A Rhombic Drive is a four-bar linkage that converts the reciprocating motion of two coaxial pistons into rotation through a pair of counter-rotating crankshafts joined by a diamond-shaped yoke. Philips Research Laboratories in Eindhoven patented the arrangement in 1953 under engineer Roelf Meijer to silence their Stirling engine programme. The two crankshafts spin in opposite directions, cancelling first and second-order shaking forces at the crankcase. The result is a Stirling engine or cryocooler that runs without external balancers — the configuration still powers spaceborne IR sensor coolers today.
Rhombic Drive Interactive Calculator
Vary piston mass, speed, and crank timing error to see the slider-crank shaking force and the rhombic drive residual imbalance.
Equation Used
F1 estimates the peak first-order shaking force a simple slider-crank would transmit. The calculator uses the 50.7 mm crank throw implied by the article's 200 g, 1500 RPM, 250 N worked example, then estimates rhombic residual force from crank timing mismatch delta.
FIRGELLI Automations - Interactive Mechanism Calculators.
- Reference crank throw is 50.7 mm, inferred from the worked example result.
- Residual force is caused by phase mismatch between equal counter-rotating crank forces.
- Only first-order reciprocating inertia is estimated.
The Rhombic Drive in Action
The Rhombic Drive, also called the Rhombic drive (Stirling) in thermodynamics literature, takes two pistons sharing a single cylinder bore — typically a power piston and a displacer — and ties each one through a connecting rod to a diamond-shaped yoke. The yoke's four corners pin to two parallel crankshafts that turn in opposite directions, geared together so their phase never drifts. As the pistons stroke up and down, the yoke flexes between a tall narrow rhombus and a short wide one, and the equal-but-opposite crank throws cancel each other's lateral inertia. You get pure reciprocating motion at the pistons and pure rotation at the output shaft, with no net side-loads on the cylinder wall.
Why bother with this much linkage when a simple slider-crank would do the same job? Because a slider-crank shakes. In a Stirling engine running at 1,500 RPM with a 200 g power piston, the first-order shaking force is around 250 N peak — enough to walk the engine across a workbench. The Rhombic Drive cancels both first and second-order forces by symmetry, which is why Philips picked it for their domestic generators in the 1950s and why aerospace cryocoolers still use it today. The two crankshafts must be timed to within ±0.5° of each other; if the synchronising gears wear or the keys slip, you'll feel an immediate buzz through the casing and see the displacer phase angle drift away from its design point of typically 90°.
Tolerances on the yoke pivots are tight. Bushing clearance above 0.05 mm at any of the four pin joints lets the yoke skew under load, which puts the displacer rod off-axis and scrapes the rod seal. The classic failure mode is a worn yoke pin showing up first as a knock at top dead centre, then as helium leakage past the displacer seal a few hundred hours later. The seal didn't fail — the linkage did.
Key Components
- Power Piston: The lower of the two coaxial pistons, taking gas pressure and feeding work into the linkage. Typical bore 70-90 mm in a 1 kW engine, with a piston-to-cylinder clearance of 15-25 µm to limit blow-by without scuffing.
- Displacer Piston: The upper piston, shuttling working gas between hot and cold spaces without sealing pressure. Its rod passes through the power piston on the cylinder centreline, so any yoke skew translates directly into rod-seal wear.
- Rhombic Yoke: The diamond-shaped four-bar plate that links both connecting rods to both crankshafts. Must stay rigid under peak gas load — a yoke that flexes 0.1 mm at the centre pivot will throw the displacer phase angle by 1-2°.
- Twin Counter-Rotating Crankshafts: Two parallel cranks geared 1:1 in opposite directions. Phasing tolerance ±0.5°; sloppier than that and the inertia cancellation breaks down, reintroducing shake at the fundamental frequency.
- Synchronising Gears: Spur or helical pair locking the two crankshafts together. Backlash above 0.03 mm shows up as an audible knock at every stroke reversal and as gradual drift of piston phase under load.
- Connecting Rods: Two rods, one per piston, each pinned to opposite vertices of the yoke. Length matched to within 0.02 mm between the pair so the piston motions stay symmetrical about TDC.
Real-World Applications of the Rhombic Drive
The Rhombic Drive lives almost entirely inside Stirling engines and Stirling-cycle cryocoolers, where vibration and side-loading on the displacer rod are the make-or-break engineering problems. You'll find it wherever a builder needs reciprocating gas-machine output without an external balancer, and where the displacer rod must run on the cylinder axis to keep dynamic seals alive.
- Aerospace IR sensor cooling: Lockheed Martin and Northrop Grumman have flown Rhombic Drive Stirling cryocoolers on missile-warning satellites to chill HgCdTe focal planes to 60-80 K. Vibration export below 0.05 N is the spec the linkage is chosen for.
- Domestic power generation: Philips MP1002CA generator set, built in Eindhoven from 1951, used a Rhombic Drive in a 200 W Stirling engine intended for radio power in remote regions. Roelf Meijer's prototype proved the concept here.
- Submarine air-independent propulsion: Kockums (now Saab) V4-275R Stirling units in the Swedish Gotland and Japanese Sōryū-class submarines use Rhombic-Drive-derived linkages to keep acoustic signature down during silent running.
- Laboratory cryogenics: Stirling Cryogenics SPC-1 helium liquefiers run a Rhombic Drive (Stirling) configuration at 1,500 RPM to deliver 1 W of cooling at 4.2 K for superconducting magnet labs.
- Solar dish concentrators: United Stirling 4-95 engine used in early SES Solar Two trials at Sandia drove a 25 kW alternator from a parabolic dish, with the Rhombic Drive carrying the imbalance loads from the four cylinders.
- Medical and analytical instruments: Liquid-nitrogen-free MRI pre-coolers and gas chromatograph detector chillers use compact 5-50 W Rhombic Drive cryocoolers because the bench instrument can't tolerate shaker output.
The Formula Behind the Rhombic Drive
The single number that defines a Rhombic Drive's geometry is the displacer-to-power-piston phase angle θ, set by the crank throw r, the yoke half-width e, and the connecting-rod length l. Get the geometry right and you hit the textbook 90° phase that maximises Stirling cycle work. At the low end of the typical e/r ratio (around 1.0) the phase narrows toward 75° and indicated power drops 15-20%. At the high end (around 1.6) the phase opens past 100° and you start losing volumetric efficiency. The sweet spot for most aerospace coolers sits at e/r ≈ 1.3, l/r ≈ 4.0.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| θ | Phase angle between displacer and power piston motion | degrees | degrees |
| r | Crank throw (radius from crankshaft centre to crankpin) | mm | in |
| e | Yoke half-width (offset from cylinder centreline to crankshaft centre) | mm | in |
| l | Connecting rod length, pin-to-pin | mm | in |
Worked Example: Rhombic Drive in a benchtop Stirling cryocooler
A cryogenics startup in Boulder is sizing the Rhombic Drive for a 10 W-at-77 K benchtop Stirling cryocooler aimed at university superconductor labs. The crank throw is fixed at 12 mm by the desired piston stroke, and the connecting rod length is 48 mm. The team wants to know how the displacer-to-power-piston phase angle varies as they sweep the yoke half-width e across the practical range from 12 mm to 19 mm, and where to land the design point.
Given
- r = 12 mm
- l = 48 mm
- e (sweep) = 12 to 19 mm
Solution
Step 1 — at the nominal design point, e = 15.6 mm (e/r = 1.3), compute the two arctan terms:
That lands almost exactly on the textbook 90° Schmidt-analysis optimum — the design will deliver close to peak indicated power for a given swept volume.
Step 2 — at the low end of the practical range, e = 12 mm (e/r = 1.0):
76° phase costs roughly 15-18% of the cycle work compared to the 90° optimum, and the cooler that should hit 10 W at 77 K will measure closer to 8.2 W on the bench. Not catastrophic, but you've left performance on the table.
Step 3 — at the high end, e = 19 mm (e/r = 1.58):
105° drives the displacer too far ahead of the power piston. Volumetric efficiency falls, the regenerator sees flow reversal at the wrong moment, and net cooling drops to about 7.5 W. The tall-narrow yoke also pushes side loads on the synchronising gears noticeably higher.
Result
The nominal design at e = 15. 6 mm gives a displacer phase angle of 90.4°, sitting right on the cycle-work optimum. Across the swept range, phase moves from 76° at the low end to 105° at the high end — a 29° spread that translates to a 25% spread in cooling capacity, with the sweet spot a narrow band around e/r = 1.25-1.35. If your bench prototype measures 90° geometric phase but only delivers 7-8 W instead of the predicted 10 W, the most likely culprits are: (1) synchronising gear backlash above 0.03 mm letting the two cranks drift under load and softening the effective phase, (2) connecting rod length mismatch between the two rods exceeding the 0.02 mm pair tolerance, which skews the yoke and shifts piston TDC asymmetrically, or (3) helium charge pressure 10-15% below design, which is not a linkage fault but presents identically as a phase-angle problem on the dyno trace.
When to Use a Rhombic Drive and When Not To
The Rhombic Drive is one of three competing linkages for double-acting Stirling machines. The decision usually comes down to vibration budget, parts count, and how much cylinder real estate you have. Compare it against a conventional alpha-configuration twin-crank slider arrangement and against a Ross yoke, the closest single-crankshaft cousin.
| Property | Rhombic Drive | Alpha-Stirling Twin Slider-Crank | Ross Yoke |
|---|---|---|---|
| Inherent vibration cancellation | First and second order cancelled by symmetry | First order only, needs balance shafts | First order cancelled, second order partial |
| Typical operating speed | 1,000-3,000 RPM | 500-1,500 RPM | 800-2,000 RPM |
| Part count (rotating + linkage) | High — 2 cranks, 4 pins, gears | Medium — 2 cranks, 2 rods | Medium — 1 crank, 1 yoke, 2 rods |
| Build cost (relative) | 1.0× (baseline, highest) | 0.5× | 0.7× |
| Side-load on displacer rod seal | Near zero — coaxial pistons | Moderate — separate cylinders | Low — coaxial pistons |
| Service life to seal replacement | 10,000-20,000 h aerospace grade | 3,000-8,000 h | 5,000-12,000 h |
| Best application fit | Cryocoolers, vibration-sensitive labs, space | Solar dish, large stationary engines | Compact terrestrial Stirlings |
Frequently Asked Questions About Rhombic Drive
Symmetric cancellation only holds if the two crankshafts stay phased to within roughly ±0.5° and the rotating masses on each crank are matched to within about 1%. The most common offender is asymmetric counterweight machining — if one crankweb came off the lathe 2 g lighter than the other, you'll see a residual first-order force at the crankcase even with perfect gear timing.
Check it by running the engine motoring (no gas charge) on an accelerometer. A balanced Rhombic Drive should read under 0.1 g at running speed. If it's reading 0.3 g or more with no gas load, it's a mass-balance problem, not a thermodynamic one.
For 50 W class benchtop work, a Ross yoke usually wins. You get most of the coaxial-piston benefit (low displacer rod side-load) at roughly 70% of the build cost and with one crankshaft instead of two. The Rhombic Drive only justifies its part count when vibration export must be below ~0.1 N, which is an aerospace or precision-metrology spec, not a benchtop one.
Cross over to the Rhombic Drive if your application sits next to an interferometer, an electron microscope, or a focal-plane array. Otherwise the Ross yoke is the practical choice.
Aim for under 0.03 mm circumferential backlash measured at the gear pitch line. Beyond that you get two failure symptoms in sequence — first an audible knock at every stroke reversal as the gear teeth re-seat, then a slow drift of piston phase under load that shows up as falling cooling capacity over the first 100-200 operating hours.
AGMA Q12 ground spur gears or better are typical for aerospace builds. Stamped or hobbed Q9 gears will work for a hobbyist engine but expect the cooler to lose 10-15% of its rated capacity within a year.
The geometric phase from the linkage equation is not the same as the effective thermodynamic phase the gas sees. Three things can shift them apart: yoke flex under peak gas load (a yoke that deflects 0.1 mm at TDC steals 1-2° of phase), helium charge pressure below design (low pressure changes the gas-spring stiffness and shifts the piston dynamic motion away from the kinematic prediction), and dead volume in the regenerator larger than design.
Quick check — pressurise the engine 10% above nominal and re-run the dyno. If capacity climbs back toward the predicted curve, it was a charge-pressure issue. If not, instrument the yoke with a strain gauge and look for flex.
The Rhombic Drive is fundamentally a beta-configuration linkage — both pistons share a single cylinder, which is the defining feature of beta layout. You cannot bolt it to an alpha (separate-cylinder) Stirling because there's no shared bore for the displacer and power piston to reciprocate inside.
If you need a vibration-cancelling linkage for an alpha engine, look at twin-opposed alpha layouts with balance shafts, or move to a four-cylinder double-acting Siemens arrangement instead.
On a Rhombic Drive, premature rod-seal wear almost always traces back to the linkage forcing the rod off-axis. The two suspects are unequal connecting rod lengths (anything over 0.02 mm mismatch tilts the yoke and walks the rod sideways through the seal) and worn yoke pivot bushings allowing the diamond to skew under load.
Pull the engine, mount a dial indicator on the displacer rod tip, and rotate the crankshaft slowly through 360°. Lateral runout above 0.05 mm at the rod tip means the linkage geometry is the cause, not the seal material. Replacing the seal without fixing the linkage just buys you another 2,000 hours.
References & Further Reading
- Wikipedia contributors. Rhombic drive. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
