ADJUSTMENTS FOR ROTATION SELECTION LOGIC
There are a number of adjustments to work through for the rotation selection and I am going to focus on getting them exactly correct, by the book, in order to achieve proper selection. Some of them will require me to power up the motor because the adjustment procedure requires the print cycle to occur at full speed.
One of the earliest adjustments is to set up the rotary pivot arm so that it as close to vertical as you can get it. The various manuals suggest measuring from the line etched in the arm at the top and bottom in order to ascertain whether it is equidistant. The snag is that there is no 'thing' to measure to from which it would be equidistant.
|Vertical etched line with no surface to match|
You can see the pivot arm with its line etched running up and down. Notice that there is no vertical surface anywhere in sight that you could measure from . The mechanisms behind the rotational pivot arm include the tilt pivot arm and they block visibility of the metal frame behind them, at any position where you can observe the pivot arm vertical line. The opening to the right of the pivot arm is quite deep and even at its far end there is no flat surface where a measurement point can be taken.
|Note the nonexistent reference structure to which the measurement is taken|
The vertical positioning of that pivot arm is accomplished by moving a turnbuckle on the black rod you can see running from the bottom of the pivot arm rightward into the depths of the machine. This is what pulls or pushes on the arm to effectuate tightening or loosening of the rotate tape, which in turn causes the type ball to spin on its axis to one of the five positive or five negative positions away from its idle 'home' spot.
When the vertical position is set by lengthening or shortening the black rod using the turnbuckle, a later adjustment has the black rod attachment moved up or down on the pivot arm to increase or decrease the displacement of the arm for a given movement of the rod. These two are interrelated and one must iterate to get them both adjusted properly. The later adjustment, the pivot point change, is intended to make the positive 5 and negative 5 rotations of the type ball turn the same number of degrees.
Before I can set the pivot point, however, I have to set the 'homing' position of the type element. This is the resting or idle position, called home. The adjustment manual requires me to half-cycle the machine using a set of selection latch activations, under machine power. A special tool, which I own, sits on the cycle clutch and forces it to step halfway through the print cycle, when the type element has achieved its full tilt and rotation. This is the dynamic half cycle tool.
|Dynamic half cycle tool|
|Initial home position setting|
I have to begin with a tilt 3, rotate 0 character from the 'upper case' side of the type element, which lets me adjust the ball. There is a setscrew under the bottom of the shaft holding the ball, which is loosened and then the ball rotated to make sure that the detent bar enters the teeth of the ball at the right point on the side of the teeth.
Next we adjust the selection mechanism until the I/O Home and Latched Home position of the ball is identical. There is a balance arm involved in this adjustment, which tweaks the mechanism so the ball remains at the home or rest position, in exactly the same spot, when either of the home types is activated.
Latched Home is a configuration where all the positive rotate latches R1, R2 and R2A are selected, e.g. they are NOT pulled down by the selection bail but R5, which controls the -5 rotation, is pulled down. With no pull on the positive latches there is no positive turn and pulling on R5 blocks the -5 action - a net zero rotation is produced.
I/O Home is a configuration where the three positive latches R1, R2 and R2A are not selected, so that they are pulled down by the bail and would cause a +5 rotation of the ball. However, because the R5 is selected and does not pull down, it imparts a -5 action resulting in no net rotation. The balance arm adjustment is what ensures that the -5 action of R5 is as equal as possible to the +5 action produced by R1, R2 and R2A.
The type element (golf ball) has teeth under each of its 22 columns (11 per 'case') with four characters vertically arranged in the column based on tilt position to achieve the full 88 character capacity of the element. A detent arm slides into the tooth to slightly turn the ball and anchor it in the precise rotary position for the desired column of characters. A similar detent arm locks the tilt mechanism to the precise degree of tilt to select which of the four characters in the column to print.
The selection mechanism pulls on the pivot arms and thereby tightens or loosens the two metal tapes for tilt and rotate. Each tape is hooked onto the carrier at one end and to a lever or rotating part inside the carrier with the tape's other end. Thus tightening the tape will pull the lever or cam in one direction, while loosening the tape lets springs move the lever/cam in the opposite direction as tension is relieved.
During the print cycle, the selection ball pulling down on latches causes the selection mechanism to tilt two pivot arms, one for tilt and one for rotate, a given amount to move the lever or cam to its approximate correct position for the chosen character. The detent arms then move in and jog the type element to its final correct position.
To work properly, the selection mechanisms have to get the tilt and the rotate position close enough that the detent arms always enter the proper tooth position to lock the ball correctly. The homing adjustments and all the checks as I adjust the selection mechanism are intended to ensure that we will have a good initial position as the detent arms begin to move in.
The tilt mechanism does work correctly. The detent arm barely causes any adjustment to the tilt as the initial position is good. The rotate mechanism, however, is where we have problems. The machine presented with a failure to move the type element enough to reach the +5 or -5 columns before the detent arm moved in, thus we locked the ball to the +4 or -4 column incorrectly.
All my adjustments are intended to get the typewriter to initially move the ball so that for any of the eleven rotate positions, -5 to home to +5, we are set so that the detent arm enters the teeth at the right point to ensure that we reliably lock the ball on the chosen column.
SELECTION LATCH RESTORATION ISSUE
The hanging selection latches are pulled forward by springs so that they will hook under the bottom edge of the selection bail. That bail moves downward during a print cycle, thus pulling down any selection latch that is under the bail edge. Choosing a character involves selectively activating some of the solenoids for R1, R2, R2A, R5, T1, T2 and AUX and any activated solenoid holds the selection latch out of the way so that it is NOT underneath the bail. Thus, an activated latch is not pulled down while an unselected latch moves downward during the print cycle.
The selection latches are supposed to have just enough clearance when the print cycle ends, based on the stopping position of the selection bail, that they will restore or pop under the edge of the bail if they had previously been activated during the print cycle. This restoration is important, otherwise a selection latch might incorrectly remain off the bail edge and pull down when the solenoid did not fire to choose it.
In order to adjust this, the manual suggests bending the metal stop tabs of any errant selector latch until it has a given clearance beneath the selector bail bottom edge when we are idle, not in a print cycle. It is very unusual to need to adjust one of these and nobody who worked as a typewriter technician (and who is reading the Golfball Typewriter Facebook group in order to respond) has ever seen a machine where all the stop tabs are wrong. I too worry that having them all out indicates some other error in the machine, thus bending tabs will put the machine further out of adjustment in reality.
The stop tabs are just bent sections of the backing plate where the mechanism is mounted, bent forward at nominally 90 degrees but then bent to push the selection latch down so that it pops under the selection bail bottom edge when the print cycle is at idle.
The selector bail pivots on a rear rod, pushed down by cams on the main operational shaft when they rotate during a print cycle. Rollers on the end of the pivot arm ride on the cams. The adjustments given for changing this part of the machine carry warnings that it should normally not need to be adjusted, but I can't see any adjustments of the position of the rod about which the selector bail pivots that might raise it at the idle point of a print cycle to let the selector latches restore.
The design of the cams that move the selector bail up and down does move the bail a bit higher than its rest position at the start of a print cycle, then proceeds to move it downward to pull on any selector latches still under the bottom edge. Thus, in this typewriter where selector latches do not restore at the end of a print cycle, they will restore at the very beginning of the next print cycle when I am hand cycling the machine.
The adjustment instructions insist on having the restore occur when the print cycle is at idle, however. The bail is held up by the rollers riding on the cam, thus there is no room for the bail to move further up; I can't hope for momentum during powered printing cycles to get the latches to restore.