A double Hooke joint is two universal joints connected in series by a short intermediate shaft, with both yokes phased so the velocity fluctuation of the first joint cancels the fluctuation of the second. Unlike a single Hooke (cardan) joint — which produces a sinusoidal speed error twice per revolution at any non-zero angle — the double version delivers true constant angular velocity between input and output. We use it where shafts must transmit torque across an angle without speed ripple, such as 4×4 front driveshafts, steering columns, and PTO drives, eliminating the 2nd-order vibration that single joints generate.
Double Hooke Joint Constant Velocity Interactive Calculator
Vary the two operating angles and yoke phase error to compare single Cardan speed ripple with the residual ripple of a double Hooke joint.
Equation Used
The single Hooke joint velocity ratio varies twice per revolution. In an ideal double Hooke joint, the second joint applies the inverse velocity error, so equal operating angles and zero yoke phase error make the output speed constant.
- Ideal rigid universal joints with no bearing friction, backlash, or shaft compliance.
- Equal angles and zero phase error represent the constant velocity condition.
- Residual double-joint ripple is sampled over one input revolution.
The Double Hooke Joint (constant Velocity) in Action
A single Hooke joint — the classic cardan cross — works fine in straight-line driveshafts but it has one nasty habit: when the input shaft and output shaft are not collinear, the output speed oscillates twice per input revolution. At a 10° operating angle the output speed swings about ±1.5% around the mean. At 30° it swings ±15%. That ripple shows up as torsional vibration, gear chatter, and accelerated U-joint wear. The double Hooke joint solves this by stacking two cardan joints back-to-back on a short intermediate yoke, with the two yoke ears in the same plane. The first joint speeds up while the second slows down by exactly the same amount, and the errors cancel.
The cancellation only works if three conditions hold. First, the two operating angles must be equal — the input shaft and output shaft must converge symmetrically on the centring point of the assembly. Second, the intermediate yoke ears must be phased at 0° relative to each other (same plane), not 90° like you'd find on a long propshaft with two separate U-joints at each end. Third, the joint must be centred — most automotive double cardan joints use a ball-and-socket centring device between the two crosses to force the bisecting geometry. If the centring socket wears, or if the yokes are clocked wrong by even one spline tooth, you lose constant velocity and you get the same 2nd-order vibration the joint was meant to eliminate.
Failure modes are predictable. Worn needle bearings in either cross produce a clunk on torque reversal. A failed centring ball causes a low-speed shudder that gets worse with steering angle on a 4×4 front shaft. Phasing errors from a careless rebuild produce a vibration that scales with shaft RPM squared and is loudest in the 40-80 km/h range on a vehicle. None of this is mysterious — the homokinetic geometry simply stops being homokinetic the moment any of the three conditions slips.
Key Components
- Input Yoke: The driving fork that connects to the input shaft via a spline or flange. It carries one of the two crosses and must align rotationally so its ears sit in the same plane as the intermediate yoke's matching ears — typical phasing tolerance is ±1° to keep velocity ripple below 0.1%.
- First Cross (Spider) and Bearings: A four-trunnion cross running in needle roller bearing cups. Each trunnion typically carries 24-30 needles at 2 mm diameter. The bearing cup press fit into the yoke ear is held to ±0.013 mm — anything looser causes the cup to walk under load reversal.
- Intermediate Yoke (Coupling Yoke): The short double-ended forging that ties the two crosses together. Both sets of ears are machined in the same plane during the same setup so the phase error between them is held under 0.5°. This is the part that distinguishes a double cardan from two separate U-joints on a shaft.
- Centring Ball and Socket: A spring-loaded ball-and-socket assembly between the two crosses that forces the intermediate yoke to bisect the angle between input and output shafts. Without it the joint can transmit torque but loses its constant-velocity property. Wear here is the most common cause of front-driveshaft shudder on solid-axle 4×4s.
- Second Cross (Spider) and Bearings: Identical to the first cross. Its phasing relative to the intermediate yoke is what cancels the velocity error generated by the first joint. A 90° phasing mistake here turns the unit into a velocity-error multiplier instead of a canceller.
- Output Yoke: The driven fork connecting to the output shaft, pinion flange, or steering box input. It must operate at the same angle as the input yoke relative to the intermediate yoke — typically held within 0.5° during assembly to keep residual ripple negligible.
Who Uses the Double Hooke Joint (constant Velocity)
You find double Hooke joints anywhere a shaft has to transmit torque across an angle without speed ripple, but doesn't need the wide angular range of a Rzeppa-type ball CV joint. The geometry is cheap, rugged, and tolerant of dirt, which is why heavy vehicles and industrial drives still use it 100 years after its automotive debut. The key trade-off — angles must be split equally between the two halves — drives where you can and can't use it.
- Automotive (Light Truck 4×4): Front driveshaft on solid-axle pickups like the Jeep Wrangler JK Rubicon and Ford Super Duty F-250 4×4, where the shaft runs at angles up to 25° when the suspension articulates.
- Heavy Truck Driveline: Inter-axle shafts on tandem-drive Class 8 tractors from Mack and Kenworth, where the two drive axles sit at slightly different heights and a single U-joint would introduce torsional ripple into the differential.
- Steering Columns: Intermediate steering shafts on passenger cars where the column must clear the engine and reach the steering rack at a compound angle — used in BMW, Mercedes, and most pickup trucks. Constant velocity here means no torque feedback ripple in the driver's hands.
- Agricultural PTO Drives: Wide-angle PTO shafts on Walterscheid and Bondioli & Pavesi assemblies driving rotary mowers and round balers, where the implement angle changes constantly as the tractor turns.
- Rail Vehicles: Cardan shaft drives between traction motors and axle gearboxes on diesel locomotives like the EMD GP38-2, where the bogie pivots relative to the carbody and the drive must accommodate angle without speed variation.
- Industrial Machinery: Roll drives on steel mill stands and paper machine calenders where the driven roll position varies during operation but speed must remain ripple-free to avoid surface marking.
The Formula Behind the Double Hooke Joint (constant Velocity)
The relationship that matters is the velocity ratio between input and output of a single Hooke joint as a function of input angle and operating angle. Run this for two joints in series with the phasing condition satisfied, and the errors cancel — but only if the operating angles are equal. At the low end of the typical operating range, say 5°, a single joint produces about ±0.4% velocity ripple, so even a sloppy double cardan rebuild would feel smooth. At the design sweet spot of 10-20°, the cancellation is essentially perfect when angles are matched, and any residual ripple comes from phasing or angle mismatch. Push the operating angle to the high end — 30° or more — and small angle mismatches between the two halves explode into noticeable vibration, because the underlying error term scales with tan²(β).
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| ωin | Input shaft angular velocity | rad/s | RPM |
| ωout | Output shaft angular velocity (single-joint instantaneous) | rad/s | RPM |
| β | Operating angle between shafts at one joint | degrees or radians | degrees |
| θ | Input shaft rotational position | degrees or radians | degrees |
| Δω/ω | Peak velocity ripple of single joint ≈ ½ × tan2(β) for small angles | fraction or % | fraction or % |
Worked Example: Double Hooke Joint (constant Velocity) in a solid-axle 4×4 front driveshaft
You are evaluating the front driveshaft on a lifted Jeep Wrangler JK with a 4-inch suspension lift. The transfer case output sits at the stock height but the front pinion has rotated upward, giving a total operating angle of 18° split between the two halves of a double cardan joint at the transfer case end. The shaft runs at 1500 RPM in 4-high at 80 km/h. You want to know what velocity ripple to expect at the front pinion if the angles are split correctly versus if the rear half ends up 3° off due to a worn centring ball.
Given
- βtotal = 18 degrees
- ωin = 1500 RPM
- β1 nominal = 9 degrees each half
- Mismatch case = β1=12, β2=6 degrees
Solution
Step 1 — at the nominal split where each half operates at 9°, the single-joint peak ripple is approximately ½ × tan2(β):
Each individual joint swings ±1.25% twice per revolution. With proper phasing and equal angles, the second joint's swing exactly cancels the first, so the net output ripple at the pinion is essentially zero — under 0.05% in practice, limited only by manufacturing tolerances on yoke clocking.
Step 2 — at the low end of the operating range, with the suspension at full droop and total angle dropping to 6° (3° each half), single-joint ripple is only ½ × tan2(3°) = 0.14%. Even a sloppy phasing job here would be undetectable in the cab. This is why stock-height 4×4s rarely complain about front driveshaft vibration.
Step 3 — at the high end, with the angles mismatched 12°/6° because of a worn centring ball that has let the intermediate yoke drift off the bisector:
That 1.71% residual ripple at 1500 RPM means a torque pulse hitting the front diff at 50 Hz with enough amplitude to feel through the floor and the steering wheel. At 80 km/h cruise this shows up as the classic Jeep front-driveshaft shudder that owners chase for years.
Result
With angles correctly split 9°/9°, the double Hooke joint delivers effectively zero velocity ripple at the pinion — the driveshaft feels as smooth as a CV-jointed shaft. At 3°/3° (stock ride height) any error is invisible; at 9°/9° (correct lifted geometry) the joint cancels properly; but at the 12°/6° mismatch caused by a sagged centring ball, residual ripple jumps to about 1.7%, producing a 50 Hz vibration pulse at 80 km/h that's loud enough to feel in your hands. If your measured vibration doesn't match the prediction, the most likely causes are: (1) a worn centring ball-and-socket allowing the intermediate yoke to drift off the bisector, (2) the pinion yoke clocked one spline tooth off during a u-joint replacement so the two halves are 30° out of phase instead of 0°, or (3) a bent intermediate yoke from a rock strike that's introducing its own runout-driven ripple independent of operating angle.
Choosing the Double Hooke Joint (constant Velocity): Pros and Cons
The double Hooke joint sits between the simple, cheap, ripple-prone single cardan and the smooth, expensive, angle-limited Rzeppa CV joint. Each option wins in a different operating envelope, and the right pick depends on operating angle, RPM, lubrication access, and cost.
| Property | Double Hooke Joint | Single Hooke (Cardan) Joint | Rzeppa Ball CV Joint |
|---|---|---|---|
| Velocity ripple at 15° operating angle | ≈ 0% (when angles matched) | ±3.4% peak, 2× per rev | ≈ 0% |
| Maximum continuous operating angle | 30° | 15-20° (vibration limited) | 47° |
| Continuous RPM rating (typical 1310-series) | 6000 RPM | 6500 RPM | 8000+ RPM |
| Torque capacity (relative) | High (same as cross size) | High | Medium (ball-track limited) |
| Maintenance interval (greased version) | 12,000-25,000 km service | 12,000-25,000 km service | Sealed for life |
| Tolerance to dirt and water | Excellent (open or boot) | Excellent | Poor (boot failure = death) |
| Relative cost (assembled) | 1.8× | 1.0× baseline | 2.5-3.5× |
| Phasing sensitivity | High — must be 0° clocked | Low (single joint) | None (inherently CV) |
| Best application fit | 4×4 front shafts, PTOs, steering | Long propshafts at small angles | FWD axles, high-angle compact |
Frequently Asked Questions About Double Hooke Joint (constant Velocity)
A new joint doesn't fix the underlying geometry problem. When you lift a solid-axle truck, the transfer case output stays put but the pinion drops away from it, and unless you rotate the pinion to point directly at the centring ball of the double cardan, the two halves end up at unequal angles. The whole point of the double cardan is angle bisection — break that and you're back to single-joint ripple.
Measure the angle of the transfer case output yoke and the front pinion yoke with a digital level. The pinion should point at the t-case output, which means its angle should equal the shaft angle — not be parallel to the t-case output as it would be on a single-cardan setup. If the numbers aren't right, you need a higher pinion angle, achieved with cam bolts, adjustable control arms, or shimmed axle pads.
Two field checks. First, with the shaft removed and held horizontally, the intermediate yoke should fall under gravity to bisect any angle you put between the input and output yokes — flop the yokes to a 20° bend and the centre section should self-centre cleanly. If it lags, hangs, or feels gritty, the ball-and-socket is worn or dry.
Second, on the vehicle, the giveaway symptom is a low-speed shudder under light throttle that gets worse with steering angle — different from a u-joint clunk, which happens on torque reversal. The shudder comes from the intermediate yoke no longer bisecting the angle, so velocity error returns. Replacement of the ball and socket usually requires the whole CV head, since the components aren't typically sold separately.
No, and the reason is angular range. A FWD outer joint sees up to 47° during full lock plus suspension travel. A double cardan tops out around 30° before the crosses bind on the yoke ears, and even then you need the centring ball to keep the geometry honest under bump steer. Rzeppas use ball-in-track geometry that stays homokinetic at any angle within their design envelope without needing a bisecting mechanism.
The double cardan also fails on packaging — it's roughly twice as long as a Rzeppa for the same torque rating, and you simply don't have that length between the diff and the wheel hub on a transverse-engine FWD car.
Phasing. The intermediate yoke ears on both ends must lie in the same plane — 0° clocking. If you pulled the shaft apart at the slip yoke and reassembled it 180° rotated, you may not realise the splines let you re-clock by 90° relative to the CV head's intended orientation, which destroys the cancellation.
Look for the alignment arrows or paint marks the OEM stamps on the slip yoke and tube. If they're missing because someone replaced a tube, find a reference picture and check that the front yoke ears at the pinion end are parallel to the intermediate yoke ears at the CV head end. Rotated 90° and you have a velocity-error amplifier instead of a canceller.
Yes — a small inertial moment that single joints don't generate. The intermediate yoke is not a point mass; it has rotational inertia, and because its angular velocity oscillates at twice input frequency (even though net input-to-output is constant), it absorbs and releases kinetic energy each half-revolution. This shows up as a 2nd-order torsional reaction at the input bearing.
For most automotive applications it's negligible — under 1 Nm for a 1310-series at 3000 RPM — but on high-RPM industrial drives above 4000 RPM it's enough to require a torsional damper or a balanced intermediate yoke. Walterscheid and GWB datasheets list this 2nd-order moment for their wide-angle PTO units.
The difference is whether the angle is concentrated or distributed. A double cardan is one assembly that cancels velocity error locally — both halves operate at half the total angle, in opposite directions, at the same physical location. Twin U-joints on a long shaft (input and output U-joints clocked 0° to each other, with parallel input and output shafts) cancel error too, but only if the input shaft is parallel to the output shaft and the operating angles at each end are equal.
Use a double cardan when input and output shafts converge on a point (like a transfer case output to a pinion that's pointing at it) — this is the 4×4 front shaft case. Use twin phased U-joints when input and output shafts are parallel but offset (like a rear propshaft from t-case to rear diff). Mix them up and you'll get vibration no matter how new the parts are.
References & Further Reading
- Wikipedia contributors. Universal joint. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
