Monday, November 13, 2023

Adjusting rotate mechanism to properly select characters on typeball; performing a transplant


When I set the solenoids to choose a position on the typeball to print, the various latches cause the tilt and the rotate arms to pivot, pulling on the metal tape that allows the fixed position movements of the arms to be transmitted to the carrier mechanism which can be anywhere from left to right on the typewriter. 

The tilt selections work fine - the arm pivots to one of three positions or remains in its idle spot, which causes the typeball to pivot to one of four tilt levels. This selects 11 of the 44 characters on each hemisphere (upper or lower case) of the ball. 

The rotate selection does not move sufficiently to turn the ball all the way to its +5 or -5 position, instead mis-selecting a character on the +/-4 for printing. That arm moves between eleven positions - five positive and five negative, with the idle position selecting the eleventh possible character at this tilt ring. Rotate by itself chooses which 4 of the 44 characters are to be printed, with the tilt further refining that down to a single choice. 

The rotate spring under the golf ball maintains some tension on the rotate tape so that when the pivot arm moves inward it lets out some tape and turns the ball in one direction away from idle. When the pivot arm moves outward it pulls in more tape, turning the ball in the opposite direction from idle. 

A pivoting shift pulley on the other side of the typewriter can rotate in or out by a fairly large distance, pulling or releasing the rotate tape which turns the golfball between the upper case and the lower case side of the ball. There is an arrow on the golfball cap which points at the idle position of one of the hemispheres. The ball sits at idle with the arrow either facing the paper or 180 degrees away when shifted. 

The idle position is correct for the golfball, in that when the rotate pivot arm is in its vertical idle position, the ball has the proper characters facing the paper. I have the correct rotate spring tension and the idle (home) positions are correct in both upper can lower case. The issue I face is that the rotate pivot arm is moving less than the required amount when it moves out or in away from the idle position. 

I have also noted that the latches for tilt and rotate are not restoring themselves just under the actuating bar which moves them, which suggests incorrect adjustment of some stop tabs. This may be contributing to the incorrect rotation of the type ball. 


There is a rod that pulls or pushes on the bottom of the pivot arm, which causes the pulley at the top to move in the opposite direction. The point of attachment of this rod is adjustable, so I will move it closer to the arm's pivot point to increase the movement of the pulley for a given movement of the rod. 


The latches which are selected to cause the movement of the tilt and rotate pivot arms are supposed to just pop back under the actuation bar when the print cycle is done. A spring holds the latch under the actuation bar so that when the print cycle takes place and the bar moves downward, it pulls the latches downward. The solenoids for selecting a character will hold the latch away from the actuation bar thus when we select a solenoid for a latch, the latch will NOT be pulled down by the actuation bar. 

The downward movement of any latches which remain under the actuation bar cause arms in the Whiffletree mechanism to rotate. The length from the latch to the pivot of each arm is designed so that they deliver different amounts of movement based on the purpose of the latch. The latch which moves the ball one rotate position from home, R1, produces the least movement of the rod. The other latches produce increased movement and causes some to add or subtract to the movement when used in concert. 

Thus, the R1, R2, R2A and R5 latches can move the rod to any of the eleven positions in and out that should cause the golfball to rotate to its eleven positions. The amount of movement of the latch depends in part on where it starts from when the actuating bar moves it downward. There are stop tabs that sit above each latch and block it from moving upward past the tab. These are to be adjusted by bending them upwards or downwards until each latch will just pop back under the actuating bar at the end of a print cycle.


I saw that the entire block which has the stop tabs is attached to the frame with four bolts. I removed the motor from the rear to access the bolts, but there was not enough movement to completely resolve the stop tab issue so I will have to figure out how to bend them later. The access route for pliers to the stop tabs has the print shaft directly in the way. It might take some ingenuity. 

After moving the rod closer to the pivot of the rotate pivot arm, I was able to get to the full -5 position but the +5 still doesn't make it past +4. I think I can fine tune this to select properly. 


To get the cord tension correct for the cords that pull the carrier left and right during space, backspace, tab and carrier return operations, I need to rotate the CR drum and tighten the screws at the point where the tension is just right. The drum on this machine is extremely difficult to rotate. I believe it is corrosion between the shaft and the drum that causes the problem, since clock oil did nothing to free it up. The one time I did rotate it and tightened the screws, it slipped back to the original position anyway.

The donor Selectric I bought has a CR drum that rotates easily, free of corrosion. Swapping it does involve some major transplant activity, as the entire shaft that connects the CR drum, the tab drum and the mainspring has to be replaced. The shaft with the CR drum mounts on a back plate through a ball bearing. The tailpiece that locks to the mainspring is a nylon piece that is press fit on the back of the shaft. The shaft does slide through a front plate ball bearing and then into the tab drum. 

In addition to the drums mainspring and shaft, there are levers attached to the back plate that activate functions like backspace. The levers on the donor machine are different from the levers on the 1053. Fortunately, I saw circlips that allow the levers to be removed and attached.

The transplant operation began. The setscrew on the tab drum is loosened so that the shaft can pull out of the drum. I unscrewed the back plate and then began to free up the various levers and other parts. One dangling lever is the activation mechanism to drive the backspace power from the operational shaft clutch, which has a spring to pull it back out of the way unless it is being selected by the clutch. 

I carefully detached the spring, pulled the lever out of the operating mechanism and maneuvered the various levers behind the machine to free up the plate. I had to remove part of a microswitch and take off the motor start capacitor mount as well. The two cords were completely removed from the drums at this point. 

The 'patient', mid transplant

Once the old plate with its CR drum was removed, I began the swap of the levers between the original and the donor plates. I have many of the special tools that IBM repair people used to work on Selectrics, including a circlip insertion tool that came in very handy here. I verified the easy rotation of the CR drum on the assembly I will be putting into the typewriter. 

Two plates with drums, ready to exchange levers

I did work on the original CR drum and shaft a bit. Even with two vicegrips, it was difficult to rotate the drum on the shaft and it felt very gritty and uneven. The donor machine has paid for itself several times over by this point.

Reinserting the assembly is a complex reversal of the removal steps. There are levers and other parts inside the machine that need to be moved to a specific orientation before the plate with the drum can be slid back into place. The tab drum on the front needs to have the setscrew positioned over the flattened section of the shaft front, while all the levers and parts are in the proper orientation, including the dangling lever, before it can be screwed into place. The small spring was replaced on the dangling activation lever, the start capacitor mount reinstalled and finally I am ready to begin the next steps.

Tomorrow I will reattach the tab and CR cords, get them wound properly, attach the mainspring, wind it to proper tension, then rotate the new CR drum to set the correct cord tension. I am hoping this is a relatively straightforward and quick task now that the bad drum is replaced. 

No comments:

Post a Comment