Two Patents Made the Cross Recess and Neither One Mentions Cam-Out

Phillips does have his own patent on the recess. It is US 2,046,343, filed 3 July 1934 and granted 7 July 1936, and its header reads “Henry F. Phillips, Portland, Oreg., assignor, by mesne assignments, to Phillips Screw Company, Wilmington, Del.” Both documents were read in full for this page. Between them they run to six printed pages, five claims, and a great deal of argument about punching.

The 1933 patent is an argument about manufacturing cost

The first paragraph states the object: a screw with an aperture that lets a single screw driving tool drive screws of various sizes. The second gives the reason the walls converge outward to a knife edge, which is that this makes available the use of automatic machinery to produce screws in quantities at reasonable cost. From there it goes after the alternatives.

Cruciform apertures had been proposed before and had been impractical. Several proposed types required casting, which is so expensive as to make manufacturing costs prohibitive. Others required broaching, and broaching pushes the metal ahead of the tool and so disturbs the distribution of metal as to render the screw head extremely fragile. Against that, the patent sets the existing trade: ordinary screws are cut, rolled or stamped by automatic machinery and cost but a few cents per hundred to make.

So the aim, stated in the document, is to permit the use of automatic machinery by altering the proposed cruciform aperture so that the usual machinery can be used with slight alterations, and so that the cost of manufacture rises only to an inappreciable extent. The listed benefits are that the head is stronger near its outer edge, that a smaller opening is needed, and that the product has a neater appearance.

The negative finding, stated plainly

The word cam does not appear in US 1,908,080. No sentence in it describes the driver rising out of the recess, releasing, slipping or limiting torque. The four claims are geometry and nothing else: wings, side walls converging radially to knife edges, converging downwardly to a common point on the axis of the screw. Torque is not a term in any of them.

That matters in both directions. It means the popular story that the recess was designed to slip has no support in the founding document. It also means our own earlier sentence, that the patent claimed no tendency to cam out, was equally unsupported. A patent that does not discuss a behaviour cannot be cited on either side of an argument about it.

The same silence turns up on the tool side. Within the publicly readable part of the ISO standard that dimensions the driver tip, the word cam does not appear either, and there is no reference to a torque at which the tip is meant to leave the recess.

The sentence people are probably thinking of is from 1936

Phillips’s own patent contains a passage that is close to the thing the cam-out debate keeps reaching for. Among the objects of the invention:

A positive lock and stabilizer under any load conditions is not the language of a deliberate slipping mechanism. If anything in either document speaks to design intent about the driver staying put, this is it, and it is worth knowing that it exists.

It is also worth knowing what it is. This is an objects-of-the-invention paragraph, the part of a patent where the applicant says what the thing is for. It is not a claim, it is not a measurement, and it is not a statement that the recess will hold at any torque. The single claim in US 2,046,343 describes grooves, flat bottom walls, parallel corner edges and their angles. It says nothing about load either. Quoting the objects paragraph as though the patent had proved something about behaviour under torque would be the same mistake we made the first time, pointed the other way.

Where the angles came from

The 1936 patent does explain how the recess geometry was arrived at, and the answer is not what a reader expects. The angular relation of all the walls to each other, and particularly to the depth of the recess, has been carefully calculated and determined in tests especially conducted to analyze the action of punches of various angular shapes in producing recesses of this particular type.

The tests named in the document are punch tests. The stated variable is the angular shape of the punch, and the stated output is a recess that can be produced by a simple punching operation with proper and equal displacement of the metal. The flank angle that a modern reader knows as the source of cam-out is described here as the outcome of an investigation into forming, and the patent offers no separate account of choosing it for how the driver behaves.

One more consequence of punching is claimed, and it is a mechanical one. The crowding and displacing action of the punch compresses part of the head material, which engenders a hardening of the metal, and the patent calls this an important feature because such hardening produces a head of much greater strength than heretofore possible. Work hardening of the recess wall, claimed as a benefit, in 1936.

How to read this kind of source

Patents have two registers and they are easy to mix up. The objects and description are persuasion, written to explain why the invention is worth granting and often worth buying. The claims are the legal boundary, and they are usually narrower and duller than the prose above them. A story about a famous product will almost always be built out of the persuasion, because that is the readable part.

• Check whose name is on it. Assignment is normal and says nothing bad about anyone, but the inventor and the assignee are different roles and popular retellings drop the distinction
• Search the actual text for the actual word before repeating that a document says something. Two of these three pages exist because we did not
• Separate a claim from an object. The claims in both patents are pure geometry, so neither patent makes any legal assertion about torque
• Treat silence as silence. Not mentioning cam-out is not the same as denying it, and it is not the same as intending it

Our page on what actually happens in a cross recess under torque has been corrected to match what these documents say. The mechanics on that page, the geometry, the down-force and the difference from the JIS recess, are unaffected, because they never depended on the patents in the first place.

This page covers step 4, the drive. The whole order is substrate, thread, head, drive, finish, documentation, and why doing it out of order is rework rather than a tweak is in specifying a screw.

Common questions

Who invented the Phillips screw?

The 1933 patent normally cited for the recess, US 1,908,080, names John P. Thompson of Portland, Oregon as the inventor, and records that it was assigned by direct and mesne assignments to H. F. Phillips of Portland, Oregon. Henry Phillips filed his own patent on an improved recess on 3 July 1934, granted as US 2,046,343 on 7 July 1936 and assigned to the Phillips Screw Company. Both documents are readable as scans from the USPTO.

Does the Phillips patent say the screw was designed to cam out?

No. The word cam does not appear in US 1,908,080 or in US 2,046,343. Neither document describes the driver rising out of the recess, slipping or limiting torque, and the claims in both are purely geometric. The 1933 patent argues about manufacturing cost, about casting and broaching, and about head strength.

Does the patent say the screw was designed not to cam out?

It does not say that either, and this page previously said it did. The closest passage is in the 1936 Phillips patent, which lists among its objects the provision of grooves and side walls affording maximum bearing surfaces and self-centering, acting as a positive lock and stabilizer between screw and driver during all driving operations and under any load conditions. That is an objects-of-the-invention statement rather than a claim or a measurement.

Why was the cross recess invented?

Read from the 1933 document, for three reasons it states itself. One driving tool could drive screws of many sizes because all the apertures share substantially the same taper. The winged aperture could be made on ordinary automatic machinery with slight alterations, where earlier cruciform proposals needed casting or broaching. And punching displaced far less metal than broaching a square-ended aperture, which the patent says had rendered screw heads extremely fragile.

How were the angles of the Phillips recess chosen?

US 2,046,343 says the angular relation of all the walls to each other, and particularly to the depth of the recess, was carefully calculated and determined in tests especially conducted to analyze the action of punches of various angular shapes in producing recesses of this type. The variable under test was the punch. The patent gives no separate account of choosing the angles for the way the driver behaves.

Does the patent mention hardening of the screw head?

The 1936 patent does. It states that the crowding or displacing action of the punch, with the resultant compressing of portions of the screw head material, engenders a hardening of the metal, and calls this an important feature of the invention because it produces a head of much greater strength than heretofore possible.

References

• US 1,908,080, “Screw”, John P. Thompson, assignor to H. F. Phillips. Filed 20 May 1932, Serial No. 612,555; granted 9 May 1933. Full specification and four claims
• US 2,046,343, “Screw”, Henry F. Phillips, assignor to Phillips Screw Company. Filed 3 July 1934, Serial No. 733,623; granted 7 July 1936. Full specification and one claim

Both patents were downloaded as PDF facsimiles from the USPTO full-text and image service. Neither has a text layer, so both were rendered to images at 200 dots per inch and read from the page. Every quotation above is transcribed from those scans, and the Google Patents links are given because they carry the same facsimile in a stable public location. The absence of the word cam is a negative finding from reading the six printed pages of the two documents; it is not a machine search of a text layer, because there is no text layer to search. The co-pending application on screw drivers referred to in the opening of US 2,046,343 was not read for this page and nothing is quoted from it. This page does not assign credit for the invention between the named inventor and the named assignee, and does not compare either document with any later patent. Our own earlier statement that the 1933 application claimed no tendency of the driver to cam out came from an encyclopedia entry rather than the patent, is not supported by the document, and has been removed.