An Epicyclic Train in slow-motion form is a planetary gear arrangement where two ring gears (or two sun gears) have nearly identical tooth counts, producing an output speed that is the small difference between two large rotations. Reduction ratios of 1000:1 to 10,000:1 in a single compact stage are routine. We use it to get glacially slow, smooth output from a fast input without stacking multiple gearboxes. You will find it inside telescope sidereal drives, clock pendulum-wind mechanisms, and precision rotary indexers.
Epicyclic Train Slow-Motion Interactive Calculator
Vary the two ring tooth counts and see the slow-motion reduction ratio, output advance, and animated differential gear motion.
Equation Used
This calculator follows the worked example: a fixed 100-tooth ring and a 101-tooth output ring have a 1-tooth mismatch, giving a 100:1 slow-motion reduction. The tooth difference sets how far the output ring advances for each carrier revolution.
- Carrier is the input, one ring is fixed, and the second ring is the output.
- Both rings use the same module and compatible tooth geometry.
- The worked-example slow-motion ratio uses the fixed-ring tooth count divided by the tooth difference.
The Epicyclic Train (slow-motion Form) in Action
The trick is differential subtraction. A standard planetary gear train has a sun, planets on a carrier, and one ring gear — fix any one of those three and you get a fixed ratio. The slow-motion epicyclic uses two ring gears stacked on the same carrier, with the planets engaging both rings simultaneously. If one ring has 100 teeth and the other has 101, the planet has to walk one full tooth around the carrier for every full revolution of the input — and that means the output ring rotates only 1/100th of a turn for every 100 turns of the input. Same idea works with two sun gears or a sun-and-ring pair with mismatched tooth counts. This is the compound planetary gear train geometry that makes Ferguson's paradox famous.
The reason designers reach for this layout is package density. A conventional spur reduction at 1000:1 needs 4 or 5 stages, dozens of gears, and a long shaft path. The slow-motion epicyclic does it in one carrier with two ring gears and 3-4 planets — the whole thing fits inside a 60 mm housing on a telescope drive. The catch is tooth-count selection. The two rings must differ by exactly 1, 2, or 3 teeth and must share a common module so the planets mesh both at once. If you pick rings with mismatched modules or a tooth difference that doesn't divide evenly with the planet count, the planets bind and the train locks solid.
When tolerances drift, the symptoms are specific. Backlash in the high-reduction stage gets multiplied by the ratio at the input — 0.1° of ring-gear backlash shows up as 100° of input-shaft slop on a 1000:1 train. Carrier-bearing runout above 25 µm makes the planets walk unevenly between the two rings, producing a periodic torque ripple at the carrier rotation frequency. And the most common failure mode is tooth-edge wear on the planets where they bridge two rings of different tooth counts — the contact pattern shifts axially across the face width as the planet rotates, and undersized face widths chew themselves out within a few hundred hours.
Key Components
- Input Sun Gear or Carrier: Drives the system at high speed. In most slow-motion builds the carrier is the input and one ring is fixed, with the second ring as output. Input shaft runout must stay under 20 µm or the planets see uneven tooth loading.
- Planet Gears (Compound or Single): Mesh with both ring gears simultaneously. Single planets need wide face widths — typically 1.5× the module — because they bridge two slightly different tooth counts. Three or four planets share the load; tooth counts must satisfy the assembly condition (Zr1 + Zr2) / nplanets = integer.
- Fixed Ring Gear: The reaction member. Bolted rigidly to the housing. Tooth count typically 100-120 for clock and telescope work. Concentricity to the carrier axis must be held within 30 µm or planet loading goes asymmetric.
- Output Ring Gear: Differs from the fixed ring by 1-3 teeth. This tooth-count difference IS the reduction ratio mechanism. Mounted on its own bearing concentric with the carrier. The output ring is what carries your driven load — telescope right-ascension wheel, clock minute arbor, indexer table.
- Carrier Plate: Holds the planet shafts at equal radial spacing. Pitch-circle accuracy of the planet bores must be within ±15 µm or the planets fight each other. Carrier bearings see the full input-side speed so they govern the lifespan of the whole train.
Who Uses the Epicyclic Train (slow-motion Form)
Slow-motion epicyclic trains live wherever you need a tiny output rate from a normal-speed input without a long gearbox. The classic uses are astronomy and horology, but precision automation, scientific instruments, and aerospace actuators all use the same compound planetary gear train geometry. The differential epicyclic ratio gives ratios that no single-stage worm or spur set can match.
- Astronomy: Sidereal tracking drive on a Losmandy G11 equatorial telescope mount — the slow-motion epicyclic produces the 1 revolution per 23h 56min sidereal rate from a stepper running at normal speeds.
- Horology: Pendulum-rewind mechanism on a Synchronome electric master clock, where the slow-motion train converts a 30-second impulse into a barely perceptible arbor rotation.
- Precision Automation: Rotary indexing table on a Weiss TC150 cam-driven indexer auxiliary axis, using a Ferguson paradox-style epicyclic to give micro-step positioning between primary cam stations.
- Scientific Instruments: Sample-rotation drive on a Bruker D8 X-ray diffractometer goniometer, where 0.001° resolution requires a 36,000:1 reduction in a compact head.
- Aerospace: Solar-array deployment drive on small satellites — a single-stage slow-motion epicyclic gives the 0.5°/s deploy rate from a brushed DC motor without a multi-stage gearbox adding mass.
- Industrial Robotics: Wrist-roll axis on a Nachi MZ07 small-payload robot, where a compound planetary stage gets the high reduction without the dead band of a harmonic drive.
The Formula Behind the Epicyclic Train (slow-motion Form)
The reduction ratio of a slow-motion epicyclic with two ring gears of different tooth counts is set entirely by the difference in tooth counts and the absolute size of one of the rings. Pick rings 100 and 101 and you get 100:1 from the carrier rotation alone. Pick 200 and 201 and you double the ratio in the same physical envelope. The low end of the practical range is around 50:1 — below that you may as well use a single-stage spur. The high end runs out around 10,000:1 because tooth-count differences smaller than 1 are impossible and rings with 5,000+ teeth become physically unwieldy. The sweet spot for clocks, telescope drives, and small indexers sits between 200:1 and 2000:1.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| i | Reduction ratio from carrier input to output ring | dimensionless | dimensionless |
| Zr1 | Tooth count of the fixed (reaction) ring gear | teeth | teeth |
| Zr2 | Tooth count of the output ring gear | teeth | teeth |
| ωout | Output angular velocity of the second ring | rad/s | rev/min |
| ωin | Input angular velocity of the carrier | rad/s | rev/min |
Worked Example: Epicyclic Train (slow-motion Form) in a polar-axis slow-motion drive on an amateur observatory dome
Sizing the slow-motion epicyclic stage between a stepper-driven input shaft and the dome-rotation worm wheel on a 3.5 m Pulsar Observatories fiberglass observatory dome. The dome must creep around the sky at sidereal rate to track an open shutter against a tracking telescope, which works out to roughly 0.0042 RPM at the dome ring. The stepper runs at a comfortable 60 RPM nominal, so we need an epicyclic that gives about 14,300:1 — but we want to look at how the ratio behaves across the dome's typical operating speed window of 30 to 120 RPM input.
Given
- Zr1 = 120 teeth
- Zr2 = 119 teeth
- ωin,nom = 60 RPM
- ωin,low = 30 RPM
- ωin,high = 120 RPM
Solution
Step 1 — compute the reduction ratio from tooth counts:
That's 120:1 from this single epicyclic stage. To hit the 14,300:1 dome ratio we pair this stage with a downstream 120:1 worm wheel on the dome ring, giving 14,400:1 overall — close enough to sidereal that the dome-control software can trim the rest with stepper microstepping.
Step 2 — at nominal 60 RPM input, compute output ring speed:
Combined with the worm wheel that gives 0.00417 RPM at the dome — exactly sidereal rate. At 60 RPM the stepper runs in its smooth midband, the planets ride centred on the face width, and the dome glides around so slowly the motion is invisible from inside the observatory.
Step 3 — at the low end of the typical operating range, 30 RPM input:
This is the speed you'd use during fine-pointing alignment at dusk. The dome creeps at 0.00208 RPM — half sidereal — and at this rate stepper detent torque becomes the dominant driver. If your stepper has 1.8° step angle without microstepping, the carrier visibly cogs once per second and the planets chatter against the rings. Use 16× microstepping minimum below 40 RPM input.
Step 4 — at the high end, 120 RPM input for slewing between targets:
Dome ring runs at 0.00833 RPM — twice sidereal. Theoretically fine, but in practice the planet-bearing temperature rises noticeably above 100 RPM input on a 120-tooth ring with 1 mm module, and you start hearing the carrier whine through the dome shell. Above 150 RPM the load on the single-tooth-difference contact zone exceeds the Hertzian limit for case-hardened 8620 planets and pitting starts within 200 hours.
Result
The epicyclic delivers a clean 120:1 reduction in one stage, with the output ring turning at 0. 500 RPM nominal. At 0.250 RPM (30 RPM input) the drive is quiet and smooth but flirts with stepper cogging if microstepping is dialled back; at 1.000 RPM (120 RPM input) the train still works but planet-bearing heat and tooth-edge contact stress climb fast — 60 RPM input is genuinely the sweet spot. If you measure dome speed 5-10% off the predicted 0.00417 RPM, suspect three things in this order: (1) a tooth-count assembly error where the ring gears were swapped during build, giving 119:1 instead of 120:1, (2) carrier-bearing axial preload above 50 N causing the planets to skew and lose effective tooth contact, or (3) ring-gear concentricity drift from thermal cycling in an unheated dome, which throws the planet mesh asymmetric and produces a once-per-revolution speed wobble at the carrier frequency.
When to Use a Epicyclic Train (slow-motion Form) and When Not To
Picking a slow-motion epicyclic over its alternatives comes down to ratio density, backlash budget, and how much torque the output has to carry. Compare it against the two layouts that compete in the same ratio band — a multi-stage spur reduction and a harmonic drive (strain-wave gear).
| Property | Slow-motion Epicyclic | Multi-stage Spur Gearbox | Harmonic Drive |
|---|---|---|---|
| Single-stage ratio range | 50:1 to 10,000:1 | 5:1 to 8:1 per stage | 30:1 to 320:1 |
| Backlash at output | 3-15 arc-min (depends on tooth tolerance) | 10-30 arc-min cumulative | <1 arc-min |
| Torque density (Nm/kg) | Medium — limited by single tooth-difference contact | Medium-high | High |
| Cost (relative) | Medium — non-standard ring tooth counts | Low — stock gears | High — specialised components |
| Lifespan at rated load | 10,000-20,000 hours | 20,000-40,000 hours | 8,000-15,000 hours |
| Package length | Short — one carrier deep | Long — stages stacked axially | Short |
| Best application fit | Astronomy, horology, low-torque slow drives | General-purpose industrial reduction | Robotics joints, zero-backlash positioning |
Frequently Asked Questions About Epicyclic Train (slow-motion Form)
The assembly condition fails. For a compound planetary with two ring gears the sum (Zr1 + Zr2) must be divisible by the planet count for the planets to drop into mesh simultaneously at equally spaced positions. 100 + 101 = 201, which doesn't divide by 4. Drop to 3 planets (201/3 = 67, integer) or change the tooth counts to 100 and 104 with 4 planets (204/4 = 51).
You can sometimes brute-force assembly by adjusting carrier hole spacing, but you lose the equal-load sharing that makes the epicyclic durable.
Tooth difference of 1 gives the highest ratio per stage but concentrates contact stress on a single tooth pair at any moment — fine for low torque clock and telescope work, marginal above 5 Nm output. Difference of 2 halves the ratio but spreads load across two simultaneously meshing teeth, doubling effective torque capacity. Difference of 3 is what you use when output torque is the constraint, accepting one-third the ratio.
Rule of thumb: difference 1 below 2 Nm output, difference 2 from 2 to 20 Nm, difference 3 above 20 Nm.
That signature points to ring-gear eccentricity, not tooth error. When the fixed ring is mounted off-centre relative to the carrier axis by even 30-50 µm, each planet sees a slightly different mesh depth as it walks around, and the output ring picks up a once-per-carrier-revolution speed modulation.
Check ring-gear bore concentricity to the carrier bearing journal with a dial indicator — anything above 25 µm TIR is your culprit. Re-shim or re-bore the housing register. Tooth-form errors show up at tooth-mesh frequency, not carrier frequency, so they have a completely different signature on a torque ripple plot.
Technically yes, the train is reversible because all meshes are spur or helical. Practically, no — the high reduction means even tiny output torque sees the input as nearly locked. A 1000:1 train needs 1000× more torque at the output than the input friction torque to back-drive at all, so static friction in the planet bearings usually wins and the system simply doesn't move.
If you need a back-drivable slow-motion stage, switch to a harmonic drive or use a worm-and-wheel with a low lead angle for the irreversibility. Don't fight the epicyclic.
Because the direction depends on which ring has more teeth. If Zr2 > Zr1 the output rotates the same direction as the carrier; if Zr2 < Zr1 it rotates opposite. Swap the two rings physically and you reverse output direction without changing the magnitude of the ratio. That's exactly the demonstration Ferguson built into his original paradox machine — three ring gears stacked on one carrier, all driven by the same planet, each rotating at a different speed and one going backwards.
Check your sign convention before you machine the housing; getting the direction wrong on a telescope drive sends the dome chasing the wrong way across the sky.
For ratios above 500:1, grind them. The reduction multiplies any tooth-form error directly into output position error — a hobbed AGMA Q9 ring with ±20 µm composite tooth error gives you ±20 µm × 500 = 10 mm equivalent error referred to the input, which feels like sloppy positioning at the input shaft. Ground AGMA Q12 or better keeps composite error under ±5 µm.
For ratios under 200:1 in non-instrument applications, hobbed and shaved is usually adequate and saves significant cost.
References & Further Reading
- Wikipedia contributors. Epicyclic gearing. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
