Gear Design Basics: Choosing Types and Modules
A beginner's guide to gear design covering classification (spur/helical/bevel/worm/planetary), module selection principles.
The design decision behind it
Module, pressure angle, tooth count and profile shift are usually fixed early, often before anyone asks whether the part can be cut cleanly at a sensible cost. That order is backwards. Treat these four as a set of trade-offs rather than independent choices.
Start from the function of the pair, not the function of one gear. A coarse module raises root thickness and bending capacity, but it also raises tooth height, sliding velocity at the flank and the blank diameter for a given tooth count. A finer module with a wider face width often carries the same load in a smaller envelope, at the price of tighter control on alignment and lead. There is no universally correct side of that trade; there is only the side that matches your housing space, your lubrication and your expected life.
Pressure angle behaves the same way. 20° is the common compromise between root strength and sliding ratio. 14.5° runs more smoothly but leaves a thinner root. 25° buys root strength and pays for it with higher sliding and more sensitivity to mesh misalignment. Profile shift is the fine adjustment on top: positive shift on the pinion restores tooth thickness and helps avoid undercut at low tooth counts, while the sum of the two shifts decides whether the centre distance stays nominal or moves.
The decision that is most often skipped is life balance across the pair. The pinion has fewer teeth, sees more load cycles per revolution and usually fails first. If the wheel is specified far harder or far stronger than the pinion, the transmission’s life is set by the pinion anyway, and the extra capacity in the wheel is paid for and never used. Decide which member you are willing to replace, then distribute hardness, shift and material accordingly.
Numbers that matter on the drawing
A gear drawing is only manufacturable when the geometry is closed — that is, when every value needed to generate the flank can be derived without guessing. The table below lists what a reviewer will look for, and the measurement convention that goes with each.
| Parameter | What to state | Convention to state with it |
|---|---|---|
| Module or diametral pitch | One system only, plus the standard referenced for the series | Module is a linear value; DP is its reciprocal — never mix on one drawing |
| Tooth count | Exact integer for each member of the pair | Include the mating part’s count so the ratio can be checked |
| Pressure angle | Nominal value and whether it is the reference profile angle | Distinguish reference profile from working pressure angle |
| Profile shift | Coefficient per member, and the intended centre distance | State whether the pair is equal-shift or angular-shift |
| Helix angle and hand | Angle, hand, and the resulting centre distance | Hand must be given for both members of a helical pair |
| Accuracy | The standard used as the marking basis, and which deviations are controlled | Cite the standard (ISO 1328, DIN 3962, AGMA 2001) as the basis; the grade itself is your call, not ours |
| Backlash | Minimum and maximum circumferential backlash at the working centre distance | Give the measuring plane and the condition (new, at temperature) |
| Datum and inspection | Which surfaces are the functional datums; over-pins or span measurement | Bore and one face are the usual datums — say so explicitly |
Two habits make drawings reviewable. First, keep the reference profile data and the inspection data in separate blocks so a change in one does not silently move the other. Second, give the centre distance and backlash as a range with the temperature or preload condition attached; a single nominal number invites a part that fits on the bench and binds in service.
Where custom manufacturing changes the answer
Because every part here is cut to a drawing or a sample, the questions that decide feasibility are geometric and fixturing-related, not catalogue-related. Four of them come up repeatedly.
Tooth count and undercut. Below the undercut limit for the chosen pressure angle, the flank cannot be generated without removing root material. Positive shift is the usual remedy, but it must be on the drawing as a coefficient, not left to the shop. If the drawing shows a low tooth count with no shift, expect a query rather than a part.
Internal teeth and interference. Internal gearing gives a larger curvature radius and lower contact stress, which is why it appears in planetary carriers. It is also harder to cut, and the tooth count difference between internal and mating external gear is bounded by interference. A drawing that ignores that bound will not run regardless of accuracy.
Narrow features and access. Thin webs, deep tooth spaces, small root fillets and hard-to-reach flanks all constrain how the tool reaches the cut. These are the parameters that move cost most, well before tolerance does. A slightly more open root or a wider clearance often removes a whole operation.
Heat treatment versus distortion. Carburising and induction hardening both move the part after cutting. If the drawing puts tight lead or profile limits on a case-hardened gear without saying how the flanks are to be finished after treatment, the requirement is incomplete. Say whether the final geometry is cut before or after hardening.
Common mistakes in drawings and enquiries
Mixed unit systems. Module on one view, diametral pitch in a note. Consequence: the part is cut to the wrong size and the error is only found at mesh. Fix: one system, stated once, referenced to a standard series.
Shift omitted on a low-tooth-count pinion. Consequence: undercut or a rejected drawing query. Fix: give the shift coefficient, or state that the supplier may select one and report it.
Accuracy stated as a single word. “Precision ground” without a standard or a controlled deviation. Consequence: no objective accept/reject basis. Fix: cite ISO 1328, DIN 3962 or AGMA 2001 as the marking basis and list which deviations are controlled.
Backlash missing or given as one number. Consequence: a pair that binds when warm or rattles when cold. Fix: min/max backlash at the working centre distance, with the condition stated.
No functional datum. Bore and face implied but not marked. Consequence: inspection argues with machining about which surface is the reference. Fix: mark the datums, and specify over-pins or span measurement for flank control.
A short checklist
- One unit system only (module or DP), with the standard series named
- Tooth counts for both members of the pair, ratio verified
- Pressure angle and reference profile angle stated
- Profile shift coefficient per member, and intended centre distance
- Helix angle and hand given for each member (helical pairs)
- Accuracy standard cited as the marking basis, with controlled deviations listed
- Backlash range at the working centre distance, with condition
- Functional datums marked; flank measurement method named
- Heat treatment and the stage at which final flank geometry is produced
- Material and hardness for each member, checked for life balance
Does a higher module always mean a stronger gear?
No. A larger module thickens the root, but it also raises tooth height and sliding ratio and lowers contact ratio, so wear and smoothness can get worse. For a given load, a finer module with a wider face is often the better answer. The right choice depends on housing space, lubrication and expected life, not on module alone.
Can I leave the profile shift for the supplier to decide?
You can, but say so on the drawing. If the shift is left open on a low-tooth-count pinion, the shop has to choose between undercut and an unrequested geometry change. Stating “shift to be selected by supplier and reported” keeps the decision traceable and avoids a rejected drawing.
How do I specify accuracy without over-constraining the part?
Cite the standard as the marking basis (ISO 1328, DIN 3962, AGMA 2001) and then list which deviations actually matter for your mesh — usually profile, lead and runout. That gives an objective accept/reject basis without demanding control over deviations the application does not use.
Send your drawing to sales@dgpinyi.com or call +86 18002857601 and we will review the geometry and manufacturability before quoting.