The 'drive' refers to the system for transferring power from the hand crank to the cams, printer and carry axes.
Power is transferred from the hand crank to the cam stack drive shaft via bevel gears at the top of the cam shaft (163). A ratchet is located to the right of the drive shaft bearing and ensures that the drive cannot be reversed. The ratchet drive is coupled to the bevel gear via a rocker clutch operated by a gut cable running from the printer. The purpose of the clutch is to disengage the drive immediately a page of printed results is completed, without any risk of overrun. The intention is to halt the calculation at an exact point in the calculating cycle to allow the stock receiving the printed impressions to be renewed. Activating the clutch uncouples the drive. The engine halts and arrests the figure wheels is a state ready for the commencement of the next run of results.
Each of the seven carry axes is driven by a bevel gears pinned to a horizontal shaft driven by an intermittent drive mechanism (165 lower right). The bevel drive for the first odd axis is omitted as the seventh difference carry mechanism was regarded as redundant The shaft coupling the bevel gear and the intermittent drive mechanism is hollow. The shaft from the handle crank passes through this hollow drive socket and is supported by the bearing on the right hand lower end frame.
The printer drive shaft which runs the length of the underside of the engine is driven by a bevel gear at the lower end of the cam shaft.
The crank consists of a drive arm and a drive handle. The handle is made from lignum vitae which, though not specified, was regarded as typical. The handle is free to turn on a shaft which is peened over at each end. A thrust washer is fitted at the outer end. No end elevation of the crank is given so the intended shape of the arm is not known. The chosen shape was taken as characteristic. The length of the arm is taken from 163 and 164, and the thickness from 159, 163 and 164.
The ratchet prevents the engine being driven in reverse which would damage the calculating mechanism. The ratchet allows the operator to start turning from a stationary position best suited for the exertion of controlled effort. The ratchet also allows the operator to nudge the engine to a particular point in the calculation cycle by a series of short pulling movements.
The ratchet mechanism is shown on 159 lower right. It consists of a ratchet wheel and matching pawl housed by a pawl wheel which acts as its casing. A leaf spring fixed to the pawl acts against the inside of the case to provide the return action. The pivoted end of the pawl nestles against a machined pocket in the side wall of the case. The machined pocket takes the drive load rather than the pawl pivot which is a loose fit The pawl is sandwiched in place by the pawl wheel cover which is rebated into the case and secured with counterbored cover screws. (159 shows the case with the cover removed but the fixing screws fitted.)
The counterbores for the pawl wheel cover fixing screws have insufficient clearance from the wall of the casing. The section view (159) shows the cover fixing screws breaking through which is uncharacteristic. Fixing screws were therefore reduced from 5/16" to 1/4" BSW, the head reduced to 7/16" and the length was shortened to avoid breakthrough. The case is cast and machined.
The outer wall of the case was extended to take 56 gear teeth to allow a four-to-one reduction in the drive. The addition of reduction gearing was in response to concern that an operator would be unable to exert sufficient force to turn the engine over. The additional parts for reduction gearing are a 16-teeth pinion (D323), a crank (D322) and a horseshoe mounting (D321) for the pinion. The crank, D322, is a repeat of the pawl wheel crank. The crank is keyed to the pinion, and the drive arm is fixed to the crank with a shoulder screw (D327). The mounting is fixed to the underside of the cam stack upper mounting plate. Using the reduction gear is optional: the handle crank can be removed by undoing the shoulder screw and refixing the drive arm directly to the ratchet wheel crank, as Babbage intended.
Without the reduction gears the engine is driven by turning the handle anticlockwise as seen facing the status plate (otherwise called the chapter disc) and the addition of the reduction gearing reverses the direction of the drive. The main loads on the drive occur at the start of the cycle and at the half-cycle. These are caused by all the figure wheel and sector wheel locks coming out simultaneously. There are thus two short peak loads in one full cycle. Lesser cyclic load occur at the start of the carry portion of the cycle i.e. at 10 degrees CB (72 degrees) and 35 degrees CB (252 degrees).
For the convenience of the operator the handle is positioned at the lowest point of 'its travel for a pulling action at roughly waist height to overcome the start-of-cycle load. Without the reduction gears the half-cycle load would then occur at the top of the arc requiring a pushing action which is awkward and could reasonably have induced Babbage to include a reduction gear of his own. Anadditional advantage of the reduction gearing is that the full and half-cycle loads both occur with the handle at the bottom of its travel and can be overcome by a pulling action which is more convenient than with a single-turn cycle.
The clutch is shown in end elevation (159 lower left) and in section (159 centre left). The clutch rocker has a short lug and an extended lug protruding from the ends of the basic oval shape of the rocker and which rock in slots machined into the rim which stands proud of the clutch body. The extended lug protrudes beyond the clutch body and, with the clutch engaged, inserts between a pair of lugs on the clutch bevel. In 159 (centre left) the clutch is engaged with the clutch rocker shown vertical and disengaged in the inclined position.
To disengage the drive, the scoop lever, operated from cable from the printer, positions the scoop cam (shown as dotted position in 159 centre left) so as to drive the extended lug clear of the driven lug when next it traverses the bottom of its trajectory. The extended lug engages with the inclined face of the scoop cam, drives the scoop lever fully home (position shown slotted on 159 centre left), and is itself driven out of engagement with the driven lug so breaking the drive.
Clutch Operation - Difficulties
The clutch rocker is likely to overrun the scoop cam after disengagement following the sudden release of load. Reengaging the clutch is problematic with the arrangement as drawn. To reengage the drive the clutch rocker is rotated to bring the extended lug opposite the bevel lugs with a view to reinsertion. However, with the rocker out of engagement (inclined position) the extended lug clears the first of the two bevel lugs but is prevented from reengaging by the scoop cam. If the scoop lever is reset (moved to the position with the scoop cam shown solid) re-engagement is similarly obstructed. Setting the rocker to the engaged position (vertical in 159 centre left) while outside the two bevel lugs causes other difficulties. As the rocker approaches the scoop cam position the extended lug will fail to clear the bevel lug and will engage with the outside face of the trailing lug (the trailing lug is the upper of the two bevel lugs as shown on 159 lower left and is the first one met by the extended lug).
If Babbage had intended the bevel to be driven from the outer face of the lug then it is likely that he would have drawn a radial face to ensure a flat bearing surface. Drawn as it is with a square shoulder, the bearing surface is a sharp edge which is uncharacteristic. A further difficulty driving the engine from the outer face of the bevel lug is that the timing of subsequent disengagements would be delayed i.e. the engine would halt in a slightly more advanced part of the calculating cycle because the extended lug would be lagging behind its intended position between the two bevel lugs. The sequence for re-engagement would be to rotate the rocker to clear the scoop cam, reset the scoop lever, reset the rocker into the engaged position while clear of the bevel lugs, rotate the rocker till the extended lug engages with the outer face of the trailing bevel lug, drive the bevel clear of the scoop cam area, halt the drive, disengage the rocker and reengage the extended lug between the two bevel lugs, continue with the calculation run. This an unlikely sequence and contra-indicated by the square shoulder of the leading bevel lug. The omission of provision for reengaging the clutch seems to be an oversight.
Since the ratchet ensures that the drive is unidirectional the trailing bevel lug is redundant. The only situation in which the trailing lug would come into play would be to reverse the drive and there is no means in evidence for disabling the ratchet to allow this. Since there is no functional reason for the trailing bevel lug this was omitted. With only the leading lug present, reengaging the clutch is simply achieved by returning the rocker to the engaged position once clear of the scoop cam area, resetting the scoop lever and rotating the drive until the extended lug meets the driven lug on the bevel. The single remaining bevel lug was strengthened by making the outer face radial rather than parallel to avoid a weak corner when casting.
During commissioning and debugging it was found that a useful way of unlocking jams was to reverse the drive, and a means of doing this using the hand crank would have been useful. In the event, reversing was accomplished by gripping the cams and backing the engine off by turning the cam stack a short distance in reverse. It is difficult to back the engine off more than a very short amount by gripping and turning the cams because of the load and the awkwardness of the movement. Backing off by hand in this way was considered preferable to disabling the ratchet drive as prolonged reverse drive of inadvertent reverse drive at an inappropriate point in the calculating cycle could damage the calculating mechanism.
The views of the extended lug are inconclusive with respect to the shape of the area that engages with and slides up the scoop cam. The end of the extended lug was radiused to present a larger contact surface with the scoop cam.
The scoop lever is pulled in by a gut cable from the printer and is reset by hand to resume calculation. With the clutch engaged there is no provision for positively locating the scoop lever which rests free against the slack of the cable. The risk of the scoop lever drifting into unwanted engagement or into an end-on collision with the extended lug of the clutch rocker was considered to be real. A counterbalance was added to provide positive resistance to the printer cable. The counterbalance maintains cable tension which prevents the printer cable jumping its pulleys. It also resists the tendency for the scoop lever to be pulled in unintentionally by the weight of the printer cable lever, by any elastic tension in the gut, or by any tendency of the gut to curl. The counterbalance is provided by the addition of a weight suspended over a pulley by a gut cable attached to an eyelet added to the opposite end of the scoop lever. The pulley is fixed to the rear horizontal framing member and is a repeat of the printer cable pulley.
Disengaging the Clutch
The clutch is disengaged by the printer when a page of stereotyping has been completed. 163 (bottom left) shows a cylindrical weight resting on a half-round gutter. The trip lever is linked to the scoop cam by a gut cord which runs over the two pulleys. At the end of a stereotyped page the release of the weight operates the trip lever which pulls the gut cord and sets the scoop lever to disengage the clutch on the next pass of the extended lug (dotted position on 159 centre).
Strictly speaking, 163 is indeterminate with respect to the shape of the weight: it is unclear whether a sphere or cylinder is intended. The representation of the receptacle as a gutter rather than a ring suggests a cylinder. This is confirmed by 166 which shows a freely suspended rod. The view shown on 166 omits the base supports. Drawing these in shows that the right hand end of the cylinder as drawn will foul the base support. The left hand end of the weight, as drawn, would foul the rewind clutch during operation. If a cylindrical weight were to be used it would have to be shortened to avoid fouling.
There is a further concern about using a cylindrical weight There are two basic formats for stereotyped results: generating successive tabular results below one another and starting the second column after completing the first (line-to-line movement): and generating successive entries alongside each other, two to a line (column-to-column movement). In both instances the weight is raised above the clear of the sides of the gutter in its unreleased position. Without the sides of the gutter to constrain the weight it seemed likely that the cylinder would tend to twist and fail to align with the gutter on release.
In view of these two concerns the cylindrical weight was replaced by a spherical bob and the gutter by a ring-shaped hoop. The spherical arrangement allows the weight to rise clear of the ring and relocate on release regardless of any twisting. The use of a spherical bob also solves the problem of the cylindrical weight fouling the base support and rewind clutch.
The lever is shown in elevation only (163) and no further original details are supplied. The lever was made in two parts - the completion trip which includes the ring and ring arm, and the trip lever which has the eyelet for the gut cord. Both are fixed to the same shaft but the angle between the trip lever and the completion trip can be adjusted by means of a grub
screw so as to take up any slack in the gut cord while maintaining the height of the completion trip.
No provision is made in the original drawings for keeping track of or indicating the position in the calculating cycle at any time. Marking the upper-most cam was considered. This was rejected as viewing is partially obstructed by the cam stack framing plates. Instead a status plate or chapter disc (so called because [ask Turvey]) was fixed to the smooth (non-ribbed) side of the phasing wheel which is in direct sight of the operator. A single rotation of the phasing wheel corresponds to one full calculating cycle of the machine. The brass chapter disc is engraved every five divisions and follows Babbage's fifty-divisions-to-the-circle convention. Engraved labels indicate main positions in the cycle: "Full Cycle" (0): "Half Cycle" (25): "Set Odd" (20); and "Set Even" (45). A swan-necked pointer is fixed to the right front upright and acts as a fixed reference. Status indication is essential for the setting up procedure. For example, at 10 and 35 divisions the odd and even axes respectively are zeroised and the setting locks inserted. The chapter disc and pointer provides this necessary position information.
Carry Shaft Drive
The ripple-through or successive carry is performed by a helical arrangement of carry arms keyed to the carry shafts. The carry shafts are driven by bevel gears pinned to the horizontal drive shaft. 163 shows eight carry shaft bevel drives. The first odd carry mechanism (extreme right) was considered redundant, and omitted. The intermittent rotary drive for the carry drive shaft is provided by the mechanism shown on 165 lower right This consists of a large phasing gear with incomplete sets of teeth and a twin tooth drive. The twin tooth drive is fixed proud of the plane of the gear and meshes with the single impact tooth fixed to the register pinion. The register pinion is fixed to a register wheel which has a circular detent for a sprung roller. The spring tube, register arm and roller are shown on to the top of 160, centre right The detent and roller holds the pinion during the idle periods in the correct position for remeshing with the phasing gear at the start of the next cany. There are two cany episodes in one full calculating cycle - one for the odd axes and one for the even. The continuous rotary motion of the phasing gear, driven by the hand crank via the ratchet drive and clutch, is converted into intermittent rotary motion of the register pinion to provide two episodes of carry shaft rotation for each full calculating cycle. The odd and even carry shafts all rotate together though only one set is active during a given episode.
The arrangement on 165 lower right shows the phasing gear and pinion positioned at the start of a carry. The two impact teeth mesh and in turn ensure correct meshing of the drive teeth. The first set of twenty teeth on the phasing gear, drive the register pinion full circle. The last tooth on the pinion is incomplete: it is angulariy truncated to clear the trailing tooth of the phasing gear so as to leave the register pinion in its idle position. The register pinion idles during the toothless portion of the phasing wheel's rotation and the second carry commences when impact teeth next mesh. Each carry occupies 104 degrees of a 360 degree cycle, and is followed by a 76 degree idle between carries.
The function of the impact teeth is to provide correct meshing of the drive teeth when engaging and to allow the use of materials tough and durable enough to withstand repeated impact The phasing gear and register wheel are cast; the impact tooth and twin tooth drive are machined from high-grade (EN16M) steel. The twin tooth drive is fixed only at two positions near the rim of the phasing gear and its cylindrical centre stands clear of the phasing wheel boss . The effect is to provide a form of shock absorbtion through the sprung action along the length of the twin tooth drive.
The impact tooth and twin tooth drive are shown fixed vwth screws though no further details given. To spare the threads of the fixing screws taking the impact the fixing holes were fitted with impact sleeves which have the action of doweling the impact tooth and the twin tooth drive in the register pinion and to the phasing gear respectively. The sleeves are fixed with cheese head screws. (The impact sleeves, D398, were added during the build. Red Book 22.11.90)
Phasing Gear Key Position
It was noticed that the line through the centre of the phasing gear and the midpoint of the keyway deviates by 1 degree from the line bisecting the angle between the lower pair of webs (165 lower right). This is a subtle and curious deviation and its possible significance takes a little interpreting. The phasing gear is driven by the bevel drive and drive socket which is hollow to allow for the drive shaft. Pinning the bevel and phasing gear would be ill-advised and the original drawings show the phasing gear keyed (165). The position of the key for the clutch rocker is shown (159) opposite the extended tug. but no position is shown for the keying between the bevel gear and drive socket. The position for the bevel gear keyway was assumed to be in line with the keyway for the clutch rocker. This assumption was made on the basis that if the position of the bevel gear keyway was other than arbitrary, it would have been positively specified. It was assumed that the keyways in the drive socket for the phasing gear and the bevel gear would be in line given the additional trouble of machining with an offset. So the argument is that the bevel and phasing gears need to be keyed because they cannot be pinned, that the bevel gear keyway is in line with the clutch rocker keyway, and that the drive socket keyways should be in line for convenience of manufacture. Working back from the clutch and drive bevel in this way fixes the position of the key for the phasing gear.
Working from the register pinion end it is evident that the relative positions of the shafts for the phasing gear and the register pinion are fixed by the layout of the calculating mechanism and the drive. In 165, which shows the configuration at the start of a carry, the position of the twin tooth drive is determined by the geometry of the impact teeth and the requirements for correct phasing of the intenmittent carry drive. The twin tooth drive needs a web to fix to for support. This then fixes the orientation of the webs and therefore of the phasing gear. So the orientation of the phasing gear is determined from the register pinion end, and the position of the phasing gear key is determined independently by working back from the clutch and bevel. There is therefore no reason why the centre line of the keyway should coincide with the line bisecting the lower two webs. The fact that the discrepancy is as little as 1" is an additional confusion: there is no reason in principle why it could not have been something more substantial and therefore more noticeable.
In many of the drawings gear teeth are characteristically referenced from the flank of the tooth rather than from the centre or from the intertooth gap. The occurrence of the 1" difference is not helped by the fact that the centre line through the keyway closely coincides with the line taken from the flank of the opposite tooth.
If the keyway in the phasing gear were machined to preserve symmetry i.e. so that the centreline through the keyway bisected the angle of the two lower webs, this would amount to ignoring the 1" offset. The implications of this would be to retard the carry timing which brings the end of the carry part of the cycle critically close to the point at which the locks come in. The 1" displacement of the keyway was strictly adhered to on the assumption that the offset was deliberate and dictated by timing needs.
Lubrication of Cam Shaft Bearings
The cam shaft is held at the top by the upper framing plate journal bearing and is supported by a thrust bearing in the lower framing plate (160 and 163). Detail of the thrust bearing is shown in 160 lower right. The bearing consists of an inverted cup-shaped collar supporting a shoulder on the cam shaft. The collar, which is keyed to the shaft, rests in a annular oil bath located in the lower framing plate. No provision is made for lubricating the cam shaft bearings: access to the annular oil bath is obstructed by the closely spaced cams in the lower section of the cam stack, and the upper bevel gear obstructs access to the upper journal bearing.
Oil is supplied to the bearings through an oiling hole in the upper bevel gear and conveyed along the length of the cam shaft to the thrust bearing through a channel machined into the shaft opposite the driving keyway. Oil collects in an unsighted oil cup fitted to the upper framing plate below the bevel gear. This acts as a reservoir and prevents oil spreading over the upper framing plate. The oiling channel in the shaft starts off shallow and deepens to full depth below the upper framing plate to assist flow. The shallow channel rotates in the oil cup and drains oil into the channel lubricating the upper journal bearing at the same time. Oil drains down the channel and collects in the lower oil bath which forms the lower half of the thrust bearing. Internal overflow lubricates the lower journal bearing. To ensure that oil is properiy distributed over the bearing surface between the thrust collar and the annular bath, two D-shaped cutouts were machined into the thrust face of the collar.
Go to next section Framing
Go to top of Babbage Technical Description