Chapter 26.2 - Installing and Calibrating the Tempo Valve
None of the original tempo linkage except for the tempo lever itself, remained in the instrument as found. So it was necessary to devise a new system
A rod was fabricated to connect the tempo lever. A flattened end was soldered on, and a hole was carefully reamed to fit the shoulder screw so that movement was easy, but no slop was left.
A wooden linkage was made, to connect this rod to the slide valve actuating rod. The ability to make small tempo changes is important for proper interpretation of the music. So the linkage joints had to be made so they work smoothly, but without any slop or histeresus.
This configuration allows the bearing hole to be adjusted so that it fits the screw snugly, with little friction but no clearance.
A small plastic scale and pointer, indicating the tempo on the scale, was made for inside the valve box. So if the linkage adjustment were to be lost, recalibration would not be needed.
The process of calibration was tedious, but necessary, because of Aeolian's practice was to call for tempos all across the scale, from 35 to 85. The test roll was marked off in feet, from 0 to 10. Using a stopwatch, the time taken to roll one to ten feet of paper, depending on what tempo range was being calibrated, was read, with the organ winded and the spoolbox pressurized. A chart, reproduced below, was made in Excel. Functions in Excel automatically calculated the error and the actually tempo. A limit function (+/-5 initially; later reduced to +/-1) in the rightmost column, reported the result; "too high" or "too low".
The calibrating screws did affect the tempo, and only small fractions of a turn were needed. Since air does flow around the screws, and there is some space between the screws and the outline of the slot, turning one screw does have some effect on adjustment of adjacent screws. In other words, the screw adjustment at any one tempo position, was by no means independent. Therefore it was necessary to go back again and again, to finally approach a linear response with all the tempos correct. The leftmost lever position, near zero, must stop the roll altogether; there is no other means to turn off the player drive when playing from the keyboard. Actually the roll stops around tempo 7 to 9, depending how big it is, but this is much lower than in a player piano, which rarely works below 30 or 40. The rightmost lever position uncovers the big reroll slot, so that tempo ends up naturally higher* than 97.