Sunday, December 31, 2023

Fixed backspace, fine tuned carrier return operation, verified solenoids will trigger functions

DONOR SELECTRIC ALLOWS EASY ACCESS TO VISUALIZE PARTS LOCATIONS

The IBM 1130 version of the typewriter has quite a bit bolted on to allow for computer operation and support the front panel pushbuttons. Sadly all the additions tend to block the view and access to the primary areas that implement the five movement functions - tab, space, backspace, return and index (line feed). 

The donor selectric typewriter I have been using does not have any of the additions and thus we can use it to look at and understand the mechanical functions we have to work on. First we look at the underside of the machine with the rear at the bottom of this picture. We can see the five long plates, which IBM calls interposers, that trigger tab, space, backspace, return and index movement functions (left to right of the picture. 


In the picture above you can see the five springs that will pull the interposers to the rear (down in the picture) when they are tripped. The two horizontal bars across the middle are the two clutch release mechanisms. The left bar is pushed by any of the first three interposers when it moves rearward; this causes the left operational clutch to take one half-rotation cycle. The right bar is pushed by either of the right two interposers moving rearward, and it causes the right operational clutch to take a full rotation cycle. 

The next picture is shot from the rear, showing the operational bracket which is what helps translate the rotary motion of the operational clutches into an down-up movement of various linkages to perform the actual movement function.  In this picture we see one of the levers dangling down and having a spring hold it forward so that the bottom lip is not under the entire bracket which is a bail that moves down and up. 

However, the interposer which has been pulled rearward after it was tripped will push its lever lip under the bail and thus as the cam on the operational clutch turns the bail cycles down to pull the lever. I have disassembled part of this machine so that we only see one of the levers in place, but another would dangle in the opening to the right of our one lever. 


Just visible at the bottom of the picture above are the springs we saw in the first picture, the ones that act to pull interposers to the rear if they are not latched forward in their idle state. 

This is another picture, taken from the side in the rear, showing the same bail that pivots down and up as the left operational clutch rotates. You can see that the one dangling lever does not have a lip projecting out to the left under the edge of the bail. If you follow the small spring fastened to the middle of the lever, the other end is the back of the interposer and it has a projection below that pushes the lever back under the bail when the interposer is pulled to the back. 

BACKSPACE ISSUE WAS A BAD SETTING, NOT A MISSING SPRING

My first guess for why backspace did not trigger a clutch cycle or function was that the spring was missing that should pull the backspace interposer to the rear. When I looked, however, it was attached and working. Instead, the problem was that the lip of the dangling lever was too high and when the interposer tried to push it rearwards, the lip jammed into the side of the bail instead of under it. 

That not only kept the lever from fitting under the bail to be actuated, it also kept the interposer from moving far enough back to trip the horizontal bar for the left operational clutch that we see in the first picture of this post. 

The mechanism that transfers the levers downward motion into moving the carrier back a column involves a couple of pivoting levers with screw adjustments. The vertical screw adjustment determines where the lever lip dangles and was set too high when I checked it. I adjusted it so that tripping the backspace interposer allowed it to move fully rearward, pushing the lever lip under the bail and pushing the horizontal bar to release the clutch for a cycle. 

The lower screw and nut control the backspace lever

As this pivoting linkage moves downward from its initial point we determined with the screw above, it pushes a second smaller pivoting linkage that will shove the backspace rack sideways. That has its own screw/nut adjustment which determines how far the backspace rack is moved when the bail pulls down on the lever. 

The adjustment instructions for this are a bit odd, as the backspace function will not successfully move the carrier back when the machine is hand cranked. It only works due to the tiny bit of additional momentum from full motor power operation. Thus one has to adjust the backspace rack so that the escapement pawl is right on the edge of falling into the prior column's tooth on the escapement rack, but not achieve it. 

Escapement pawl holding the carrier at a column position

I wasn't able to capture a good picture of the pawl at the point where it is just shy of engaging the next column back, but imagine the pawl (triangular object projecting upward in this picture) sitting on the left edge of the tooth just to the left of where it is engaged in this picture. I successfully completed the adjustment and the backspace function is now ready to go. 

CARRIER RETURN MECHANISMS CONTROLLING THE MARGIN BAR

The way that carrier return works is tied into the margin bar, on the front of the typewriter, which is used to detect when the carrier reaches the column set as the left margin. As the right operational clutch pulls down on a lever to rotate some linkages for starting the carrier return, it also pulls on a rod that releases the margin bar to pop rightward; it is normally latched to the left. 

When the carrier strikes the left margin, it drives the bar leftward. An air piston is used to decelerate the carrier rapidly. The left movement of the margin bar pulls on a second rod, releasing the carrier return function. 

I had to adjust both so that they worked reliably, as otherwise the carrier will jam to the left edge of the machine instead of disengaging correctly at the chosen left margin column. This was done and return is now ready to go. 

TWEAKED ALIGNMENT OF PUSHBUTTON MECHANISM FOR RELIABLE OPERATION

I tested the pushbuttons and slide the parts around to ensure that it would reliably trigger tab, space or return when the associated pushbutton on the front of the typewriter was pressed. 

PULLED ON EACH SOLENOID ARMATURE TO VERIFY IT TRIGGERS ITS FUNCTION

The solenoids of the operational magnets are underneath the machine. When energized, the pull down on a link which trips the associated interpose. Thus the interposer can be tripped either by a pushbutton, if one is assigned to that function, or by a computer command. I checked that each armature would reliably trip its interposer and drive the movement function. 

View from underneath of operational magnet assembly

side view underneath showing solenoids and armatures

CHECKING ADJUSTMENT OF TWO MICROSWITCHES ON OPERATIONAL MECHANISM

On the sides of the operational magnet assembly are two microswitches and pivoting levers. These are activated by the interposers for tab and carrier return, respectively. Both of these are relatively long movement functions, compared to space and backspace which last the same time as printing one character, or index which takes exactly twice the time to print a character. 

A tabulate will move the carrier rightward until it reaches a column that has tab set, or the right margin whichever comes first. That is a variable number of columns from the current position of the carrier, thus a variable time delay. Similarly, a carrier return moves the carrier leftward until it reaches the left margin. The time it takes depend upon the current position when return is triggered and the distance to the left margin setting. 

These two microswitches are part of a circuit that is used to feed back to the computer times when the machinery is busy on one of these long functions. The computer will hold off on triggering the printing of the next character or any other movement commands until the long function has completed.

The time that the operational clutch is busy is usually shorter than the total time of the long function, so the 1053 uses a pair of switches per function to block the computer from initiation up to the completion of the long movement function. When the movement is first commanded by the computer, through energizing one of the operational magnets, the interpose slides to the rear to start off a clutch cycle. 

The interposer closes the switch on the operational magnet assembly and the pivoting bar holds the switch closed until the end of the clutch cycle. This is because the interposer is forced forward and relatched in the midst of the clutch cycle thus the switch otherwise would open only partway through the cycle. The pivoting bar swings out of the way when the clutch check pawl engages to hold the clutch idle at the end of a cycle. Thus the microswitches on the operational magnet assembly are active from initiation through the end of a clutch cycle. 

A second set of microswitches are placed on the linkages which are latched while the long movement is underway, so they stay closed long after the operational clutch completes its cycle that started the movement. A switch on the left side near the rear latches when the tab torque bar is twisted and is released when the carrier hits a tab or right margin. A switch on the right side near the rear is closed when the carrier return is latched on and opens when the margin bar is struck by the carrier reaching the left margin and pulling a rod. 

The combination of the two microswitches in each pair ensure that the computer is blocked from the time that the interposer is first released rearward until the long function ends at a margin or tab stop. I verified that the switches for the tab were working properly but the switch for carrier return on the operational magnet assembly is not releasing properly when the clutch cycle ends. 

RIGHT HAND OPERATIONAL CLUTCH CHECK PAWL IS NOT FALLING INTO SLOT

I determined that the pivoting bar which holds the carrier return switch closed during a clutch cycle is not releasing because the check pawl on the right operational clutch is not snapping into the slot as it should. The pivoting bar is blocked from releasing until the pawl falls into place. 


the right operational clutch, whose right side is show here, has a disk that contains the slot which is adjusted relative to the rest of the clutch. An eccentric screw, visible in the picture above, is turned until the check pawl falls into the slot as the clutch ends its cycle. The smaller bolt in the picture, and a counterpart 180 degrees around the disk thus hidden in this picture, must be tightened once the eccentric adjustment is complete.

My initial difficulty is the thinness of the bolt heads. When I put a 5/32 wrench on the eccentric, I can't get a 1/4 wrench to grip on the tightening bolt. This requires special very thin wrenches. I have several thin wrenches for use with Selectric but I do not have these sizes in a very thin format. I will have to figure out a method to turn and hold the eccentric while tightening the outer bolt. 

x



Saturday, December 30, 2023

Operational clutch trigger work on the 1053 console printer

GETTING BACK TO THE SELECTRIC

It has been a very busy few weeks with holidays, family obligations, volunteer work and my neighbor in the strip mall pushing hard to get me to relocate so he can expand his chiropractic practice. 

That involved lots of work evaluating properties and working on requirements so that I minimize the disruption to my shop work. Too, he will pay to do the moving of everything and I have to find a suitable replacement without any additional cost burden other than perhaps slightly higher rent. 

I have organized and placed most parts in labeled containers, which helps me while working and is also going to make a move be convenient as there will be no special packing necessary. There are so many tradeoffs and subtle requirements due to the heavy weight of some equipment, access door minima, power and air conditioning necessities, plus of course keeping the travel time between home and shop to a reasonable value. 

Commercial real estate figures can be quite confusing - some include public area space into the square foot they quote for a unit - and there are various lease types some with big downside exposures. Even after adjusting to the actual floor space, the usable area can be quite different. If a space is configured as multiple offices, there is wasted space in hallways and it takes a lot of planning to be sure that all my equipment and tables (19 of them) will fit comfortably. An open rectangle, on the other hand, is like my current shop and offers the most flexibility. Anyway, back to the good stuff working on computers for now.

OPERATIONAL CLUTCH TRIGGERS

The Selectric based console printer uses motor power to accomplish five movement functions - space, backspace, tab, carrier return and index (line feed). There are two clutches for these functions, one that completes its role in one-half a rotation of the main operational shaft and the other taking an entire rotation at a time. 

The carrier return and index functions need the full rotation to complete their work and are both implemented on the right clutch which turns 360 degrees when tripped. The other three are completed in a half rotation, 180 degrees, and share the left clutch. 

A long front to back plate is assigned to each of the five functions. It is held forward with a spring loaded latch, but when the front is pushed down it is pulled rearword by a spring. Any of the left three plates will trip the left clutch to take a half rotation. At the same time that a plate moving to the rear trips the clutch, it moves a lever so that when the clutch rotation moves a bail up and down, it pulls on that lever. 

Thus, the left clutch has three levers, controlled by the three long plates which trip the clutch. The particular plate that was activated causes its associated lever to be pulled down by the mechanical movement driven by the clutch. 

One lever will twist the escapement torque bar causing the carrier to be released for rightward motion which is stopped at the next column position. This is the space function. To push on the plate and let it move rearward, the console printer has two mechanisms. There is a solenoid with a link that pulls down on the plate to release it. There is also a lever driven by the Space pushbutton that pushes down at the front of the plate. Either of these activations of the plate cause the clutch to operate, the escapement lever to be pulled down by the bail, and the carrier move over one column. Just to add to the complexity a bit, there is a third way that the escapement torque bar is twisted, in addition to the lever powered by the operational clutch, This third method is a pull rod that is driven by a cam on the print shaft, so that at the end of a cycle printing one character, the carrier then moves right one column. 

The next lever when pulled down with twist the tab torque bar, which causes the carrier to lock all the pawls out of the rack teeth so that the carrier glides to the right. When a tab position is set at a given column, it pushes a release on the carrier so that the pawls snap back into the rack; this stops the carrier at the position which was set. 

The third lever when pulled down will force the backspace rack to the left, which forces the escapement pawl out of its rack and physically moves the carrier back about one column. The escapement pawl drops into that rack tooth one column to the left and we have accomplished a backspace. 

The right (full rotation) clutch has two plates associated with it, for carrier return and index. Just as with the left clutch, when either of these plates is released to move rearward, it trips the clutch to spin once. It also pushes a lever so that it will be moved down by a bail driven by the rotation of the right clutch. 

When the index lever is pulled down, it pulls on a set of ratchet teeth to move the platen (roller upon which the paper sits) up one line. Since each of the five plates has a solenoid that can trip it to move rearwards, the processor can command any of those plates by energizing its solenoid.

When the carrier return lever is pulled down, it activates a complex set of mechanical parts, including arming the margin bar to be tripped when the carrier reaches the column set as the left margin. It latches the carrier return operation to continue until the margin bar is tripped. The latched carrier return continues after the right clutch has finished its rotation, taking as long as necessary to move the carrier to the left margin. 

It pushes a nylon shoe down on a coil spring which has only one end fixed to the operational shaft. By stopping the coil from turning, it winds up, squeezes on the operational shaft and thereby turns the drums that pull the carrier to the left since the drum is now locked onto the operational shaft rotation.  When the carrier strikes the left margin that releases the latch which pulls the shoe from the spring. That disconnects the rotary energy of the operational shaft from the drum. 

At the end of the rotation of either of the operational clutches, a lever pushes the plates forward where they should latch into the idle position awaiting the next time the front is pushed to release a plate. One defect this console printer exhibited is that some of the functions were continually tripping the operational clutches and causing continuous movement. 

MISSING SPRING TO LATCH BACK IN THE IDLE POSITION, CHALLENGING ACCESS

The reason for the failure to stop after one rotation was that a spring at the front of the plate was missing. This spring pivots a small mechanism up to cause the plate to latch into its most forward position until pushed down at a future time. No spring, no latch, thus is remains rearward and trips the clutch repeatedly. 

These springs sit in an area with very limited access. All the parts that were added to allow three pushbuttons to trigger three of the five plates take up space right in front where I need to be. There is a button for Tab, one for Space and one for Return. 

The spring for the Tab function was missing, so that when a tab was requested by either solenoid or pushbutton, it kept retriggering. You can see the missing spring below, once I opened up access. 

missing spring

DISASSEMBLY OF PUSHBUTTON ASSEMBLY TO GAIN ACCESS

pushbutton parts are in the way

After hours fighting with spring hooks, powerful lights and almost no visibility, I realized that I would have to disassemble all the pushbutton related parts to free up space.  I pulled off all the parts and the difference is clear. 

parts moved out of the way
It was very easy at this point to attach the spring and verify that it latched the tab plate when it was returned forward at the end of one left clutch cycle. 
The spring is in place, pulling the latch to hold the plate forward

ALL FUNCTIONS LATCH BUT BACKSPACE NOT FIRING OFF CLUTCH

I reassembled everything but did notice that when I released the backspace plate, although it did release from the latched position, it did NOT trip the clutch. I am guessing that the spring which pulls that plate rearward is missing or disconnected. 

This is surgery that must take place in another area called the operational bracket which is near the rear of the machine but blocked by the mainspring and all the linkages that are turned when a lever is pulled down by a bail driven by a clutch. I have been successful attaching springs to the levers that are pulled by the bails, so I expect that I have enough access to accomplish this fix on my next shop visit. That should complete repair of all the movement functions. 

Thursday, December 21, 2023

Cords installed and holding; moving on to adjusting other functions of the 1053

SUCCESS IN TENSIONING THE CORDS

I was able to put the cords on the machine and apply tension. It is not perfect, a bit looser than the target scribe lines, but good enough to move forward.

BEGINNING VALIDATION OF THE MOVEMENT FUNCTIONS

The pushbuttons reliably trigger space, tab and carrier return activation, as a first check. I will exercise the machine further to be sure this reliably works as intended, one column per space and a full tab movement to the next position with a set tab.

In addition to the three functions for which there are buttons on the console printer, the solenoids can trigger backspace and index (line feed). I will verify that these operate as well as I can while manually cranking the machine. According to the manuals, a backspace only works properly under motor power.

ADDITIONAL AREAS TO ADJUST AND VERIFY

I have to finish installing my black thread for the ribbon color mechanism and adjust it so that the 1053 can choose programmatically whether to print in black or red ink. This involves fashioning a hook end for the thread and installing it inside the carrier to the ribbon lift mechanism. 

The remaining issue I have to resolve is the inability to achieve a +5 rotation. I suspect this is a rotate spring tension issue, but it may be some other adjustment. Once that is set, it would be time to run the machine under power to test out its printing and movement operation. 

Cause of escapement drum 'skipping' discovered and corrected, back to setting up cord tension

SKIPPING CAUSED BY PARTIALLY ENGAGED CARRIER RETURN SPRING

The sporadic grabbing or skipping that I was experiencing is due to the carrier return 'shoe' still partially engaged pressing on the spring to drive the escapement drum in the carrier return direction. The shoe should be away from the spring to allow the spring to rotate freely around the operational shaft. When the shoe presses on the spring, it winds up so that it is clutching on the shaft and causing the spring to turn with the main motor (operational shaft). 


The C.R. Clutch Arbor in the diagram above is turning with the operational shaft, but the C. R. Pinion is not directly connected to the C. R. Clutch Arbor. Instead, the coiled C. R. Clutch Spring turns over the Pinion but does not have enough grab or force to turn it. 

When the Nylon Shoe is pressed against the far end of the C. R. Clutch Spring, however, the spring is quickly wound tight by the C. R. Clutch Arbor, thus gripping around the C. R. Clutch Pinion so that it turns with the operational shaft. This pinion turns the drums and shortens the return cord to pull the carrier back to the left side of the machine. 

The Carrier Return Clutch Arm in the picture above is what moves the Nylon Shoe so that it either presses against the C. R. Clutch Spring or is not in contact with the spring. The adjustments on the Carrier Return Latch are what determine the grabbing force when engaged and the position away from the spring when not engaged.

ADJUSTING THE SETTINGS SO DRUMS TURN FREELY WHEN NOT IN CR MODE

Nylon Shoe contacting spring
Activating rod to Nylon Shoe

At the rear of the typewriter are these two adjustments for setting the action of the shoe to engage or release the carrier return clutch, implemented as a spring over the operational shaft. I loosened the lower adjustment and allowed the mechanism to move the shoe further from the coil spring when the return is not engaged. This eliminated the sporadic grabbing or skipping, which was the problem affecting my cord tension setting over the last week or two. 
Adjustments for CR clutch engagement


Thursday, December 14, 2023

Tab and return cords installed again but something still wrong, not holding tension

NEEDED MORE HANDS ONCE AGAIN

Adjusting the tension on the cord system requires that the escapement drum be held from turning, the return drum be rotated to shorten the combined cord distance, the return drum be pushed back against the rear bearing, the shaft that these ride on be pushed fully forward to remove play, remove the tension from the sliding pulley and then the screws on the return drum hub be tightened. 

The space where the return drum sits is restricted, where I can just about get one finger in to turn the drum and the tool in to tighten a screw. I don't believe that hypothetical additional fingers could get into the area at the same time. 

I did come up with a clamp to hold the sliding pulley all the way to the left, removing any tension it might apply. Then if I judge the cord length correctly, when the clamp is removed the pulley will be pushed out to the scribe lines where it belongs. 

I can combine actions, pulling the shaft forward and stopping the escapement drum from turning since these both happen in a different area below where the carrier travels. It just requires some care to be exercised since the two metal tapes (for tilt and rotate) run above the escapement drum and could be damaged while I work with the drum. 

Conceptually, this doesn't sound too hard. However, for someone whose fine motor skills were low on the bell curve, and who layers on the slight shakiness of someone in his seventies, this took more effort than it seemed. 

LOCKED IN PLACE WITH PROPER TENSION ON THE SPRING LOADED PULLEY FOR A SEC

I prevailed and got everything set up. The pulley was sitting with its outer edge right over the scribe lines that indicate proper cord tension. As I move the carrier by tab or carrier return, there is some odd clicking or stuttering around the tab drum and it released some tension each time until the spring loaded pulley was fully released. 

At this time I really don't understand what is going wrong to cause it to slip. Until I know the cause I can't correct it.

Wednesday, December 6, 2023

Continuing with tab/return cord installation and adjustments; resolving the missing ribbon color shift tape

ATTACHED SPRING ONTO BACKSPACE OPERATIONAL LATCH

The linkage to the backspace operational latch had to be put into place and the spring reattached that will keep it deactivated until the backspace function is triggered. This involved tedious manipulation with spring hooks through narrow openings. 

As is my luck, the spring got away from me and hid inside the machinery. After 45 minutes of hunting from every angle I finally discovered where it had wedged and retrieved it. I then went back to the spring hooks and fought it until it surrendered. The spring is attached and working properly.

Finally, there are adjustments that must be made again, such as the trigger condition for a backspace which is a screw with locknut that sits so high on the bracket that during removal, I had to lower it to clear obstacles. This has to be set up right in order for the machine to move just one column backwards reliably. I will address these adjustments a bit later.

REINSTALLING THE TAB AND RETURN CORDS

This is the same process I have done many times already, the only tricky part now is at the end after I have the mainspring pre-wound and the cords attached properly, I have to loosen the return drum, rotate it to tighten up the cords and tighten down the screws to achieve the proper cord tension.

The core tension arm is a pulley that is forced outwards with strong spring force. When the cords are properly tightened, they pull the pulley inwards partway so that the cord tension offsets the spring tension. There are scribed lines on the bracket where the pulley moves and the cords must be adjusted so that the far end of the arm is right at the scribed lines. 

This involves holding the rest of the mechanisms steady while turning the return drum to pull on or loosen its cord. Turning it against the spring tension of the cord tension arm is infeasible, so I use a clamp to hold the tension arm all the way in, relieving its pressure. I then adjust the return drum so that the cord holds the pulley at the scribed lines, minus a bit that I estimate, then lock down the drum by tightening the two screws on its hub. The clamp is removed and we assess the tension arm against the scribe lines.

This is a bit iterative but the desired result is achieved when I can space, tab and return with the cord tension arm remaining at the scribe lines.

When I thought I had it all set and removed the clamp, I saw the escapement/tab drum hop forwards, skipping teeth on the operational shaft gears. I then realized that when I move the return drum and tighten it down, I also have to move it back against the bearing in front of the mainspring, otherwise there will be slop that allows the gears of the tab drum to disengage. 

At this point the cords are all properly attached but the tension is again inadequate and the return drum is not snug against the bearings. When I get back to the shop I will resolve all that and should have a properly tensioned set of cords. The mainspring must be wound to the right tension once I have the cord properly installed, then we should be done with this extended task.

At this time of year, I have other obligations that reduce my time in the shop to only 3 or 4 hours per day at best, which contributes to the plodding pace of the 1053 repairs.

RIBBON COLOR SHIFT MECHANISM AND MISSING TAPE

The IBM 1130 console printer makes use of bicolor cloth ribbons, which are inked with red and black ink on the upper and lower halves of the fabric. By controlling the height that the ribbon lifts during a character print, the 1053 can print in either red or black. The color is controlled by program commands, firing one of two solenoids to shift to red or shift to black.

mechanism inside carrier
lever which changes tension of the tape



The implementation of the ribbon color makes use of a flat nylon tape that runs across a series of pulleys, much like the metal tape for tilt and rotate runs around pulleys. One end of the tape is hooked to the right of the carrier, the tape runs to the right edge, around a pulley, back in front of the carrier to the left edge, then down to the solenoid lever at the bottom left. The tape then goes back up to another pulley on the left side and runs to the left side of the carrier. On the carrier is a small pulley to route the tape back to hook to a pull mechanism that controls the ribbon lift. 

pulley on carrier

ribbon lift lever circled

tension adjustment

right pulley

solenoid lever

The original nylon tape was missing and there is no source of replacement ribbon. The IBM original had appropriate metal hooks on the ends to attach to the carrier. An adjustment screw moves one of the pulleys in and out to fine tune the tension so that it reliably moves between black and red halves of the ribbon. 

TESTED LACING CORD AS A SUBSTITUTE

I temporarily tied some lacing cord onto the carrier and ran it through all the pulleys to test its fit. It does appear to stay on the pulleys and work appropriately. The carrier moves left and right without impacting the tension on the cord, while the solenoids tip the lever up and down to add tension which is transmitted into the carrier's ribbon lift lever. 

MUST FASHION GOOD HOOKS ON EACH END AND DO A FINAL ADJUSTMENT

I believe a small ring, a miniature version of the rings for key chains, will fit through the hole in the ribbon lift lever. I need to find or make one. The other end just has to sit over a post on the carrier, it is less critical. When I have workable hooks I will get them fastened securely to the lacing cord and put them onto the 1053. The final step will be to adjust the position of the right pulley to assure that we select red or black as intended.

Monday, December 4, 2023

Swapped again to put new CR drum; fighting to reinstall a small spring

DISASSEMBLE THE CARRIER MOTION PART OF THE MACHINE, AGAIN

With the return drum holes partially stripped and not holding on the shaft, I had to disassemble the carrier motion mechanism to fix the issue. Cords hooked to both sides of the carrier wind around a pair of drums on a shaft which also attaches the main spring of the typewriter. 

During a carrier return operation, the shaft is rotated by motor power to pull the carrier all the way to the left margin. This also winds up the mainspring, whose energy will be released during each space or tab movement of the carrier to the right. 

I had to remove the circled parts, which in turn require removal of other parts to gain access including the motor start capacitor, its bracket, and linkages to the backspace operational latch.



STATE OF THE AXLE WHEN I PULLED IT APART

The front drum (escapement/tab) fits on the shaft with a setscrew tightened to a flatted section of the shaft. The rest of the mechanism pulls out the rear of the typewriter. The mainspring is carefully unwound and removed. The return drum is slid off the front of the shaft to remove it.

I found that the portion of the shaft where the return drum fastens (via two screws set 90 degrees apart around the hub) had some very sticky, gluey substance I had to clean off first. I suspect that a prior repair of the donor Selectric had added glue to help keep the drum from rotating. 

CLEANED THREADS ON DRUM SO SCREWS WORK WELL

The drum from the original mechanism had thread holes on the hub in decent condition. I carefully cleaned them out and worked the screws through them to ensure they would tighten down well on the shaft when I reinstalled everything. 

I slide the return drum onto the shaft and did a test with the two screws lightly tightened. The drum appeared to hold well, which is a good sign that it will work properly when I am applying full tension of the cords as my final adjustment. 

REASSEMBLING THE DRUM ASSEMBLY INTO THE MACHINE

Getting the parts back in place took some time, as I had to put the plate back in position with the return drum, add the mainspring, adjust the position of the escapement/tab drum to mesh properly with the operational shaft gears and tighten the setscrew. 

Next, the linkage to the backspace operational latch had to be put into place and the spring must be reattached that keeps it deactivated until the backspace function is triggered. This involves tedious manipulation with spring hooks through narrow openings. 

Spring around notches on the vertical lever, hooks to the rear

I have a set of tools that make it possible to attach and detach springs deep inside the machinery of the Selectric. There are hooks that grab one end of a spring and pull it into tension. These are useful when the far end is anchored and we just need to put the near end onto its holder. There are pushers that are used when the near end of a spring is attached and the far end must be pushed away into tension to fit over its holder. 

Finally there is a tool that fits through the body of the spring and clamps onto the far end. It allows a completely loose spring to be pushed into some cavity and then the far end hooked onto its holder, Before removing this tool, one would grab the near end with a spring hook and stretch it out to fit on the near end holder. 


The latch to the right of the backspace latch has its spring attached - hard to see in the picture but the spring fits around the vertical lever at the side notches and stretches directly backwards to a mounting point behind. Its role is to pull the lever backwards so that it is NOT pulled by the operational bail. The only way it is pulled down is if the activation of a backspace function forces the lever forward against its spring while triggering the operational clutch to take a cycle. Apologies for poor photo quality.

Far end attachment point circled

The surgery involves the tools inserted through every narrow openings, with limited opportunity to see what is going on or even to light it properly.

How I spend my hours at the shop

All too often, the spring pops loose and I have to start over. This is the kind of issue that arises which takes a process (reinstalling the drum, the cords and setting the tension) from what seems like a 30 minute task to a day or more of work. 

The area where I had to remove and replace the drum assembly is loaded with adjustments, quite a few of which are disturbed by the work and will need to be set properly once the cord attachment task is complete. 

The screw which wouldn't clear the horizontal bar until screwed way down






Saturday, December 2, 2023

Installation of new console loader on museum's 1130 system

REPLACEMENT FOR CUMBERSOME RELAY BOARD BASED LOADER

The first version of the loader, which was a quick and dirty design at the request of another museum, required three relay boards (2@8 plus 1@4 relays), an Arduino Mega 2560 and a board with resistors, all of which were stuffed under the desktop, in front of the keyboard and behind the typewriter panel. 

I recently designed a version using optical isolated switches where all of the switch functions plus the resistors easily fit on a PCB the same outline as the Arduino. Plugged in as a 'shield', the stack is not much taller than the Arduino itself.

I designed and manufactured the shield PCB with connectors for cables that run to the rest of the machine, to simplify maintenance on the 1130 in the future. The Arduino stack will sit behind the 1053 front panel, under the typewriter mechanism, thus being completely out of sight. A set of sixteen wires run up to the console entry switches (CES) on the same front panel and end in a connector to hook to the board.

CES switches, Prog Start and Load IAR buttons

Another smaller cable with a connector to the Arduino stack runs out of the front of the 1053 alongside the IBM cable for the CES signal lines, underneath the desktop. That cable is hooked to the Program Start and Load IAR buttons, as well as bringing 12V to the Arduino stack. 

Exit notch bottom right for IBM cable, USB cable and my wires

I first disassembled the existing relay board implementation before wiring in the two cables that will hook to the Arduino stack. The area looked much neater already.

WIRING SIXTEEN CONSOLE ENTRY SWITCHES

The CES toggle switches have sixteen wires running down to the 1130's logic cages, where the electronics can read the state of the switches. +12V feeds to the common terminal of all the switches and the normally open contact has a 470 ohm resistor in series to the wire running to the 1130 electronics. 

I soldered my wires to the 16 normally open connections such that my loader can deliver +12V through 470 ohm resistors to the 1130 electronics to virtually set the CES to a 1. I installed the connector and verified connectivity from the CES switch to the proper contact on the connector of the Arduino shield.

WIRING THE PUSHBUTTONS AND POWER

This design will activate as soon as the IBM 1130 powers up, but will not take control of the machine unless requested by a specific command code over the USB Serial link. When idle, not controlling the 1130, it provides the +12V to the common terminal of the Program Start pushbutton, replacing a wire connection to 12V installed by IBM. 

When the loader is activated by the special command character, it removes the +12V to the common terminal of Program Start. The normally closed and normally open contacts are powered alternatively by my shield to indicate button up or button pressed conditions. Thus, my shield drops 12V from normally closed and delivers it to normally open as a virtual button push. 

The Load IAR button is only fed +12V power to the common terminal when the rotary mode switch on the machine is set to either Load or Display mode. The normally open contact runs down to the 1130 electronics to trigger the setting of the IAR from the contents of the CES when the button is pushed. 

My shield will route the connection from the rotary mode control that is soldered to the common terminal of Load IAR, taking it to my shield and returning the output of the optical switch to the normally open contact of the Load IAR button. Thus, when I connect the two wires from that button, the shield is virtually pressing the Load IAR button. 

I grabbed +12V and ground from the power supply of the 1130 and delivered it over the cable to the Arduino stack. Thus, when the machine powers up, the Arduino will come up and go into its idle state, delivering +12V to the Program Start common terminal so that the operator can push this manually. When activated we disable Program Start button and instead virtually push it, plus virtually push the Load IAR button and virtually set the 16 CES. 

MOUNTED IN PLACE INSIDE TYPEWRITER CONSOLE PRINTER

I found a good location to mount the Arduino stack. There is an empty volume behind the 1053 front panel, where the stack will fit underneath the typewriter mechanism. I only needed a few inches deep and high while the existing area is comfortably larger than my needs. 

I mounted the Arduino stack on a bracket which itself is mounted to the back surface of the 1053 front panel. This panel is removed from the rest of the 1053 in order to move the typewriter mechanism for service while keeping the panel with its CES close to the electronics. Thus, when the front panel mounts on the bottom are unscrewed from the 1053 base, the panel is pulled forward and upward to move it away from the rest of the 1053. The two connectors to the CES and pushbuttons can stay attached as the move with the front panel.

I simply drilled a couple of holes in the front panel, low enough to be hidden by the table top when closed. The bracket is mounted and fits exactly as I envisioned.

Inside 1053, rear view of front panel, top view looking towards keyboard

The final mounting decision was for the USB connector. When someone wants to control the 1130 and load memory, they plug into the USB connector and open a serial connection to the Arduino. The command lines that are sent over that link command the loader to set CES and push buttons, but require an activation command character to start the process and a second transmission of that command to deactivate at the end of a loading session so that the Prog Start button is again functional. 

I found that the USB cable would route out of the front panel near the bottom along with the IBM and my signal wires, then it can lie alongside the right side of the console printer cover, low where it won't be seen. The cord is hung behind the 1130 so that it can be picked up and plugged into the serial communications device when it is time to use the loader.