The Comptometer was a key driven machine consisting of a number of columns of keys numbered from one to nine. The largest of number of columns of keys the author saw was twenty columns and such machines were located in the Australia Post Office and Australian Customs Departments. The normal size of such machines was of ten and twelve columns. Because Australia worked in £’s, shilling and pence, at the time, the American built machines were built especially to cater for this type of currency. A separate section will cover this mechanism and for the main part we will concentrate on the standard machine of ten and twelve columns by nine keys. These machines being capable of addition and subtraction (using the nines complement) multiplication by repeated addition and division by repeated subtraction, again using the complement. Skilled operators were required to operate the machines and because of this ‘Comptometerists Schools’ were held in each Australian States to provide clients with a steady pool of skilled operators. In Melbourne the school was first run by Miss Doris Hedley and later by Miss Verna James.

If so how do we make it transfer a carry over in to the next column each time the gear makes a complete turn? How do we ascertain that any residual positions can be cleared to a zero position when required?

THE ROCK FRAME

Earlier models of the Comptometer had permanent meshed gears and this continued up to the model ‘F’ when the idea of continuously engaged gears was dispensed with. From the model ‘H’ onwards all units incorporated a ‘Rock Frame’ and this was a section that held the position of the Accumulator and the data it represented as long as the ‘Rock Frame’ was engaged. When the rock-frame swung out by an actuation of the clearance handle the residual tension in the carry over sprint returned the mechanism to a known position we will call zero. The rock-frame mechanism was, as far as I am aware, unique to the Comptometer for its introduction made the complete carry over section and accumulation storage section removable for service.

The illustration is an illustration of the rock frame mechanism on a model ‘K’. For our purpose, at this stage, the mechanisms is almost the same for models ‘H’ through 3D11 and 992. The segment is not the segment lever segment, which has a flatter slope, and in this illustration Segment Rack is rounder in shape than the segment lever in manual models however, this is only due to the shorter fulcrum on the electric models. There are control differences such as the error control mechanism which will addressed later. The Rock-Frame is shown engaged.

When the clearance handle is pulled forward the shaft 35-A cranks round clockwise. The toggle 37-A breaks from its rigid position and the carrying gear (consisting of the gear guard 26-F) is mounted upon comes out of mesh with the accumulator carrying pinion 23-A. Because the gear is spring loaded via the carrying gear spring 25-A and the escapement 27-A is held by the carrying detent 26-A or B the tension in the spring, representing stored data between 1 and 9, is cleared at the gear returns to it’s zero position; against the zero stop illustrated below.

At this stage we have accumulation, in one column, and zeroisation in any column that had a loaded carry spring. Our next consideration is much more complex. The reason for the complexity is that we can not enter data into an accumulator that is in motion. Secondly we have not even contemplated a method of carrying into a stationary higher order; the next column to the left. We with call the function the carry function and the parts that complete this operation will be the Rock-Frame carrying mechanism and the accumulator in the next column to the left of the gear moving from 9 to 0.

There are exclusions to the carry taking place. Firstly if the segment lever roll is down. The latch, or detent, ‘H’ on the bell-crank prevents the escapement being released until such time as the roll returns to it’s up position. At that time the parrots beak part of the bell-cranks stops the rotation of the accumulator. As the escapement is released by the parrots beak ‘G'; toggling the detent ‘H’ the escapement turns. Flicks a locking dog from under tail ‘B’ and as the cranks dips the pawl ‘F’ moves forward releasing the cranks hold on the lantern wheel pins via the ‘parrots beak’ ‘G’ and releasing the energy in the spring into motion, driving the bell crank forward thus pushing the lantern wheel pin of the next higher order. Thus a one is added to the next column on the lefts accumulator. The column transferring from 9 to 0 completes its operation going to zero and the next higher or increases by one significant digital. That is increases by one completing the carry. The escapement continues its rotation presenting a valley in it’s cam face. The Roll ‘D’ drops into the valley allowing the return spring attached to ‘C’ to pull the crank back against stop ‘A’. The Parrots Beak stop ‘G’ locks behind a lantern wheel pin stopping any further rotation and a dog locks under the tail ‘B’ of the bell crank holding it in its up position.

It should be noted that when the lantern wheel is actuated by the carry pawl over rotation is stopped by the detent ‘E’ entering between the pins in another position on the lantern wheel. This timing is important as one has to make certain the carry can rotate the wheel far enough to introduce the carry before being stopped by the detent.

This process is repeated in each column providing a mechanism by which addition by direct action can take place, subtraction can take place by use of the nines complement. Similarly repeated functions of this process will provide multiplication and division. Routines were also provided for extracting the square root of a number.


ERROR CONTROL MECHANISM.

At the start of this description of the operational mechanism of the manual Comptometers models 'H' through 3D11 I described the operation of the Segment Lever. At that time I mentioned a parallel lever mounted inside the main segment lever forming a pair of levers. The inner levers function appears to be mainly for the purpose of error control. Between the main lever and secondary lever at the rear of the machine is a 'Segment Lever Bell Crank’ This bell crank pivot on one lever and is actuated by the second lever. In its rest position it is held towards the back of the machine by a long spiral spring. An elongated slot attaches the segment lever bell crank to a mechanism called ‘the trigger’ and is retained in a home position by a small spiral spring. It is strange how time changes ones understanding of things mechanical and the cunning nature of their design. Had attachment to the bell crank been solid or firm the spider like trigger mechanism would not have withstood the rapid keystrokes and the movement would have been staccato instead of smooth. Obviously much thought went into adopting this method of coupling.

The idea of the trigger mechanism is that as the inner lever of the Segment Lever pair goes down the top of the bell crank moves towards the front of the machine. A horizontal retainer resting on an upper protrusion or vertical arm (A) on the trigger moves to teeter ready to drop behind the horizontal retainer. As the key takes down the Segment levers as a pair a level arm on the trigger which rests on the pair allowing the front of the trigger to dip. This allows the trigger itself to be latched forward as the upper arm (A) which was teetering on the edge of the horizontal retainer falling forward to latch there. As the front end of the trigger is coupled via a ‘Y’ yoke to a spring loaded pair, hook and pawl, called a pinion ratchet reverse lock and an accumulator locking hook the mechanism is now positioned to assure that if the operation is not completed the accumulator will be held from coming up and adding an amount into the carrying gear spring. As the shortened operation would have resulted in an error the trigger is made to hit a square shaft which in turn releases a saddle latch under the segment levers that are in the up position. This gives an immediate physical signal to the operator that an error was about to take place, as all the keys that completed the normal operation were held from starting a new operation.

Envisage the Segment Lever in the column that would produce an error is held in limbo as the Accumulator is held either by the accumulator locking hook or pinion ratchet reverse lock. If should be realised that the error has been flagged to the operators by the other keys locking however the column containing the error is allowed to continue its stroke. On completion the error is corrected. The Segment Lever on coming up kicks the horizontal retainer from behind the trigger and the bell crank is free to pull the mechanism back to its normal position. As this stage the column that had had an abnormal operation will also lock as the saddle latch drops under the hook on the segment lever to stops it next operation.

The trained operator is immediately aware of the circumstance of the error and presses the error control button (red on earlier models, green/gray on later models) this pushes the square shaft back under its delicately balanced pawl and removed all the saddle latches from under other column. The operator can then continue with the operation.

In a later model called the 3D11 this operation was further refined and the trigger mechanism was reset in a semi automatic manner. This allowed the operator to continue the operation without removing her/his hands from the key tops. The operation was essentially as I described however the saddle latch was of a different design and the option to use automatic reset was selected by a reset mechanism attached to a key designated as a multiplication/divide) key.

All models from the Model ‘J’ onwards had a Latch and latch lifter added to them to stops the accumulators moving out of place when a clearance operation was in force. This latch and latch lifter was not on the model ‘H’. On the illustration of the carrying gear (earlier page) you will note the carrying detent guard was attached to the gear with studs. On the model ‘H’ the guard was flat metal and spot welded. It tended to crack at the bend and one assumes this is the reason for the improvement.


MACHINES DESIGNED FOR OPERATIONS IN STERLING CURRENCIES

The decimal machines were obviously designed to fulfill the requirement of countries using such currencies and weights and measures. In countries requiring a different radix, for example English Pounds, Shilling and Pence and Tons, Hundred Weights and Stones ways had to be conceived of to accommodate such differences. That is, at least, if one wanted a global market for the product.

Sterling currency machines required quarter pence (or farthings, as they were still legal currency in those days; radix four). Pence, Radix twelve and as Ten shillings resulted from a ten shilling note and there being two ten shilling notes to the pound, a radix two was required for the Ten Shilling Column.

This was a tall order considering the mechanics of the system. Felt & Tarrant produced Accumulators, Carrying Gears, Escapements etc all geared to accommodate the varying radixes required. The Ten Shilling (or radix 2). A concept that presents no difficulty with the binary language of computers presented a very special difficulty with the mechanical machines. The difficulty was not in the production of a machine to work as a radix two machine but in the mixing and matching of the various radixes. A unique solution was adopted consisting of an auxiliary gear and matching accumulator. This effectively made each one added into the ten shilling accumulator rotate the carrying gear a quarter rotation. You will recall the carry gear only did half a rotation to represent a storage on ten digits. This two quarter rotations resulted in the conversion of two ten shilling notes in to one pound.

In the wake of this concept came a bonus using the fours complement 1 your could subtract or divide farthings. Using the 12’s complement 1 you could subtract and divide pence and using the 2’s compliment 1 the ten shilling could be used in dynamic calculations associated with such machines if the operators were skilled in such operations. In sterling currency countries all operators were trained in these skills.

I understand that machines in Tons, Hundredweights, Stones and Pounds existed in the Melbourne Railyards. I vaguely recollect these machines during service visits however, thankfully, I was never called upon to repair any of them. They were large machines actually built into a metal desk and often had many columns more than the average machine. To stop the frames distorting they had a huge solid base of about one centimeter thickness. I often wonder what happened to such unique machines.

TIPS AND TRICKS TO KEEP THEM WORKING

A/ A prime requisite of mechanical equipment with very close tolerance pivot points is regular oiling with a suitable oil of the correct viscosity. Special lubricant we recommended by the manufacturer to service these units. Felt & Tarrant produced a very light oil that would not congeal as long as one adhered to a regular service routine. In Australia we used Shell Alavania #3 on motor bearings, Shell Ossagol V on sliding parts such as levers, pawl ends and detents. Shell Risella 917 for pivot points and where greasing bearing would result in sluggish action and for Locking Dogs’, Shafts and special pivot points we used an extremely light oil not unlike modern sewing machine oil. I am very conscious of the necessity of continuously lubricating shafts and pivot points with this very light, sewing machine consistency oil and failure to do this on a regular basis soon resulted in machines becoming so sluggish as to be unusable by skilled operators. Modern oils are possibly superior to the oils and lubricants available in the fifties and sixties, indeed I very much doubt if many of these older lubricant are still available. Regular lubrication will still be essential if one is to keep the machines in top working order. Important lubrication points were.

A Down the side of the accumulator ratchet to lubricate the pawl.
B On the escapement to allow it to work along the shaft.
C On the pivot studs of the Bell Crank.
D Each side of the Numeral wheel to spread along the numeral wheel shaft.
E MOST IMPORTANTLY on the locking dog pivot post. If this starts to bind it becomes necessary to wash the old oil out from the pivot post with ammonia. Thoroughly flush out the residual ammonia and thoroughly re-lubricate the pivot stud. In severe cases it became necessary to put a spring hook into the locking dog spring hole and rock the dog to and from to work its bearing loose.

NOTE : Many shafts hold springs as well as acting as pivot points for moving parts. When removing any shaft for cleaning with smooth wet and dry emery cloth it is necessary to follow through with a new shaft of the same diameter; taking care not to drop any springs. Cleaning congealed oil off shafts was a part of regular maintenance.

When removing a rock-frame for service it helps to fit a safety shaft in the holes provided through 56-C. See the Carrying Mechanism drawing. The intermediate gear shaft is then removed as any parts on the shaft either dropped out of place or removed. The rock-frame can then be removed by toggling it forward through the front of the frame.

On electric machines the speed of the motor is critical. It should be 180 rpm at the drive shaft. Do not touch the end of the armature shaft with a metal rev counter; they can be ‘live’. Always disconnect the power before adjusting the governor contacts ( at least unless you have the correct tool to make the adjustment).

I have seen reference to washing machines out in hot water containing detergent and forcing this into machines under pressure. If you would care to remove the keyboard of one of these machines you would see why this is a No No. Some springs are very fine and would be damaged by the pressure. Accumulator pawl springs would take time to dry out and would allow the spring and pawl to rust. I could go on a length with reasons why one should not take this approach. During decimal conversion we stripped down four machines a day. Converted them to decimal from sterling and reassembled them. Prior to working on them we removed gunk with a mixture of Shellite and Kerosene. Afterwards all shaft were remove one at a time, by the use of follow through rods. The shafts were sanded down with fine emery. Re-oiled and put back in the same position as they were removed. The accumulators and bearings were oiled with light oil and the triggers adjusted by the use of one key in each column before the machines were re-assembled. After conversion there was so little work required on machines that we spent time learning other things.

Shafts requiring particular work were the Carrying gear shaft. The Intermediate gear shaft and the Numeral wheel shaft. Locking dogs and accumulators also required particular care.

This document may be copied and released for Educational purposes. It may be copied and given to enthusiasts interested in the mechanics and remembered history of the product. It may not be produced in whole or part for the purpose of profit.