A European industrial pump manufacturer converted a billet-machined housing to A360 die casting to reach 8,000 units a year, holding 6 bar pressure tightness, an H7 bearing bore and a 0.05 mm sealing face.
The housing carries a water-glycol circuit at 6 bar working pressure. Machining the bearing bore and the cover sealing face opens the casting interior, so microporosity that would have been harmless below the skin becomes a potential leak path. A single interconnected path scraps the part after all machining value has been added.
The Ø52 bore takes a pressed bearing at H7 (+0.030/0 mm), and the cover face needs 0.05 mm flatness for its elastomer gasket, with 0.04 mm perpendicularity between the two. As-cast capability is ±0.1 mm and 0.1 mm flatness per 100 mm, so neither feature can come off the die.
The incumbent route machined the housing from 6082 billet at roughly $41 per part, with 22 minutes of spindle time and about two thirds of the purchased aluminum becoming chips. At the planned volume the machining line, not the price, had become the constraint on the customer's build plan.
The pump sets are skid mounted in coastal plants and food processing facilities, washed down regularly, and fitted with stainless fasteners. The customer's specification called for 2,000 hours neutral salt spray on the coated exterior with no blistering at edges or fastener interfaces.
A380 is the lower-cost alloy and machines better, but A360 holds copper below 0.6% against A380's 3.0 to 4.0%, which materially improves corrosion performance, and it is rated excellent rather than good for pressure tightness. On a wet 6 bar housing that difference is worth the alloy premium. Its 317 MPa tensile strength and 3.5% elongation gave adequate margin at the mounting flange.
Flow and solidification simulation placed a three-point fan gate on the mounting flange side so the fill front sweeps away from the bore, with two overflows positioned to catch the gas-rich leading metal in material that is later trimmed off. Squeeze pins fed the two heavy bosses. The last-to-freeze region sits outside every machined face on the part.
Cast on a 630 ton cold chamber machine at a 2.31 kg shot weight and a 74 second cycle. Die temperature held at 210 to 240 °C with conformal cooling routed close to the bore core, and intensification at 85 MPa held for 6.5 seconds to feed solidification shrinkage in the flange section.
Three cast pads raised 1.2 mm, all in the fixed die half so they share one shrink history and never cross the parting line, carry the first operation. A single setup on a 4-axis horizontal centre bores the Ø52 H7 with a PCD-tipped bar, fly cuts the sealing face for flatness and drills and taps the eight M8 holes. Allowance is 1.2 mm on bore diameter and a deliberate 0.5 mm on the sealing face to keep the cut inside the dense chill layer.
Every housing is air-decay tested at 8 bar, 1.33 times working pressure, on a fixture that seals the same faces the assembly does, with a 0.5 cm³/min reject threshold. Parts that fail go to vacuum resin impregnation and are retested. That route was qualified and approved with the customer before production release rather than negotiated after a failure.
The customer builds circulation pump sets for industrial process cooling. The housing had been machined from 6082 billet since launch, which suited the first few hundred units a year and stopped suiting anything once the product line found its market. When they approached us they were at roughly 3,000 units annually with a plan to reach 8,000, and their machining cell could not absorb that without a second shift and another machining centre.
Two things made this a straightforward casting candidate: the geometry was already close to castable, and the volume sat well clear of the 1,000 to 2,000 unit threshold where tooling starts to pay for itself. One thing made it interesting. The part had to hold pressure after being machined, which is where pump housing conversions succeed or fail.
A380 would have been the cheaper and easier answer. It has the best castability and machinability of the common die casting alloys and the lowest cost, and on a dry bracket it would have been the obvious specification.
Two properties ruled it out. A380 carries 3.0 to 4.0% copper against A360’s 0.6% maximum, which is the main reason its corrosion resistance is rated fair rather than good, and this housing lives in coastal and washdown environments under a coating that will eventually be scratched. And A360 is rated excellent for pressure tightness against A380’s good, reflecting its better feeding behaviour and lower tendency to interconnected microporosity. On a part where a leak means scrap after all the machining value has been added, that is worth more than the alloy premium.
At 317 MPa tensile strength and 3.5% elongation A360 also gave enough margin at the mounting flange, the only meaningfully loaded feature.
This was the engineering work that mattered. Machining removes the dense chill layer, roughly 0.3 to 0.5 mm of the best material in the casting, exposing whatever lies beneath. Gas and shrinkage porosity concentrate in the regions that freeze last, so the problem reduces to keeping those regions out from under the bore and the sealing face.
Once a gate is cut in steel that geometry is fixed, so it had to be settled before tooling. Flow and solidification simulation went through four gating iterations. The accepted design puts a three-point fan gate on the mounting flange side, sweeping the fill front away from the bore and into two overflows sized to catch the leading, gas-rich metal, with squeeze pins feeding the two heavy bosses. Sectioned first articles confirmed the model: every machined face on the part sits over material that solidified early.
The machining allowance was set to match. The bore takes 1.2 mm on diameter, enough to remove the cast draft and taper of the cored hole. The sealing face takes a deliberately minimal 0.5 mm so the cut stays inside the chill layer, which is the standard approach on any face that has to remain pressure tight.
A casting arrives with no flat, square reference anywhere on it, so the datum scheme was defined on the part drawing before the die was designed. Three cast pads raised 1.2 mm establish the primary plane, and all three sit in the fixed die half. Nothing in the reference frame crosses the parting line, which matters because parting line mismatch of 0.1 to 0.15 mm is normal and would otherwise appear as fixture error on every part.
The bore and the sealing face are machined in one setup on a 4-axis horizontal centre, so their perpendicularity comes from the machine rather than from fixture repeatability, which is worth about 0.05 mm on its own. A single-insert fly cut produces the sealing face, removing insert-to-insert height variation from the flatness result.
Every housing is air-decay tested at 8 bar against a 0.5 cm³/min threshold, on a fixture that seals the same faces the finished assembly does. Testing at 1.33 times working pressure is deliberate, giving margin against a marginal path that would open later in service.
The more useful decision was agreeing the failure route in advance. Parts that fail go to vacuum resin impregnation, which seals interconnected microporosity without affecting dimensions, and are retested. Qualifying that with the customer before production release meant the first non-conforming part was a routine rework rather than a program discussion. First-pass yield settled at 96.4%.
Piece cost came in at $12.40 against roughly $41 machined, a 70% reduction and about $229,000 a year at 8,000 units. Machining dropped from 22 minutes to 4.2 minutes per part, releasing the capacity the customer had been planning to buy.
Across the first twelve months and 8,140 units shipped there were no leak escapes, and the Ø52 H7 bore held Cpk 1.67 over thirty consecutive lots. The coated exterior was qualified at 2,000 hours neutral salt spray on a duplex epoxy primer and super-durable polyester topcoat over trivalent chromate. That is well above the 500 to 1,000 hours a standard single-coat powder system delivers, and it was panel-qualified for this program specifically rather than assumed.
The part still runs on the original single-cavity tool. At 8,000 units a year a second cavity does not pay for itself, and that arithmetic is worth revisiting only if the volume roughly doubles.
| Part weight | 1.42 kg (3.13 lb) |
|---|---|
| Envelope | 186 × 142 × 98 mm (7.32 × 5.59 × 3.86 in) |
| Alloy | A360 (AlSi10Mg), 317 MPa UTS, 113 W/m·K |
| Wall thickness | 4.5 mm nominal, 3.0 mm minimum |
| Machine | 630 ton cold chamber, 2.31 kg shot weight |
| Cavity count | Single cavity, one slide for the discharge port |
| Annual volume | 8,000 units |
| Critical tolerance | Ø52 H7 bore (+0.030/0 mm), 0.05 mm sealing face flatness |
| Finish | Chromate conversion plus duplex powder coat, RAL 7016 |
| Piece cost | $12.40, down from about $41 machined (−70%) |
|---|---|
| Annual cost reduction | About $229,000 at 8,000 units |
| Leak escapes, first 12 months | 0 across 8,140 units shipped |
| Leak test first-pass yield | 96.4%, remainder passed after impregnation |
| Bearing bore capability | Cpk 1.67 on Ø52 H7 over 30 consecutive lots |
| Coated exterior | 2,000 h ASTM B117 on the qualified duplex system |
Send your drawing and we will tell you what this geometry costs in your annual volume.
Request A Quote
A collaborative robot manufacturer moved a joint housing from machined prototypes to ADC12 die casting at 12,000 units a year, cutting 180 g per joint while holding ±0.02 mm concentricity between the harmonic drive seat and the encoder bore.
An LED lighting brand replaced a bonded two-piece design, an extruded heat sink adhered to a die cast housing, with a single A413 casting, removing the thermal interface and taking 11 °C off the LED junction temperature at 60,000 units a year.
Send your 2D drawing and 3D model and a manufacturing engineer will review the geometry, confirm the alloy and return pricing with a DFM report. No obligation, no sales scripts.