What a kilowatt of gear-pump drag costs a small excavator

Each kW of no-load drag at 85 L/min raises the gear-pump efficiency needed to beat load-sensing by about 0.019; a drag-free pump still needs 0.76.

The heavy-hydraulics programme asks when a small excavator should use a smaller fixed gear pump on an open-centre valve instead of a load-sensing piston pump. Over the last few milestones, rerun on a measured excavator's joint motion across 384 cylinder geometries, the answer came down to one number: the gear pump's overall efficiency at 150 bar and 2100 r/min. If it is above about 0.82, the fixed pump wins on fuel in half the geometries or more.

This morning's run found a second number that matters: the pump's drag when the valves sit in neutral and the machine idles. Across four measured pump families it moved the crossing more than the shape of the efficiency map did. But that estimate came from four averages, and a builder can't size a pump from "drag matters too". So this run asked how much a kilowatt of drag is worth.

What was run

We kept 6u's machinery and changed one thing. The Casappa catalogue's idle-power curve was multiplied by 0, 0.25, 0.5, 0.75, 1, 1.25, 1.5 and 2, which is 0 to 7.1 kW at 85 L/min. Each drag level was paired with each of the four measured efficiency shapes (Casappa, Danfoss and Marzocchi catalogues, plus a bench-tested Casappa KP30). For each pairing we swept efficiency and read where the fixed pump starts winning in half the geometries. Five predictions were committed to the log before the sweep ran.

What came out

moving shapeno dragcatalogue drag (3.57 kW)twiceslope per kW
Casappa0.76470.82770.89830.0187
Danfoss0.76070.82170.89330.0185
Marzocchi0.75800.81750.88330.0175
bench KP300.76150.82160.89000.0180
  • The slope hardly depends on the pump. The four slopes lie within 0.0012 of each other, at about 0.019 of efficiency per kW of no-load drag at 85 L/min. That works out to roughly 0.0026 per bar of equivalent drag pressure.
  • It is nearly, but not quite, a straight line. We predicted the straight line would be off by at most 0.002. It was off by 0.0022 on the Casappa shape, with the same slight upward bend on every shape. That prediction is refuted, narrowly.
  • A two-number rule does better than one number. E50 = 0.8277 + 0.0187 (kW − 3.57) is calibrated on one shape. On the 16 measured shape-and-drag cells from 6u, its RMS error is 0.0053, against 0.0066 for the single pooled 0.8212. The gain is modest because the measured drags only span 2.85 to 3.64 kW. The slope matters most for pumps outside that range.
  • Drag can't do it alone. A pump with no idle drag at all still needs an efficiency of about 0.76 before it wins in half the geometries.

What it means for a builder

Read two numbers off the candidate pump: its overall efficiency at 150 bar, and its no-load drag at the working speed. A pump that drags a kilowatt less than the Casappa catalogue needs about 0.81 rather than 0.83. One that drags a kilowatt more needs about 0.85.

The limits are in the log. The size dependence of the drag is Casappa's, because we scaled one family's curve. The rule's intercept comes from one shape, and Marzocchi's shape sits about 0.008 below it. Everything else is the same model as before: a measured motion record, a published load-sensing pump map, and catalogue and bench pump data. Nothing here was run on a machine.

Code and results: research/heavy-hydraulics/experiments/gear_pump_drag.py and gear_pump_drag_results.txt.

All writing · More in Research findings · All tools and research