From the Rolls-Royce experimental archive: a quarter of a million communications from Rolls-Royce, 1906 to 1960's. Documents from the Sir Henry Royce Memorial Foundation (SHRMF).
Method for checking and calibrating speedometers.
| Identifier | ExFiles\Box 33\4\ Scan003 | |
| Date | 23th August 1913 | |
| THE AUTOCAR, August 23rd, 1913. 327 XC1956 Useful Hints and Tips. A Method of Checking and Calibrating Speedometers. THERE have been discussions lately in the pages of The Autocar concerning the accuracy or inaccuracy of speedometers, and it occurred to me that the following might be of interest. I have often wished to have a check upon the accuracy or otherwise of my speedometer, which is a Stewart magnetic one of last year's make, and starts reading from zero; but the roads in the district of Kent where I reside do not afford many level stretches, nor could I place any reliance on mile posts, so that for some time I had to rely upon the maker's reputation until it occurred to me that I could obtain a fairly reliable check by jacking up the back wheels, running the engine at varying speeds, and counting the revolutions of the back wheel—this particular method is, of course, only applicable to cases where the speedometer is run off some part of the mechanism abaft the engine. I found I could keep the engine running at varying speeds fairly steadily but when both back wheels were off the ground, they did not revolve at the same speed, so, as I was single-handed, except for my young son, whom I placed on the front seat to watch the variations of the speedometer, I could not take simultaneous readings of the two back wheels. I accordingly decided to let one wheel rest on the ground and took the revolutions of the other, which, owing to the action of the differential gear, ran twice as fast as the two should. The instrument used to count the revolutions was a little counter called the Ideal—to be obtained of Messrs. G.{Mr Griffiths - Chief Accountant / Mr Gnapp} W. Burton Griffiths and Co., of Ludgate Square, London—and all that there is to do to use it is to make a small counter-sunk hole in the centre of the hub cap, in which the point of the counter could grip, then hold the counter against the wheel when the engine had settled down to a steady speed and time it for one minute or half a minute. Half minute readings were taken at the higher speeds, as, naturally, the engine was racing. Readings were taken as near every five miles as the engine could be made to run steadily up to 46 m.p.h., which was six miles faster than my car can go safely or comfortably. The back wheel was, at 46 m.p.h., making 1,000 revolutions per minute, and the engine 2,500. The readings thus obtained were plotted on squared paper with m.p.h. and revolutions of wheel as abscissæ and ordinates, vide accompanying diagram, the readings obtained being shown by the small circles. The speeds and revolutions calculated from the size of the tyre are also shown as a straight line passing through zero. The tyres are 810 mm., their circumferences were measured and found to agree with the nominal sizes to within ⅛in., the tyres in question being a Michelin steel-studded and a Kempshall non-skid. The number of revolutions thus calculated worked out at 633 per mile. An examination of the diagram shows that this speedometer is really very accurate in its gradations, and has apparently a constant error of from 1½ to 2 miles per hour all along its scale. Observe, however, that no account has been taken of the compression of the tyre when it carries its load, which, from some measurements I have made, makes some difference. The results obtained, up to the present, were somewhat curious and conflicting, so that it cannot at present be stated exactly by how much it affects the speedometer, but it is hoped to make this the subject of a further communication. The results so far obtained pointed to tyre depression having a not inconsiderable effect in the direction of making the speedometer readings tally more nearly with the observed readings at the higher speeds. A straight line through the observed readings—shown dotted—indicates that there is apparently a zero error, so that at very low speeds the percentage error is great, e.g., at 5 m.p.h. its error is 25%. The next step was to check the distance recorder; this was simply done by taking a reading of the back wheel revolutions whilst the speedometer indicated one mile, care being taken to start counting the moment the recorder jumped to any figure. By calculation this was 633, and the observed number was 646—or a difference of 2%, which, when account is taken of the depression of the tyre, is practically correct. On the whole the tests of this speedometer proved it to be very satisfactory, and quite accurate enough for every day use. J.J.K.-S. [Diagram Caption] Diagram showing (1) calculated speeds of car as compared with (2) speedometer readings during test. No allowance is included for tyre depression. Size of tyres 810 mm. = 8.34 ft. circumference. Measured circumference of tyre = 8.33 ft. Calculated number of revs. per mile = 633. [Diagram Labels] Top Axis: ENGINE R.P.M (TOP GEAR) 500, 1000, 1500, 2000, 2500 Left Axis: SPEEDOMETER M.P.H. 10, 20, 30, 40, 50 Bottom Axis: WHEEL R.P.M 100, 200, 300, 400, 500 | ||