Conversion of the prototype Myford lathe

This page records the steps in the development of a prototype MC-KIT application, in this case a lathe. Similar principles would apply to the conversion of any machine tool, basically involving the replacement of manual handwheels and gears with electric motors and shaft-encoders. Depending on the design of the machine, some imagination may be needed to devise a suitable way of mounting these components. The solutions shown here are not necessarily the only possible way, but these have been found to work well in practice.

This Myford ML7 lathe forms part of the JCE mechanical workshop, serving as a main prototype for the MC-KIT development project for machine-tool conversion. It is a typical lathe of the 1960's, which has already reached its 30th birthday. It is a medium-small lathe (diameter up to 7 inches), which has proved very popular in workshops specializing in small products and precision-machined parts, both amateur and professional, in the UK, and has been exported world-wide in large numbers, particularly to English-speaking countries. This example shows the simple and robust construction of the machines of this period. With many years of service behind it, there has been no reduction in the quality and precision of the products it is capable of turning - with care, down to less that 0.001" (0.01mm) - however, it is of course completely manually controlled. It thus serves well as a base for developing the computerized control system.

The following pictures show the process of converting the mechanical parts. The electrical and computer parts form the basis of the MC-KIT system, which are intended for use on a variety of machine tools where the mechanical conversion is done in a similar way to that shown: namely, replacing manual handwheels and gears with electric motors and shaft encoders. In each case, the main problem is finding a way of mounting the motor / encoder. This will vary from one machine to another, but the following illustrates the general principles.

BEFORE MODIFICATION

The lathe before conversion, showing the traditional manual controls.

BEFORE MODIFICATION

View of the headstock showing the gear-train and changewheels of the longitudinal leadscrew drive. The first task is to replace these gears by an electric motor, which will allow the computer to control the leadscrew feed, and also mount shaft encoders both on the headstock spindle, and on the leadscrew spindle. This will allow the computer to measure the position of both spindles, and replace all the mechanical gearing with the much more flexible electronic control.

Here, all the mechanical leadscrew gearing has been removed. The lathe is still functional, only now the leadscrew has to be turned by hand, using the graduated handwheel at the tail end (right-hand end) of the lathe. This is important in this case - since we do not have another lathe available, we will go on using this one to make some of the parts (fortunately fairly simple) of the new system. Thus the development will have to take place by well-planned steps so as to always have the basic lathe functions available (a bit like modifying an aircraft while it is still flying!). Of course, workshops equipped with more than one lathe can use one to make the parts for the other one being converted - but not every workshop has more than one, so it is useful to see if it is possible for one lathe to make its own parts for its own conversion.

Measurement of the handwheel torques

If the types of electric motor, necessary for controlling the longitudinal and cross feedscrews, have not already been determined, we need to check the torque necessary to turn the handwheels under typical conditions. This will allow suitable sizes of motor to be selected. This torque is very variable, it depends on the general state of the machine adjustment and lubrication, and varies from one moment to the next while the machine is working. However, it must fall within certain limits, bearing in mind that the handwheel is designed to be turned by the human hand. The objective if to measure the typical torque which is sufficient to turn the handwheel, then choose a size of motor which with its reduction gearing (both in the motor itself and also external gearing e.g. belt drive) will give at least 4 times this torque when fed with full voltage. There must be a reserve of torque so as to be sure that the computer will always be able to turn the motor to the calculated speed and position, under normal conditions - normally the motor would be working well below its full capacity. In the case that such reserve of torque should be insufficient, probably the lathe needs maintenance or lubrication, or is not being operated correctly - e.g. cutting too deep. There is no advantage in using a very oversized motor to force the movement, this could cause other problems. Rule - if it doesn't want to turn under reasonable torque (human hand strength), see what the problem is, don't force it until something breaks!

Feed Torque

Measurement of torque on cross-feed handwheel.

This is done simply by pulling with a spring balance on a steel ruler (or bar of some kind) attached to the handwheel, thus measuring the force necessary to make the handwheel start to turn, at a certain distance from the centre of rotation - thus giving the torque (force x distance). This is not an accurate measurement, rather it is to get an idea of the maximum torque needed, so as to choose an adequate motor type from the motor manufacturer's performance tables.

Lead Torque

Measurement of torque on leadscrew handwheel.

This is done in exactly the same way as in the case of the cross-feed handwheel, but this time the ruler is attached to the leadscrew handwheel. The force was a little greater in this case, suggesting that a larger type of motor be used for the longitudinal leadscrew.

To give a rough idea for this lathe, the typical torques to start the handwheels turning under normal conditions were measured as: 8.85 lbs.ins (1 Nm - Newton-metre) for the cross-feed, approx. 18 lbs.ins (2 Nm) for the longitudinal leadscrew. These values seem typical for a small lathe, perhaps larger values would be expected for a large machine, but not very much greater - otherwise it would mean the handwheels were difficult to turn by hand.

Selection of motors for experimenting.

The following factors were taken into consideration: Desired maximum speed of the handwheel, between 1 and 2 turns per second (60 - 120 r.p.m.). This indicates a motor of up to 6000 r.p.m. with a total reduction gearing of, say, approx 60, of which the worm reduction gearing (integral with the motor) is 25:1, with a further reduction of 2.5:1 in the toothed belt drive to the leadscrew. For example, for the leadscrew a motor was chosen with a maximum torque at the worm-gear output of approx. 27 lbs.ins (3 Nm) (at the nominal 4000 r.p.m.), which gives approx.+ 27 x 2.5 = 67 lbs.ins (3 x 2.5 = 7.5 Nm) at the leadscrew (whose typical torque was measured at 18 lbs.ins (2 Nm)). This is at the nominal motor voltage - a voltage increase of 50% above nominal is allowed for short periods, which give a potential torque of approx. 88.5 lbs.ins (10 Nm), thus giving the desired reserve of torque.

In practice, even before the motor was mounted, it was tested at half the nominal voltage, and the leadscrew turned easily even though the motor was held by hand so as to maintain tension in the drive belt, thus giving reason to be optimistic that the final system would work. In fact, as a rough guide, after mounting the motor as shown below it proved impossible to stop the leadscrew turning by using hand pressure on the manual handwheel, corresponding to a much greater pressure than would ever be used in practice even for the heaviest job, and just about at the limit of human hand strength! Basically, the above approximate calculations are useful as a starting-point, but the final aim is simply to have enough motor torque to keep the leadscrew turning easily under any normal conditions.

Mounting the leadscrew motor / shaft encoder

Not having available any drawings or dimensions for the lathe, the best way is to start with a metal mounting plate of approximately the right size and machine it by stages to fit onto the lathe.

Plate cutting

The starting point is simply an aluminium plate 1/4 inch (6.35 mm) thick, and an electric jigsaw. (the plate could have been mild steel, this does not matter - this plate just happened to be in stock). It will form a sufficiently rigid base for this application. It was cut a little larger than the final dimensions, to allow for finishing after mounting.

In order to mount the plate, it is necessary to make various holes of different diameters to fit onto the available mounting points - the studs and spindles of the headstock. Due to the difficulty of measuring exact positions, even with a vernier calliper, all at the same time, it was decided to start with the large hole that fits over the bearing of the headstock spindle. This being done with precision, it would serve as a reference for the other holes. Really, a milling machine would have been best to make this large hole. Not having a milling machine available, the improvisation shown was used - the plate bolted onto the lathe carriage, using the Myford vertical slide accessory. The boring tool was fixed into the 4-jaw chuck (which allowed the radial position of the tool to be adjusted exactly, with much patience) - thus inverting the normal arrangement of machining with a lathe. The hole, previously roughed out with the jigsaw, was finally machined to the desired tolerance of 0.010" (0.25 mm) larger than the external diameter of the headstock bearing shell.

To exactly mark the position of the remaining holes, it is enough to turn a point on a screw which fits a stud hole, screw it into the hole, and .....

Mark hole position
place the plate, with the previously machined hole fitting over the headstock spindle bearing shell, and using studs in holes already made, then give a light tap over the pointed screw shown above (rubber hammer, please!) - the pointed screw makes a centre mark in the new position, which serves as a guide for drilling (naturally, starting with a small diameter drill and increasing the drill diameter to enlarge the hole diameter up to the final value). In this way, with care, the plate is made to fit exactly over all the fixing holes, with only small tolerances being necessary (of the order of 0.008" (0.2 mm) drill size over stud size). Packing washers are used between the lathe body and the mounting plate, to adjust this plate exactly at right angles to the headstock axis (visual alignment with setsquare is sufficient, extreme accuracy not required).

Headstock spindle encoder

This encoder, which serves to measure the angular position and speed of the headstock spindle, must be of a form which leaves free the hollow centre of the spindle (which is used when feeding through material in the form of long rods). This requirement is not met by most commercially available encoders, which couple to a small shaft which would block the hole through the spindle. Therefore, a codewheel is made to fit onto the outside of the spindle. This could be in the form of a thin disc mounted on a hub, but in this case it appeared simpler to machine the whole thing from a solid disc of aluminium alloy. After turning the disc, leaving a thin rim standing out, 48 slots were machined in the circumference, to be used to interrupt the light beams of optical sensors, thus generating the electronic pulses for measuring the rotation of the spindle. Why 48? Because it so happened that 48 slots of a convenient size for an easily available optical sensor (and for which the milling cutter was already available in the workshop), fitted exactly into the circumference of the disc. Any number could be used since the computer will perform the necessary calculations, but 48 also has the advantage of being exactly divisible by 2,3,4,6,8,12 which is very useful to allow exact positioning at commonly met angles. These slots are being machined in the picture, using the Myford angular divider accessory.

Code wheel

The encoder disc, machined from a solid aluminium alloy block such as that on the left. Note the 2 grub-screw holes in the hub, for mounting on the headstock spindle in the position previously occupied by the gear wheel. The centre hole is machined to exactly fit the heastock spindle diameter.

Mounting the motor - 1st stage

Now, with the mounting plate fixed onto the lathe, and also with the headstock encoder disc and longitudinal leadscrew toothed pulley being mounted (this latter being a standard stock item, bored out to fit exactly on the leadscrew extension), one can see the final arrangement of the components, and evaluate the best position for the motor, experimenting with different lengths of toothed belt, thus planning the mounting components.

A triangular motor plate is then made to hold the motor, which is screwed on by its worm gearbox. Three pillars are made to hold this plate at the correct distance from the mounting plate, so the toothed belt runs straight between the motor pulley and leadscrew. The pillars have threaded holes at each end (the lathe is still operational enough to make the pillars!). The two top holes in the motor plate are elongated, to allow rotation around the bottom hole - this allows easy installation and adjustment of the toothed belt.

Leadscrew and Headstock spindle encoder mounting

In order to mount the leadscrew shaft encoder, a second triangular plate is made, two of its holes sharing the same pillars as the motor plate, the third having another pillar (bottom right).
Here, the two opto-sensors have been mounted to read the Headstock Spindle Encoder. Adjustment is provided, so when correctly positioned, they generate 2 square waves, 90° out of phase, to read both speed and direction of the spindle. (They are connected here to a temporary circuit board which in turn is connected to the computer for development purposes).

Under construction - details of shaft encoder mountings to follow.

Back to Technical home page