From Mechanics, January 12, 1884 issue, pages 38-40

Screw Thread Absurdities.

From the Journal of the Amateur Mechanical Society of Great Britain we make the following extracts from a paper by James Edmunds, M. D., a well-known authority upon the subject of ornamental lathes, whose remarks forcibly illustrate the loose practice and absurdities that arise from the absence of proper standards in screw threads:

The screws still used by the makers of ornamental-lathe apparatus exhibit an almost unique survival of what are called “country threads” when they pass into engineers’ shops for repair. In the year 1846 Charles Holtzapffel, in the admirable third volume of “Turning and Mechanical Manipulation,” which he then issued, said that these fractional screw rates ought to be superseded by aliquot rates; but the fourth volume of this work was issued in 1879 without touching the question, and these inconvenient screws are still being used in the manufacture of geometric lathes. It has therefore become necessary for the users of these lathes to unite for the purpose of considering the question, and to say whether still another generation of amateurs shall be allowed to encumber themselves unwittingly with apparatus constructed with these screws.

For the purpose of holding a chuck on to a mandrel nose, or of fixing two objects together, screws of fractional rates are perfectly good. Whether a screw has 36 threads to the inch or 36.1 is of no consequence to the smith or the carpenter who buys screws for mere holding purposes. Nor is it of immediate moment to the amateur who limits his work to artistic turning, and uses only the graduations of the division plate. For a slide-rest screw used as a mere motor, one of 9.45 or of 13.09 threads to the inch is as good as one of 10. But for micrometric regulating purposes—making a piece of philosophical apparatus, cutting a screw of a given pitch, or graduating a surface—such fractional screw is useless, while a 10-rate screw is perfectly good for mere motorial or holding purposes, and also serves for scientific work. Aliquot screw tools cost no more than fractional ones; they are easier to make and to replace, they are simpler to work with, and they have multiform utility. Fractional screws have their utility limited to the mere exhibition of brute force; they have long been superseded among scientific engineers, and amateurs are entitled to the benefit of the reform.

Screws of fractional rate possess no advantage over screws of aliquot rate, while every purpose which they serve can be served equally well, if not better, by aliquot screws. Aliquot threads may be defined exactly in simple terms ; these “ornamental threads” are definable only by reference to a set of arbitrary screw tools, whose measurements have to be given “approximately,” and can only be approached in decimal fractions, mostly of two places. Aliquot threads may easily be understood and verified by their owners. They may at any time be accurately originated by means of a point-tool and a small number of simple change-wheels to co-ordinate the revolutions of the mandrel with those of the slide-rest screw. Such change-wheels at the same time serve for the spiral work of the ornamental lathe, and are already in the hands of amateurs or can be added to their lathes at small cost. Aliquot threads bear a simple ratio to each other, whereas the present ornamental threads are practically incommensurable with each other, and are useless for micrometric and scientific purposes.

The “ornamental screws” have already been superseded in slide-rests by the screw of 10 to the inch, and with vast advantage. But the Holtzapffel back center is still made with the No. 4 screw—i.e., a thread of 13.O9 to the inch. Consider the inconvenience of this. In boring out a chuck or other object the screw of the back center is often needed as the register of the depth bored. If its rate be 10 to the inch, every turn is .1 inch, and the number of turns gives the depth bored in tenths of an inch. If the handle of the screw be marked with a circle divided to hundredths, the depth is automatically shown to .001 inch. But with the present screw, each turn gives a depth of 1-13.09 inch. Now, to get the reciprocal of 13.09, and to multiply that by the number of turns which have been given to the back-center screw, requires the work to be stopped for a complex calculation. What is thus shown of the 13.09 screw is equally true of alJ these screws when other than mere motorial or holdfast service is required. There is not even an adequately defined standard with regard to any one of the elements of these Holtzapffel screws. Their thread rates, as given, are qualified by the term “approximately;” their thread-angles, as given, are qualified by the term “about;” the truncation of their threads is done by rule of thumb; the tools actually issued by the firm are shown on critical measurement to vary among themselves both in thread-rate and in thread-form.

The following table contains a digest of all the data which I can discover to have been published by the Messrs. Holtzapffel in reference to their screws for ornamental lathe work. (The threads of these screws are described as “the deep, having an angle of about 50°,” and “the shallow, having an angle of about 60°.” ):

Comb-screw Tools.
Threads per inch.
Taps and Dies
outside thread
1 6.58 A 1 in. = 1.0000
2 8.25 B 7/8 in. = 0.8750
2 8.25 7-inch center
mandrel nose
9/8 in. = 1.1250
3 1 9.45 6-inch center
mandrel nose
15/16 in. = 0.9375
3 1 9.45 5-inch center
mandrel nose
13/16 in. = 0.8125
3 1 9.45 C ¾ in. = 0.7500
4 2 13.09 4-inch center
mandrel nose
¾ in. = 0.7500
4 2 13.09 DD 5/8 in. = 0.6250
4 2 13.09 D 0.5600
4 2 13.09 E ½ in. = 0.5000
5 3 16.5 F 0.4500
6 4 19.89 G 0.4100
6 4 19.89 H 0.3600
7 22.12  
8 5 25.71 I 0.3300
8 5 25.71 J 0.2900
8 5 25.71 K ¼ in. = 0.2500
9 28.88 M 0.2100
10 6 36.10 L 0.2400
10 6 36.10 N 1/5 in. = 0.2000
10 6 36.10 P 0.1800
11 39.83 O 0.1900
11 39.83 Q 0.1625
12 55.11 R 0.1500
12 55.11 S 0.1350
12 55.11 T 0.1200
12 55.11 U 1/10 in. = 0.1000


Many of the data contained in the above table were elicited for the first time in reply to questions publicly addressed by myself to Mr. J. J. Holtzapffel in the columns of the English Mechanic for February, March and April, 1882. Further correspondence upon these screws will be found in the English Mechanic for February, March, April and May, 1883.

The first column of Table I gives the numerical denotements by which Charles Holtzapffel denoted these threads. These denotements have no relation to the rate of their thread.

The second column gives the numerical denotements by which the mandrel guides are denoted. These denotements, again, have no relation either to the rate of their thread or to the numbers used to denote the same threads when upon other screw tools.

The third column shows the number of threads per inch “approximately” of the tools upon each line.

The fourth column gives certain alphabet denotements by which these threads, when cut upon taps and dies, are denoted.

The fifth column gives the diameters—outside the thread— of the tap screw tools. Their diameters at bottom of thread have never been defined.

If we take into consideration the rates of these ornamental lathe threads, it cannot be contended that a screw of “approximately 55.11 threads to the inch” is better than one of exactly 55; that a screw of “approximately 39.83 threads to the inch” is better than one of exactly 40; that a screw of “approximately 36.10 threads to the inch” is better than one of exactly 36; that a screw of “approximately 28.88 threads to the inch” is better than one of exactly 30; that a screw of “approximately 25.71 threads to the inch "is better than one of exactly 25; that a screw of “approximately 19.89 threads to the inch” is better than one of exactly 20; that a screw of “approximately 13.09 threads to the inch ”is better than one of exactly 13 or 12 ; that a screw of “approximately 9.45 threads to the inch” is better than one of exactly 9 or 10, or 9.5.

The character of the denotements by which the same rates of thread upon differently-shaped tools are distinguished appears as if intended to produce confusion. The comb-screw tools are marked by arbitrary numbers ranging from 1 to 12. The mandrel guides are marked by other arbitrary numbers, the two sets of numbers being different, and each set having no relation to the thread-rates of the tools. Thus, the 9.45 thread is cut by a comb tool branded “3;” by a mandrel guide branded “1;” by a tap tool branded “C.” The 13.09 thread is cut by a comb tool branded “4;” by a mandrel guide branded “2;” by a tap tool branded sometimes “D,” sometimes “E,” and sometimes by another letter— these three alphabet denotements being branded respectively upon taps of the same rate, but of different diameters. The 16.5 thread is cut by a comb tool branded “5;” by a mandrel guide branded “3;” by a tap tool branded “F.” The 19.89 thread is cut by a comb tool branded “6;” by a mandrel guide branded “4;” by a tap tool branded “G” or “H,” according to diameter. The 25.71 thread is cut by a comb tool branded “8;” by a mandrel guide branded “5;" by a tap tool branded “I” or “J ” or “E,” according to diameter. The 36.10 thread is cut by a comb tool branded “10;” by a mandrel guide branded “6;” by a taptool branded “L” or “N” or “P,” according to diameter. In short, the nomenclature of these screws is such as to suggest that their complications have been devised in order to prevent their purchasers from understanding them.

Again, if an amateur, in cleaning his lathe apparatus, loses a screw, he must either obtain its special screw tools or have the apparatus packed up and sent to the Holtzapffel factory in order to have this screw replaced. In case he resides in the country or abroad, the loss or fracture of a Holtzapffel screw is indeed a serious business. In some cases he may drill out the old thread, tap the hole with an aliquot thread and fit in a new bolt. But often such holes may not be drilled out so far as to receive a bolt whose central stem is as large as the outside of the original thread. There remains the alternative of originating the thread by means of change-wheels, and, if the thread be a practicable one, this will at once be done by any working engineer. It is only necessary to arrange a “setting” (changewheels and guide-screw), which contains a multiple or sub-multiple of the rate to be cut— i. e., the primes, and such other factors as will produce its number. The screw thus lost or broken is, say, a “J” screw or an “N” screw. But what is a “J” screw or an “N" screw? This question is one which, until the table herewith was published, no one but Mr. Holtzapffel could answer. Indeed, on February 17, 1882, Mr. Holtzapffel intimated that Mr. Evans “was in complete ignorance of the respective values and dimensions of these taps and dies,” and could obtain them only through him.* Now, if, upon the high authority of Mr. Holtzapffel, this was then the benighted condition of Mr. Evans— who had been manufacturing these screws all his life— what must have been the condition of the clients who bought the screws? Thanks, however, to the table, it may now be seen that a “J” screw is “25.71 threads approximately” to the inch, and .29 inch in diameter— outside the thread; the “N” screw is “36.1 threads approximately to the inch,” and .2 inch in diameter— outside the thread. This being all perfectly clear, the amateur goes with a light heart to the mahogany box of “ornamental” brass change--wheels which was prescribed for his outfit by Mr. Holtzapffel. But he only drifts helplessly into a series of puzzling and unfruitful calculations. For making spirals and gimcracks in ivory and black wood these ornamental wheels answer as well as any others, but for work that requires scientific mensuration other than in combinations of 2, 3, 5 and 53, they are useless. The ordinary engineers’ set of 22 change-wheels contains all the primes below 20, and cuts not only any reasonable thread, but also five of the Holtzapffel threads, if compound trains of wheels be not too much for the operator. Nos. 2, 3, 4, 5 and 6 can thus be cut “approximately.” But the Holtzapffel wheels are so planned as not to cut any of the Holtzapffel threads.

It will be seen that the rates are qualified by the term “approximately,” and that the angles by which the threads are defined are qualified by the term “about.” On the value of these qualificitions I can throw no light. “J. K. P.”— an eminent practical authority upon amateur lathe tools —states that one of his Holtzapffel screwtools of “about 50° ” really measures 45°, and that his “16.5” tool really measures 16.25 threads to the inch. If there be such instances of the gauging of the Holtzapffel screw-tools it is difficult to imagine for what purpose their rates have been set out upon paper to two places of decimals “approximately.”

With the Holtzapffel lathes we get a box of ornamental change-wheels, 15 in number, and having teeth as follows: 15, 16, 18, 20, 24, 36, 48, 50, 53, 60. 60, 72, 96, 120, 144. These wheels, on analysis, are found to contain only the primes 2, 3, 5 and 53. The 53 is useless; 7 and 47 are absent, and, in short, no one of the Holtzapffel screws can be cut by means of the Holtzapffel changewheels. The engineer has a simple set of wheels rising by fives from 20 to 120, and these contain all the primes below 29, but he has no 47. We must then give up the attempt to cut this screw, or we must make the additional wheels. Had the pitch been 6.50 instead of 6.58, the engineer has the primes 2, 5 and 13 ; 100/65 gives the velocity ratio, and 100 on mandrel with 65 on guidescrew would cut the 6.50 thread. The pitch of 6.58 is infinitely inconvenient, and would be spit upon by an intelligent gasfitter; but, just as flint guns are still manufactured as good enough for savages, so these screws are still manufactured as good enough for amateurs. * * * * *

In order that a screw-thread may be enduring under frequent fixing and unfixing— as upon a mandrel nose— the surfaces of the threads must take an extensive bearing against each other. Now, the Whitworth thread section contains a double curve, which, even in carefully-made screws, will be found to have but narrow belts of real bearing contact; whereas the V-threads may easily be made to take a very extensive bearing against each other, and will therefore be more enduring. In rounding off the Whitworth thread the tool must have a curve of different radius for every variation in size, and a really true Whitworth thread would require the most careful preparation and adjustment of symmetrical counterpart tools for each rate of thread. Practically, this rounding off for the Whitworth screw tools is only gauged by the workman’s eye, and, as Sir Joseph explains, his thread was arrived at as a compromise thread, and for works in which cast iron is a predominant material. These facts do not affect the value of the Whitworth thread for rough purposes, for strong works in cast iron, for standard engineering boles and nuts, or for ordinary screws made with multiple-point hand tools But they tell against adopting the Whitworth thread for micrometric and scientific purposes, and for fine works in homogeneous material. On the other hand, the v-thread truncated to any extent can be made with exactitude in all sizes by a suitably truncated point tool of the proper angle, aided only by careful calipering of the cylinder.

[The form of thread recommended at the conclusion of Dr. Edmunds’s paper is a V-thread having an angle of 50°, the thread on the bolt being flat on the top and sharp on the bottom, while that on the nut is flat on the smallest and sharp on the largest diameter, corresponding in this respect to the plan adopted by the Pratt & Whitney Company in their new screw-gauges.]


Sorry. No information on contributors is available for this page.

home | library index | search |  contact drawing of envelope |  contributors | 
help | privacy | terms of use