A gear pump manufacturer was cutting its 12-tooth drive gears from bar stock. Each gear: turn, hob, deburr, and harden — about 4 minutes of machining time, $6 in material, and $3 in labor. A powder metal (PM) supplier quoted the same gear sintered to net shape at $4.20 per piece. The pump maker tested 100 PM gears. They measured backlash, tooth profile, and durability against the machined version. The PM gears passed. The pump maker switched. The gears cost 40% less and the pump’s warranty failures didn’t change.
How a PM gear is made
Powder metal gears start as metal powder — iron or steel powder with alloying elements (carbon, nickel, copper, molybdenum) mixed in. The powder is compacted in a die under high pressure (400-700 MPa) into the gear shape, teeth and all. The green (unfired) compact is then sintered — heated to about 1120°C in a controlled atmosphere, where the powder particles fuse into a solid part. The gear comes out of the furnace at net shape.
The gear teeth are formed by the die cavity, not cut. This is the whole economic argument: no hobbing, no shaping, no skiving. The tooth profile is a direct copy of the die. The tolerances are set by the die precision (and the sintering shrink, which the die compensates for).
The density question
The catch with PM gears is density. A sintered PM gear at 6.8 g/cm³ (typical for a basic iron-copper-carbon mix) is about 87% of wrought steel’s 7.85 g/cm³. The pores reduce the gear’s strength — the tooth bending fatigue strength at 6.8 g/cm³ is roughly 60-70% of wrought steel. That’s why PM gears historically failed at the tooth root in high-load applications.
The fix is density. Modern PM processes push density up:
| Process | Density (g/cm³) | % of wrought | Relative cost |
|---|---|---|---|
| Standard press-and-sinter | 6.8-7.0 | 87-89% | 1.0x |
| Sinter-hardening | 6.9-7.1 | 88-90% | 1.1x |
| Double press, double sinter | 7.3-7.5 | 93-95% | 1.4x |
| Warm compaction | 7.3-7.5 | 93-95% | 1.3x |
| Powder forging (P/F) | 7.8 | 99% | 1.8x |
| Metal injection molding (MIM) | 7.6-7.8 | 97-99% | 2.0x (small parts only) |
The pump gear was a sinter-hardened material at 7.0 g/cm³. The tooth root fatigue strength measured about 75% of the machined 4140 gear. But the pump only loaded the gear to 40% of its fatigue limit — the safety factor was 2.4x. The PM gear was fine for the duty.
The pump gear’s real test
Twelve-tooth, 2.5 mm module gear in a 6 cc gear pump at 1800 RPM, 150 bar. The tooth root stress at the rated pressure was about 180 MPa. The PM sinter-hardened gear (7.0 g/cm³, hardened to 40 HRC) had a root fatigue strength around 480 MPa. Safety factor 2.7. The machined 4140 gear (hardened to 42 HRC) had a root strength around 620 MPa. Safety factor 3.4. Both ran a 2000-hour endurance test. Neither broke. The PM gear wore the same amount — 0.02 mm on the tooth flank — because both were hardened to similar hardness and the lubrication was identical.
Backlash was the interesting one. The PM gear came out of the die with the tooth profile copied exactly. The shot-peened… no, PM gears in this pump were used as-sintered. The backlash measured 0.03-0.05 mm, matching the machined gears. The pump’s noise was within 1 dB. The pump maker’s customers couldn’t tell the difference.
When PM gears are the wrong call
PM loses in three situations.
1. High torque shock loads. A gearbox that sees torque spikes (a crusher, a punch press drive) will fatigue the PM tooth root faster than wrought steel. The porosity is a crack initiation site. Use wrought steel or powder forging for shock-loaded gears.
2. Very high precision (DIN 5 or better). Sintered gears run DIN 7-8 as-sintered. The shrinkage during sintering is about 1-2% and varies slightly across the part. A precision gear needs either a sinter-hardening plus rolling (the teeth are rolled after sintering to improve the profile) or a final machining operation. That adds cost and removes the net-shape advantage.
3. Small quantities. The die costs $20,000-$60,000 for a gear. At 100 pieces per year, the die amortization is $200-600 per piece — worse than machining. PM makes sense at volumes over 10,000 pieces per year, where the die cost spreads thin. The pump maker ran 50,000 pumps a year — the die paid for itself in the first month.
The design rules for PM gears
- Minimum tooth count about 12 (smaller teeth are hard to fill in the die)
- Add a 0.3-0.5 mm radius at the tooth root (the die can’t hold a sharp corner, and the stress concentration drops anyway)
- Keep the face width uniform — the die needs parallel surfaces
- Avoid undercuts — the gear must eject from the die axially
- Include a hub or boss if possible (eliminates a separate part)
- Specify the density, not just the material — the strength comes from the density
The hidden win: part consolidation
The biggest PM advantage isn’t the gear itself — it’s that the pump’s drive gear, in the machined version, was a gear pressed onto a shaft with a keyway. The PM version was designed as one part: gear with an integral shaft stub and an internal bore, molded as a single piece. The assembly step (press, key, snap ring) disappeared. The pump had one fewer part, one fewer tolerance stack, and one fewer failure mode. The cost savings were 40% on the gear and 15% on the assembly. That’s the PM story — it’s not just “cheaper gear,” it’s “fewer parts.”
PM gears win when the volume is high, the load is steady, and the tolerance is DIN 7 or looser. The pump gear at 50,000 pieces a year, 40% load, and 0.04 mm backlash was a textbook PM application. The die cost is real, but so is the 40% savings and the one-part gear-and-shaft consolidation. Don’t put PM gears where torque shocks hit the root. Do put them where the duty is steady and the quantity is high.
