Investment Casting Robot: Lost Wax to ABB

An investment casting robot takes over the slowest and messiest job in a precision foundry. It dips clusters of wax patterns into ceramic slurry, rolls them in sand and hangs them to dry, coat after coat, until a hard shell forms around the wax. This post follows that job from lost wax casting in the ancient world, through the dental labs and wartime turbine blades that turned it into an industry, to robot shell building cells with ABB IRB 6600 and IRB 7600 arms. The last part covers the jet engine boom that has foundries scrambling to build more shells.

Molten steel poured into a ceramic investment casting shell, filling the space the wax patterns left. Photo: 5oclockhustle / Wikimedia Commons (CC BY-SA 4.0)

How lost wax casting became an industry

Lost wax casting is about 5,000 years old. A craftsman shapes a model in wax, covers it in clay, heats the mould so the wax runs out, and pours metal into the empty space. Objects from the Cave of the Treasure at Nahal Mishmar, in the Judean Desert, have been carbon dated to around 3700 BC. Much later, metal workers in the Kingdom of Benin cast detailed brass plaques the same way, and the method has never stopped being used for sculpture and jewellery.

A museum print of the Nahal Mishmar hoard as it was found in the Cave of the Treasure, shown at the Israel Museum. Photo: Chamberi / Wikimedia Commons (CC BY-SA 3.0)

The modern industrial process came from dentistry. Barnabas Frederick Philbrook of Council Bluffs, Iowa, described casting with wax patterns in 1897, and William H. Taggart of Chicago published his work on wax compounds and casting machines in 1907. During the Second World War the aircraft industry needed precise parts in alloys that were hard to machine, so it adopted investment casting for turbine blades and other engine parts. The ceramic shell replaced the old clay: a cluster of wax patterns on a wax sprue, called a tree, is coated in layers until the shell is 5 to 15 mm thick.

Benin brass plaques, cast by the lost wax method, at the British Museum. Photo: Joyofmuseums / Wikimedia Commons (CC BY-SA 4.0)

For decades that coating was done by hand. A worker lifted each tree, dipped it in slurry, let it drain, rolled or showered it in sand and hung it on a rack, and each coat made the tree heavier. Drying between coats can take 16 to 48 hours in total. ABB's case study of Franklin Bronze, a Pennsylvania foundry opened in 1878, describes six workers dipping about 100 moulds a day by hand. The same foundry still had a five axis dipping machine whose technology dated from the 1980s, and TPC Components in Hallstahammar, Sweden, ran its shell line with a Unimate robot for many years.

A casting tree of wax turbine blades for the Walter M601 engine, coated in ceramic, 1979. Photo: Ltosnar / Wikimedia Commons (CC BY 4.0)

Robot shell building and the jet engine rush

Jet engine makers need more castings than suppliers can pour, and the big players are buying capacity. On 8 September 2026 GE Aerospace agreed to buy Consolidated Precision Products (CPP) of Cleveland for $11.75 billion. CPP makes investment and sand castings in superalloys, titanium, aluminium and other metals, employs about 6,600 people at more than 20 sites and has supplied GE for over 15 years. The deal is expected to close in the second half of 2027. On 6 October, Pursuit Aerospace announced it had bought Creasey Castings and was investing to expand casting capacity.

A two piece single crystal turbine blade from NASA's Energy Efficient Engine programme, 1979. Photo: Jack Darginsky, NASA via Wikimedia Commons (public domain)

Shell making is often robotized, ABB notes, and it has published two cases. In 2005 Shell-O-Matic of Montreal installed a shelling system at Franklin Bronze built around an ABB IRB 6600. The robot dips each tree in slurry, coats it with sand and steers the hanger through the drying room on a schedule set by a supervisory computer. Three workers now made about 200 moulds a day, man hours dropped from 56 to 32 a day, and the foundry said shell quality improved because every tree was dipped the same way.

An ABB IRB 6600 in the Shell-O-Matic shelling system at Franklin Bronze, Pennsylvania, with racks of coated wax trees. Photo: Gene Puskar / ABB

At TPC Components, ABB replaced the old Unimate with an IRB 7600 long arm version, with 3.5 m reach and a 150 kg handling capacity, without rebuilding the line. Output rose from 1,400 trees a day to 1,800, with 2,300 expected after tuning. The robot has ABB's Foundry Plus protection, sealed to IP67, because the alcohol based slurry is explosive and must not get inside the arm. Today the IRB 6700 covers the size class of the IRB 6600, and dipping paths can be planned in RobotStudio before a cell is built.

An ABB IRB 7600 holds a wax tree in a falling sand curtain to stucco the shell at TPC Components, Sweden. Photo: ABB

What comes next for robotic investment casting

Jet engine demand means more shells, larger trees and tighter control. Turbine blades with internal cooling channels and single crystal alloys leave little room for a thin spot or a crack in the shell, so foundries want every tree dipped at the same angle and speed. The Franklin Bronze cell already read a barcode on each tree in 2005, picked the matching dipping program and printed a report for every finished shell; that kind of tracking is likely to become standard. Robots also already grind and finish castings after the shell is knocked off.

The ABB robot's gripper lowers a wax tree into a tank of ceramic slurry at TPC Components. Photo: ABB

The rest of the foundry is the next step. ABB's own TPC case notes that some foundries were looking at automating the mounting of wax patterns onto trees, a fiddly job that is still mostly manual. After pouring, someone still has to knock off the shell, cut castings from the tree and inspect them, and those jobs are candidates too. Similar changes have reached sand foundries, as our posts on foundry core making and die casting describe.

Cast turbine blades photographed at NASA's Lewis Research Center in 1979. Photo: Donald Huebler, NASA via National Archives / Wikimedia Commons (public domain)

Cost and skills decide how fast this happens. Many investment casters run short series, so a robot cell has to switch quickly between parts, and the people who know how a shell should look when it drains are hard to replace. For people at home the process shows up in jewellery and art castings, where small studios still dip by hand. In factories, the payback numbers in ABB's cases, a year and a half at TPC and about two and a half years at Franklin Bronze, help explain why so many shell rooms already run with robots.

A 31 mm basset hound carved in blue wax as the master model for lost wax jewellery casting. Photo: Mauro Cateb / Wikimedia Commons (CC BY-SA 3.0)

Innovation

AI solutions for effortless ABB robot control.

Automation

Robotics

ceojohntran@chatwithrobot.net

+84905311611

© 2025. All rights reserved.

qtran1215@gmail.com