up to around 2 Hz cycling and visible degradation above 5 Hz.
If you need higher speed, stiffen the bars (move from 6 mm flat stock to 10 mm with lightening pockets), shorten the linkage, or accept that you're using the wrong mechanism for the duty cycle.
That's normal and expected. The mathematical straight-line condition is exact only when the geometry is correct, but real builds have small errors that grow with displacement from the design centre. The error term scales roughly with the square of distance from the stroke midpoint, so the first 60% of the stroke around centre stays clean and the last 20% at each end shows most of the deviation.
Design around it: oversize your stroke by 20-30% and use only the centre band, or accept the end-of-stroke degradation if your application only needs precision in the middle range.
References & Further Reading
- Wikipedia contributors. Straight-line mechanism. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
