Congress Gave Them Six Months to Standardise the Screw Thread

The document is the Progress Report of the National Screw Thread Commission, approved 19 June 1920 and issued 4 January 1921 as Miscellaneous Publications No. 42 of the Bureau of Standards. It is a United States government publication, so it is in the public domain and a full scan is on the Internet Archive. Everything quoted here is from that scan.

The commission had nine members. The Director of the Bureau of Standards chaired it. Two came from the Army, two from the Navy, and four were appointed by the Secretary of Commerce, two of those nominated by the American Society of Mechanical Engineers and two by the Society of Automotive Engineers. Before the first six months were up, it became apparent that it would be impossible to complete in a satisfactory manner the work outlined, and Congress amended the act to extend them.

Why 1918

Appendix 1 gives the reason without decoration. The need had been growing with machine tools, automobiles, agricultural implements, typewriters and sewing machines, and then the difficulties encountered in the program for procuring munitions in the recent war demonstrated the vital necessity of standardized screw threads. Engineering societies, the Bureau of Standards and manufacturers petitioned Congress, and the act followed.

The same appendix opens by naming its predecessors: Joseph Whitworth in 1841, and William Sellers in 1864, who promulgated through the Franklin Institute the thread that became the United States Standard series. That is the paper this site has already read, in sixty degrees was easier to gauge. The 1921 report is the sequel to it, written by a federal commission with a deadline instead of by one manufacturer with a gauge.

The English predecessor it names has been read too: Whitworth took fifty-five degrees as the measured average of what English workshops were already cutting, which makes three foundational documents that pick a geometry by surveying the trade.

What was finished

The report defines a terminology and a set of symbols, fixes a thread form, and establishes thread series including a national coarse-thread series and a national fine-thread series, together with pipe and hose coupling threads. Those two names appear in this document in 1921. This page does not trace them forward into the later Unified revisions, which we have not read.

It also establishes four distinct classes of screw-thread fits, with subdivisions, and is careful about how they should be used. The examples given under each class are for illustration only, and it is not the intention of the commission to arbitrarily place a general class or grade of work in a specific class of fit. Each manufacturer and user is free to select the class best adapted to his particular needs.

The class that got a name and no numbers

Class IV is the one worth stopping on, because it barely exists. It covers threaded parts a quarter of an inch in diameter or larger which are to be set or assembled permanently with a wrench. Studs, in other words. Because the material matters so much to how such a joint behaves, it was split into two subdivisions differing in the amount of the allowance, which the report writes as interference: subdivision A for light sections at moderate stresses such as aircraft and automobile engine work, subdivision B for heavy sections at heavy stresses such as steam engine and heavy hydraulic work.

The nut tolerance runs plus from basic, the screw tolerance runs minus from the maximum screw, and the screw sits above basic by the amount of the interference. Everything is in place. Then comes the sixth paragraph of the clause, under its own heading:

A class of fit, defined, named, split into two subdivisions, given the direction of every tolerance, and then handed to the reader with no numbers in it. It is an unusually honest piece of drafting. It was not dropped from later editions either: by the 1928 revision it has its numbers, as class 5, in an appendix, still marked tentative.

The tolerances were sized to the taps that already existed

One heading in the tolerance clause says the whole thing by itself: Class I and Class II Tolerances Permit the Use of Commercial Taps. The tolerances established for the loose fit and the medium fit permit the use of commercial taps now obtainable from various manufacturers. For the close fit, where the hole is wanted near basic size, the report recommends a selected tap.

Read that next to 1864. Sellers set the thread angle at sixty degrees because sixty was easier to obtain than fifty-five and closer to American practice, and he brought a gauge to the meeting. Fifty-six years later a federal commission set the tolerance bands so that the taps already on the shelves would produce work inside them. Both documents choose a geometry by asking what the shops could already make. The numbers on a modern drawing inherit that, whether or not anyone specifying them knows it.

The Go gauge rule, written out in 1920

The report is also where the gauging principle gets stated plainly. The tolerance limits established represent, in reality, the sizes of the Go and Not Go master gages. Errors in lead and angle occurring on the work can be offset by a suitable alteration of the pitch diameter. If the Go gauge passes the work, interchangeability is secured, and the profile may differ from the gauge in pitch diameter, lead or angle. The Not Go gauge checks pitch diameter only, ensuring the fit will not be too loose.

Two smaller rules from the same clause are worth carrying. The pitch diameter tolerance on a screw is the same as on the nut of the same class. And the tolerance on the major diameter of a screw is twice the pitch diameter tolerance for its class.

The clause on specifications, a few pages earlier, has the sentence that a buyer will recognise: qualities should be stated in definite terms of known measurable standards, defined by the largest tolerance limits compatible with satisfactory operation, and every specification should be so concise that no dispute regarding the limiting lines of acceptance can arise. That is the same problem this site keeps meeting from the other end, in standards that name a condition with no referee test for it.

The delegation, and the last paragraph

Section VIII is about what was left undone, and it is the part that reads differently a century later. In July 1919 the commission sent a delegation to Europe to confer with British and French standards organisations. The report records the outcome in one clause: no definite agreements were reached on international standardisation of screw threads, though the engineers and manufacturers in both countries were anxious to cooperate.

The commission still thought it could be done, and its final recommendation before the chairman’s signature says how. An international standard should be established considering the predominating sizes and standards in use, and the possibilities of providing a means for producing this international screw thread by the use of either the English or the metric system of measurement. Signed by the Director of the Bureau of Standards as chairman, Washington, 26 June 1920.

The list of unfinished work in the same section is also worth reading as a list. Threads still needing standardisation: brass tubing threads, instrument threads, Acme, square and buttress forms. Related things that would help American manufacturers: taps, dies, bar stock sizes for cut threads and for rolled threads separately, the dimensions of bolt heads and nuts, sheet metal and wire gauge sizes, and tap drill sizes. Two of those are pages on this site, a hundred years later: what decides the spanner size and what a sheet metal gauge number actually measures.

What a reader can take from it

• A class of fit is a commercial decision as much as a geometric one. Two of the four were sized so that taps already for sale would work
• The Go gauge has absorbed lead and angle error since 1920. A thread that passes Go may differ from the gauge in profile, and that was the stated intent rather than a loophole
• A standard can define something and admit it has no numbers for it. Class IV was published that way, with the gap labelled in its own heading
• The metric question was open, and the answer was to build for both. The commission’s closing recommendation was a thread producible in either system, and that recommendation is the part that did not happen

The preface is the line to end on. Congress extended the commission for two years, recognizing the impossibility of bringing out a report of this character which in the first issue is entirely free from error or inconsistency. Criticisms were to be addressed to the commission at the Bureau of Standards. A standards document that opens by saying it is probably wrong somewhere is a rarer object than a wrench fit.

This page covers step 2, the thread. 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

What was the National Screw Thread Commission?

A body of nine people created by an act of Congress approved 18 July 1918 to ascertain and establish standards for screw threads for the Army, the Navy and the public. It was chaired by the Director of the Bureau of Standards, with two members from the Army, two from the Navy, and four appointed by the Secretary of Commerce from nominations by the American Society of Mechanical Engineers and the Society of Automotive Engineers.

How long was the commission meant to last?

Six months. Section 6 of the 1918 act says the commission shall cease and terminate at the end of six months from the date of its appointment. Before that term expired it became apparent that the work could not be completed satisfactorily, and Congress amended the act to extend it. The preface to the 1921 report records a further extension of two years.

Why was it created in 1918?

The report says the need had been growing with machine tools, automobiles, agricultural implements, typewriters and sewing machines, and that the difficulties encountered in the program for procuring munitions in the recent war demonstrated the vital necessity of standardized screw threads. Engineering societies, the Bureau of Standards and manufacturers petitioned Congress, and the act followed.

What was a Class IV wrench fit?

A class of screw thread fit for parts a quarter of an inch in diameter or larger that are to be set or assembled permanently with a wrench, with a negative allowance the report calls interference. It had two subdivisions, A for light sections at moderate stresses such as aircraft and automobile engine work and B for heavy sections at heavy stresses such as steam engine and heavy hydraulic work. The report states that the commission did not have sufficient information or data to include values for tolerances and allowances for wrench fits, so the class was published without numbers.

Why were the 1920 thread tolerances chosen?

For Class I and Class II, so that existing tooling would work. The report has a heading reading that the Class I and Class II tolerances permit the use of commercial taps now obtainable from various manufacturers. For Class III, close fit, where a hole close to basic size is wanted, it recommends a selected tap instead.

Did the commission try to agree an international thread?

It sent a delegation to Europe in July 1919 to confer with British and French standards organisations, and the report states that no definite agreements were reached on international standardisation of screw threads. Its closing recommendation was that any international standard should be producible using either the English or the metric system of measurement.

What does the Go gauge actually check?

On this report’s reasoning, the fit as a whole. It states that the tolerance limits represent the sizes of the Go and Not Go master gauges, that errors in lead and angle on the work can be offset by a suitable alteration of the pitch diameter, and that if the Go gauge passes the work, interchangeability is secured even though the profile may differ from the gauge in pitch diameter, lead or angle. The Not Go gauge checks pitch diameter only.

References

• Progress Report of the National Screw Thread Commission, authorized by Congress 18 July 1918, as approved 19 June 1920; issued 4 January 1921 as Miscellaneous Publications of the Bureau of Standards No. 42. Full scan and text on the Internet Archive
• Sellers, 1864, read from the Journal of the Franklin Institute, the paper this report names as its American predecessor

The report is a United States government publication and is in the public domain. It was read from the scanned text of the Internet Archive copy with identifier progressreportof42nati, which is not access restricted. Quotations are transcribed as printed, including the typographical slip in section 6 of the act. No value from any of the tolerance tables is quoted on this page. The table pages in this scan are badly garbled by optical character recognition while the prose is clean, so only the prose has been used, and the tolerance rules are described by direction and structure rather than by number. Later editions of the report exist, from 1925, 1929 and 1933, and none of them was read; nothing here describes what they changed. The 1948 Unified thread agreement is outside this page entirely. The names national coarse-thread series and national fine-thread series appear in this 1921 document, and this page does not trace them forward into the later Unified revisions, which would require reading those. The six month term is reported as what the act said; the report does not attribute any quality of its own content to that deadline, and neither does this page.