Industries / Industrial equipment and services / Industrial machinery / Compressors
Screw compressors
Two meshing rotors with spiral lobes trap gas and squeeze it as it moves along them; this is the usual machine for factory air.
6
Milestones achieved, from 1878 to 2006.
2
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1878
A screw blower, never built
On 24 March 1878 Heinrich Krigar of Hannover was granted German patent 4121 for a "Schraubengebläse", or screw blower, with two helical lobes on each rotor. It could make less than 2 psi (0.14 bar), and the machining of the time could not produce it commercially, Sullair writes.
LessonThe idea waited more than 50 years for rotors that could be cut accurately enough.
Credit
Heinrich Krigar
1935
Lysholm's screw compressor
Alf Lysholm, chief engineer of AB Ljungströms Ångturbin in Stockholm, developed the screw compressor while looking for a compressor for gas turbines, and the firm patented it in 1935. Howden of Glasgow became the first company to license and sell it, in 1938. The first rotors were cut with shapers, fly cutters, end mills and files, a rotor rebuilder notes. SRM, as the firm was renamed in 1951, tested 25 rotor combinations between 1947 and 1950 and licensed the design widely.
Credit
Alf Lysholm and Ljungströms Ångturbin (later Svenska Rotor Maskiner, SRM); Howden
1954
Oil floods the rotors
Early screw compressors ran dry, with timing gears to keep the rotors apart. Spraying oil into the rotors cools the gas and seals the gaps, which allows higher pressure ratios, less noise and less rotor wear. Sources disagree on when this came: Wikipedia says Howden and SRM developed the first oil-flooded screw compressor in 1954 and Atlas Copco sold the first one for air in 1957; Sullair says the mid-1950s; Howden's history in CompressorTECH² says the 1960s. Atlas Copco says it bought the rights to Lysholm's compressor in 1954.
en.wikipedia.orgamerica.sullair.comcompressortech2.comatlascopcogroup.com
Credit
SRM; Howden; Atlas Copco
1970
An asymmetric rotor profile
The first screw rotors had symmetric lobes. SRM's patents of 1946 (symmetric), 1970 (asymmetric) and 1982 (its "D" profile) became the standard references for rotor shapes, wrote N. Stosic, a professor at City University, London, and SRM claims 750 patents. An asymmetric lobe seals better, which raised efficiency. Sullair, which signed its first license with SRM on 14 July 1965, brought asymmetric rotors to market in 1969, by Wikipedia's account.
Credit
SRM; Sullair
1999
Screws take the market
By the late 1990s screw compressors were 80% of all positive displacement compressors sold and 50% of those in operation, Stosic wrote, and they covered almost all of air compression. Precise milling and grinding of rotors held tolerances below 10 µm and gaps between the lobes to 30 to 50 µm at an economic cost, so internal leaks fell to a small fraction of earlier designs. He named Holroyd of Britain as the largest maker of screw rotors.
Credit
Holroyd and other rotor makers; SRM and its licensees
2006
Oil-free air, certified
ISO 8573-1 sets purity classes for compressed air, and Class 0 is the strictest of its six. In 2006 TÜV, the German testing association, certified Atlas Copco's Z series oil-free screw compressors to Class 0 after finding no trace of oil under any of its test conditions; Atlas Copco called it a world first. Atlas Copco says it launched the stationary Z series in the 1960s, opening compressed air to textile, food and medicine work.
Credit
Atlas Copco; TÜV
Watch for
Water-cooled single screw
Green-Y says its single-screw airend, lubricated, sealed and cooled with water instead of oil, is up to 20% more efficient than conventional oil-free machines. It says its 40 hp unit needs less than 22 hp per 100 cubic feet a minute at 100 psi (7 bar), measured to ISO 1217, and that the technology has run more than 10,000 hours in industry, with its first unit at a Swiss maker of pneumatic and automation equipment.
Achieved whenan outside lab confirms the energy figure on a production unit.
Claimed by
Green-Y, July 2026
India's water-injected screw
Kirloskar Pneumatic says its HYDRINO, launched in Bengaluru on 25 February 2026, is India's first oil-free screw compressor that uses water for cooling, sealing and lubrication, and that it meets ISO 8573 Class 0. It was developed with a government grant and mentoring from the Pravriddhi program at the Indian Institute of Science.
Achieved whenan independent test body certifies the Class 0 rating and a named customer reports the unit in service.
Claimed by
Kirloskar Pneumatic, February 2026
Open
01
Measured screw rotor profiles
A public library of 3D scans of rotor pairs from 20 retired screw airends (the rotor and housing unit) of different makes and sizes, bought by the team for research, with the cross-section profile extracted and the main design numbers computed by open code: lobe counts, rotor diameters, center distance, wrap angle, volume per turn, length of the sealing line and size of the leak gap at the lobe tips.
The briefCredit · unclaimed02
An airend wear atlas
A labeled public set of photographs of torn-down screw airends (rotors, bearings and housings), each tagged with running hours, oil type and the rebuilder's finding of cause, plus a baseline model that suggests the cause from the photos.
The briefCredit · unclaimed
Open 01 · the brief
Measured screw rotor profiles
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public library of 3D scans of rotor pairs from 20 retired screw airends (the rotor and housing unit) of different makes and sizes, bought by the team for research, with the cross-section profile extracted and the main design numbers computed by open code: lobe counts, rotor diameters, center distance, wrap angle, volume per turn, length of the sealing line and size of the leak gap at the lobe tips.
Why it is still open
Rotor shapes are patented and closed. Stosic wrote that public knowledge of screw compressors is limited and that "a lack of comparative experimental verification" holds back checks of design models. Design tools are commercial: one open Python framework on GitHub drives the commercial programs SCORG and GT-SUITE, and PFYA, a Russian-language calculation service for dry screw machines that went online in October 2026, says its results are calculated and not yet checked against measurements. Searches of GitHub and Zenodo found no set of measured profiles.
Built with what exists
Retired airends from rebuild shops; a structured-light 3D scanner or a university's coordinate measuring machine; open mesh and CAD tools such as Open3D and FreeCAD; Python for the profile math.
First step
Buy one retired airend from a rebuilder, scan both rotors, and publish the extracted profile with the code.
Done looks like
20 rotor pairs published with scans, extracted profiles and computed numbers; for three of them, the computed volume per turn is within 3% of the maker's published figure; one pair scanned on a second scanner agrees within 50 µm.
Open 02 · the brief
An airend wear atlas
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A labeled public set of photographs of torn-down screw airends (rotors, bearings and housings), each tagged with running hours, oil type and the rebuilder's finding of cause, plus a baseline model that suggests the cause from the photos.
Why it is still open
Rebuilders post a few example photos; one says that "by far the most common cause" of early airend failure is contaminated or insufficient lubricant. No labeled image set of airend failures was found on GitHub or Zenodo or in a web search. The nearest public work uses vibration, such as a 2022 GitHub project that classifies faults in an oil-injected screw compressor from vibration spectrograms.
Built with what exists
Phone cameras and a light box; rebuild shops' teardown reports; a pretrained image model fine-tuned with PyTorch.
First step
Ask one rebuilder to photograph the next ten airends it tears down, following a fixed list of shots, and to share its findings sheet for each.
Done looks like
600 photos from 60 airends supplied by at least three rebuilders, each with hours and cause; two rebuilders agree on the cause for at least 85% of airends; the model names the agreed cause for at least 75% of airends from a shop it was not trained on.