A six-gear equal-diameter bevel doubler is a compound bevel train that uses six identical miter bevel gears arranged so the input shaft drives a stack of three meshed pairs, and the output rotates at twice the input speed. Unlike a single-stage 2:1 spur or worm reduction running in reverse, this layout keeps every gear the same diameter and uses geometry — not tooth-count differences — to multiply speed. It exists where you need clean speed doubling in a tight right-angle envelope. Output runs at 2 × input RPM with torque halved minus mesh losses, typically delivering 92–96% efficiency per stage.
Six-Gear Bevel Speed Doubler Interactive Calculator
Vary input speed and mesh lash to see the 2:1 output speed, idler relative speed, and backlash stack-up.
Equation Used
The six equal-diameter bevel gears do not get their 2:1 ratio from tooth-count differences. The output speed is twice the input speed because of the carrier and idler kinematics. Backlash is estimated as three mesh clearances added together.
- All six bevel gears are identical diameter and tooth count.
- The carrier kinematics set an exact 2:1 speed ratio.
- Backlash is summed linearly across three meshed pairs.
- Speed calculation is kinematic and does not include efficiency losses.
How the Doubling Speed with Equal-diameter Bevel Gears (six-gear) Actually Works
The trick of the six-gear doubler is that every bevel gear is identical — same pitch diameter, same tooth count, same pressure angle. You cannot doubt-check the ratio by counting teeth between input and output, because the ratio comes from how many gears spin on free shafts versus fixed shafts, and from the carrier geometry. Three of the six gears sit on the input/output axis. The other three are idler bevels mounted on a carrier or floating shafts. As the input rotates, the idlers walk around their meshes, and the carrier-plus-mesh geometry forces the output to complete two revolutions for every one input revolution. The math collapses to a clean 2:1 because the equal diameters cancel everything except the carrier kinematics.
Why build it this way instead of a 1:2 spur step-up? Two reasons. You get a right-angle output without bending the load path through a separate miter pair, and you get balanced radial forces because the idler bevels are arranged symmetrically around the centreline. That symmetry is what lets the doubler run at higher input speeds — 3000 RPM input is reasonable in a hardened-steel build — without the side-loading problems a single-pair step-up would dump onto the input bearing.
Tolerances bite hard here. Backlash stacks across three meshes, so if each pair has 0.05 mm of circumferential play you'll see roughly 0.15 mm at the output — enough to feel as a perceptible lash if the doubler drives an indexing head. The bore on every gear must match the shaft to within H7/g6 fit, and the cone-distance error between any two paired bevels must stay under 0.02 mm or you get tooth-corner contact, which sounds like a sewing machine running dry within 50 hours of use. Common failure modes are idler-shaft bushing wear (idlers see 2× the relative speed of the input), pitting on the loaded flank of the centre-stage gear, and carrier flex if the carrier plate is thinner than 0.4 × the gear face width.
Key Components
- Input bevel gear: Sits fixed on the input shaft and meshes with the first idler. Carries the full input torque, so it's typically the most heavily case-hardened gear in the set — 58–62 HRC on the tooth surface with a tough core around 30 HRC.
- Three idler bevels: Mounted on free-spinning shafts in the carrier. They rotate at roughly 1.5× input speed and transfer motion between the input-side and output-side gears. Their bushings or needle bearings see the highest relative velocity in the train and wear first.
- Carrier plate: Holds the three idler shafts in fixed relative position. Must be stiff enough that idler-shaft deflection stays under 0.05 mm at full load — a common rule is carrier thickness ≥ 0.4 × gear face width in steel.
- Intermediate bevel gear: The middle gear in the stack that bridges the idler array to the output gear. Same dimensions as the input gear but typically sees lower mean torque because losses have already taken 4–8% off the power flow.
- Output bevel gear: Spins at 2 × input RPM. Mounted on a separate output shaft supported by its own bearing pair. Bearing selection here matters more than on the input — doubled speed means bearing L10 life drops by roughly the cube of the speed ratio if you don't upsize.
- Housing and bearing pairs: Constrains the input and output shafts coaxially or at right angles depending on the variant. Bearing preload of 5–15 µm axial is typical to keep cone-distance error inside the 0.02 mm window across thermal swings.
Where the Doubling Speed with Equal-diameter Bevel Gears (six-gear) Is Used
You see this layout where designers need a clean speed-up in a compact, symmetric envelope and want to avoid the noise and efficiency penalties of a single-pair 1:2 step-up. It shows up most often in machine tools, light industrial spindles, and a handful of historical textile machines where input shaft speed was fixed by line shafting but spindle speed needed to double.
- Machine tools: Speed-up head on a Bridgeport-style vertical mill — converts the standard 60–4200 RPM spindle range to 120–8400 RPM for small-diameter end milling without changing the main motor.
- Textile machinery: Doubling drives on early Northrop automatic looms where the loom shaft ran at one speed and the bobbin-winding spindle needed double that speed.
- Marine drives: Auxiliary alternator step-up on small inboard diesels — takes a 1500 RPM crank PTO and delivers 3000 RPM to the alternator within a confined right-angle envelope.
- Robotics: Wrist roll-axis doublers on industrial robot arms similar to older Fanuc M-series wrists where the upstream gearmotor RPM was fixed but the end-effector needed faster rotation.
- Printing presses: Inking-roller speed multipliers on Heidelberg sheet-fed presses — the main drive runs at sheet-feed RPM and the inking train needs double speed for proper ink film distribution.
- Test rigs: Tachometer calibration jigs — a known input RPM driving a precisely 2.000:1 doubler gives a verified second reference point for calibrating shop tachometers without a second motor.
The Formula Behind the Doubling Speed with Equal-diameter Bevel Gears (six-gear)
The output speed of a six-gear equal-diameter bevel doubler is fixed by geometry, not tooth count, so the formula is short. What varies in practice is what the practitioner can actually feed into it. At the low end of typical operation — say 200 RPM input on a hand-cranked or low-speed industrial drive — the output sits at 400 RPM and mesh losses are negligible because tooth sliding velocity is low. At the nominal range of 1000–1800 RPM input the doubler runs in its sweet spot, with efficiency holding above 92% and bearing temperatures stable. Push input above roughly 3000 RPM and the formula still says 2×, but the idler-shaft bearings start to dominate the loss budget and the predicted output torque you back-calculate from input power will fall well short of theory.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Nout | Output shaft rotational speed | rev/min (RPM) | RPM |
| Nin | Input shaft rotational speed | rev/min (RPM) | RPM |
| Tin | Input torque applied to the input bevel | N·m | lb·ft |
| Tout | Output torque available at the output bevel | N·m | lb·ft |
| η | Combined mesh and bearing efficiency across the three stages | dimensionless (0–1) | dimensionless (0–1) |
Worked Example: Doubling Speed with Equal-diameter Bevel Gears (six-gear) in a watchmaker's lathe live-spindle booster
Designing a bolt-on speed-doubling head for a Schaublin 70 watchmaker's lathe. The main spindle runs at a maximum of 4000 RPM but a customer wants to run 0.3 mm carbide micro-end-mills that need surface speed equivalent to roughly 8000 RPM at 0.3 mm tool diameter. The doubler uses six identical Module 0.8, 24-tooth steel bevel gears with η ≈ 0.94 across the three meshes. Input torque from the lathe spindle at full motor load is measured at 2.4 N·m.
Given
- Nin,nom = 4000 RPM
- Tin = 2.4 N·m
- η = 0.94 —
- Module = 0.8 mm
- Z (each gear) = 24 teeth
Solution
Step 1 — at nominal 4000 RPM input, compute output speed:
Step 2 — compute output torque at nominal, accounting for the 6% loss across three meshes:
That's the design point — 8000 RPM at the cutter with just over 1 N·m available. For a 0.3 mm carbide end-mill that's wildly more torque than the tool can absorb before snapping, so the doubler is not torque-limited at this scale; it's speed-limited by the bevel teeth themselves.
Step 3 — at the low end of the typical lathe operating range, 1500 RPM input (a common setting for centring and indicator work):
At 3000 RPM the doubler is barely working — tooth sliding velocity is around 1.5 m/s, well below the 8 m/s threshold where Module 0.8 steel bevels start to need oil-mist lubrication, and you can safely run grease-packed. Idler bearings are barely warm.
Step 4 — at the high end, push input to the lathe's redline of 4000 RPM and beyond if a VFD overspeeds the motor to 4800 RPM:
9600 RPM at the cutter sounds attractive but the idler bevels are now spinning at roughly 14,400 RPM on their bushings. Sintered-bronze idler bushings rated for 6000 RPM continuous will overheat within minutes. Above 4000 RPM input you must switch to needle-roller idler bearings or the doubler self-destructs from idler-bushing failure long before the gear teeth complain.
Result
The nominal output is 8000 RPM at 1. 128 N·m — exactly what the micro-milling job needs. At the low end of the lathe's operating range (1500 RPM input → 3000 RPM output) the doubler is grease-lubricated and runs cool indefinitely, while at the high end (4800 RPM input → 9600 RPM output) the limit is no longer the gears but the idler bushings, which will fail before the teeth show any wear. If you measure output speed lower than 2× input by more than ~1%, the most common causes are: (1) a slipping setscrew on the input bevel hub letting the gear lag the shaft under load, (2) carrier flex from a too-thin carrier plate allowing one idler shaft to walk out of mesh during torque spikes, or (3) a cone-distance error above 0.02 mm between any paired bevels causing tooth-corner contact and intermittent slip. If output torque is more than 10% below the predicted 1.128 N·m, suspect idler-bushing drag from undersized clearance — bronze bushings need 0.025–0.04 mm diametral clearance at this speed, not the 0.015 mm typical for slow-speed work.
Choosing the Doubling Speed with Equal-diameter Bevel Gears (six-gear): Pros and Cons
Speed doubling can be done a dozen ways. The six-gear equal-diameter bevel doubler is one of the more elegant but also one of the more part-count-heavy options. Here's how it stacks up against the two practical alternatives a designer actually considers — a single-pair 1:2 spur step-up, and a 1:2 planetary step-up.
| Property | Six-gear bevel doubler | 1:2 spur step-up pair | 1:2 planetary step-up |
|---|---|---|---|
| Maximum input speed (steel, oil-mist) | 6000 RPM | 10,000 RPM | 8000 RPM |
| Efficiency per stage at nominal load | 92–96% | 97–99% | 94–97% |
| Backlash at output (typical, three-mesh stack vs single mesh) | 0.10–0.20 mm | 0.04–0.08 mm | 0.05–0.12 mm |
| Part count (gears + carrier) | 6 gears + carrier | 2 gears | 4–5 gears + carrier |
| Output orientation flexibility | Right-angle native | Inline only | Inline only |
| Radial load on input bearing | Low (symmetric idlers) | High (single-side mesh) | Low (symmetric planets) |
| Relative cost (small batch, machined steel) | High | Low | Medium |
| Best application fit | Right-angle speed-up in tight envelope | Inline step-up where space is open | Inline step-up where backlash and concentricity matter |
Frequently Asked Questions About Doubling Speed with Equal-diameter Bevel Gears (six-gear)
The 2:1 ratio is geometric, so the only way the output undershoots is mechanical slip or lash. The most frequent culprit on a freshly built doubler is the input bevel hub setscrew bottoming on a flat that's not perfectly aligned — the gear creeps relative to the shaft under torque and gives you what looks like a fractional ratio loss on a tachometer.
Diagnostic check: mark the input shaft and input bevel with paint and run for 5 minutes under load. If the marks misalign, lock the gear with a key or a Loctite 638 retain — not a setscrew alone. If the marks stay aligned, your loss is true mesh slip from cone-distance error or backlash being driven open by reversing loads.
Almost never, if you genuinely have space. A spur pair gives you 97–99% efficiency in one stage versus 92–96% across three meshes, and one-tenth the part count. The doubler earns its keep only when you need a right-angle output and a speed increase in the same envelope, or when you can't tolerate the input-bearing side load that a single-pair step-up creates.
Rule of thumb: if your envelope allows a spur pair with 1.5 × output gear diameter of clearance, use the spur. The doubler is for cases where you're stuck at right-angle and the alternative is two separate gearboxes bolted together.
The idler shafts are spinning at roughly 1.5× input speed and the idler bushings see the highest PV (pressure × velocity) of anything in the train. Bronze bushing wear opens the idler shaft clearance, the idler walks slightly off cone-distance, and you get tooth-flank entry impact — that's the new noise.
Pull an idler and check radial play with a dial indicator. Anything above 0.06 mm at the gear OD means the bushing has worn past spec. Either tighten clearances by replacing bushings or — better on rebuild — switch to needle rollers, which extend service life roughly 5× at the same speed.
Geometrically yes, mechanically rarely. Running it as a 2:1 reducer means the output (now slow) shaft carries 2× the input torque minus losses, and the carrier and idler shafts are now sized backwards — they were dimensioned for the high-speed/low-torque side. You'll overload idler shafts that were specified for the doubler duty.
If you want a 2:1 reducer in the same envelope, design it as a reducer from the start: thicker idler shafts, larger idler bushings, and a stiffer carrier. A doubler reversed is a reducer that's about 60% of the torque rating of a purpose-built reducer of the same size.
For a hobby build running below 1000 RPM input you can get away with 0.05 mm cone-distance error per pair. Across three pairs the stack-up means one of the meshes will run on the corner of the tooth flank rather than the full face, but at low PV the gears self-burnish and find a working contact pattern within the first hour.
Above 2000 RPM input, 0.05 mm is too loose. Tooth-corner contact at higher sliding velocity scuffs the case-hardened layer, and once the hardened skin breaks through you lose teeth in 20–50 hours. Tighten to 0.02 mm for any high-speed industrial use.
Don't try to measure end-to-end backlash by holding the input fixed and rocking the output — you'll get a number that's the sum of three meshes plus carrier flex, which is meaningless for diagnosing which mesh is the problem. Instead, lock each idler in turn with a temporary pin and measure the lash at the output. Subtract pair-by-pair to isolate the worst mesh.
Total stack lash should land between 0.10 and 0.20 mm at the output gear OD for a Module 0.8 set. Above 0.25 mm something is wrong — usually one carrier hole drilled oversize or a bushing that's already past wear limit.
References & Further Reading
- Wikipedia contributors. Bevel gear. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
