Notes on Workshop Techniques


INDEX

THE LATHE

1. Initial Setup

The best lathe in the world is going to function poorly unless it's correctly setup in the first instance. Even a new lathe will not cut parallel unless it's levelled properly, and the surface finish that can be achieved will be much improved by reducing vibrations transmitted to the work and tool from the motor and lathe gearing. When I got my first lathe (a Myford ML10) for a long time I put up with it not turning quite parallel, not realising it was an easy situation to correct. Even if your own lathe has been installed for some considerable time it's worth going through the test procedures to check it's alignment. None of the procedures involved are particularly complex, and you don't require expensive tools to get a good end result.

The first thing to look to is the base itself. Ideally, a steel cabinate stand firmly bolted to a concrete floor should be used. On this stand are normally placed cast iron mounting jacks (raising blocks) which make adjustments fairly straight forward. This combination provides the most stable platform and one which, once adjustments have been made, will ensure that the lathe retains the accuracy of it's setting. All is not lost however if, like me, restricted space means your machine has to be mounted on a wooden bench. Provided the foundation is firm - this means that the wooden bench is mounted on a solid floor - very acceptable results can still be achieved. The only caveat is that regular checks will need to be made to ensure that the initial settings don't change. You should use seasoned wood for the bench, sturdy legs at least 3" x 3" and no more than 24" apart, well braced, and with a 1-1/2" to 2" thick planed top. All exposed wood surfaces should receive several coats of an oil-resistant protective varnish (I used 5 coats of polyurethane). New wood is likely to gradually dry out in the workshop atmosphere, warping as a result and changing the delicate lathe settings as it does so. Also, if it's a new bench and you put something heavy on it it's going to take quite some time to 'settle' into place (the structure being quite flexible), so you might have to make checks at weekly intervals initially until you are sure changes have stopped happening. Similarly, if the bench is subjected to wide variations in temperature and/or humidity then parts will expand or contract, and the (mis-)alignment of the lathe will follow these changes. For this reason wooden flooring is to be avoided if at all possible, and if it's not use a steel cabinate stand. Therefore, to some extent climate will influence whether or not a wooden bench is a viable option.

If your lathe did not come with mounting jacks it's quite possible to make your own provided there are mounting holes in the feet to accept them. First trick is to measure the exact distance between the hole centers of the pair at the tailstock end, and also the pair at the headstock end. I would then use a substantial metal block to mount each pair of studs (I should be looking at something like 12" x 4" x 1/2" mild steel section for each block). Mark off the center line and drill and tap 3/8 BSF for the two studs at their correct center locations. Keep these tapped holes square to the base! Examine your bench mounting area and select the position for 4 holes on each base to accept coach bolts which will be used to secure the base to the bench top. Drill these holes now, and also the holes in the bench - you can make these holes a little over-size to allow for some adjustment. (The lathe will have to be moved out of the way to do this). Make the studs from 5/8" or 3/4" AF hex steel, pieces about 3" long will be convenient to work with. First, make the nuts up. You will need (for each stud) the following items: an adusting nut with sufficient meat to get a good spanner on it (say, 1/2" long), a steel support washer 3/32" thick and 3/4" diameter, a bedding washer made from aluminium or soft brass 1/16" thick and 3/4" diameter, and finally a locking nut and washer. For the studs, turn the major part of the diameter (about 2" length) down to the nominal size for the mounting holes (probably 5/16" or 3/8" diameter) then screwcut at that diameter 32 TPI to accept the adjusting nut you have already made. This nut needs to be a good fit without slop. Turn around in the chuck and thread the other end 3/8" BSF for a length of 1/2" leaving about 1/4" of hex to form a shoulder and allow you to screw them into the bases. Screw the 4 studs into the two bases with a drop of Loctite on the threads and tighten them up with a spanner. Mount the bases on the bench and bolt them well down, use wide, thick washers on the underside if the bench top is made of wood. Screw on the 4 adjusting screws topped by a steel washer and bedding washer on each. Then you need to lower the lathe onto the studs - be careful, use a block and tackle, a pair of trolley jacks, or several pairs of helping hands. That's it - now you can use the fine thread of the adjusting nuts to ajust height and do away with bits of shimstock. Just remember to initially tighten the locking nut to seat the rough cast base of the lathe into the bedding washer, then loosen it again before making final adjustments. An alternative to having the coach bolts pass through the base between the mounting holes (and therefore beneath the lathe) is to arrange for these holes to be outside of the studs. It's going to look less tidy but at least you can then slip the lathe over the studs and then slide it into position before bolting it to the bench top. The choice is yours.

Lathe Jacking Studs

The process of setting the lathe up is a logical one, and the first step is to check that the foundation is as level as you can possibly get it. For a steel cabinate this means adjusting the screws at the bottom until the base is level when measured by a 3ft spirit level across the mounting blocks. In the case of a wooden bench you might either be using similar cast mounting blocks, or (as in my case) 2" thick slabs of hardwood screwed and glued to the bench surface. The hardwood blocks I carefully planed and sanded until they were as level as I could measure with the tools at hand. If you are not using jacking blocks you need to take more care as it's more difficult to correct for slight inaccuracies later.

The lathe is now to be mounted on the prepared base. Chuck a length of 3/4" or 1" silver steel (or precision ground mild steel - something with a nice smooth finish to it) about 10" long. Apply a DTI to the top of the bar at the headstock end and rotate the chuck until the maximum and minimum deflection can be identified. Turn the chuck so the median is set at the top and then move the DTI down to the free end of the bar. Hopefully the reading will be the same - if not this will be dealt with later. Tighten the mounting bolts sequentially, like you were tightening up cylinder-head bolts, to an even torque and watch that the DTI does not move - if it does it indicates you are distorting the bed by bolting down to a slightly uneven surface. If the movement is gross (10 thou or greater) you will need to correct for that now before continuing. First, try adjusting the tension applied to different bolts to see if just one corner is responsible. If the reading suggests the test bar is moving *up* then you are pulling the tail of the lathe down, so either use the jack screws to raise the tail a bit or put a temporary shim in place under the tailstock end. If the reading suggests the opposite, then compensate accordingly (you can't drop the tail unless using jacking screws so you will need to shim the headstock end). When the readings show less than a couple of thou movement (and it would be better to get much closer than this - ideally zero) you can move on to the next stage which will fine-tune the setup.

Using a metal rod to check for distortion whilst bolting down.

The next job is to correct any small amount of twist, and in the process ensure that the lathe can turn truly parallel. In the professional machine shop this job is performed using extremely sensitive levels (accurate to 0.003" in 10 inches) which the amateur is highly unlikely to possess. We will therefore rely on measurements taken on actual turned work, the parallelism of which is itself sensitive to the bed alignment, and which we can accurately measure with a simple micrometer. There is a fundamental difference between the two methods - using either a level or the parallelism of turned work. The first assumes the lathe is basically accurate and if levelled on a true surface will be setup correctly. The second assumes no such thing and positive steps are taken to ensure that the lathe turns parallel regardless of whether it's set exactly level or not. This will become clear later. A MT shank parallel test bar is a very handy item for testing the alignment of head and tailstocks, but for the moment we can do without one, though it may be required later if other problems show themselves. What follows assumes the lathe is in generally good condition without serious wear in headstock bearings and that the slideway gib strips have been adjusted correctly. If you are in any doubt about this you should adjust or replace worn parts before going any further (see here for making routine adjustments).

Step one: take an 8" length of some free-turning material about 1" dia., (F/C mild steel or aluminium alloy) and grip it in the chuck leaving about 6" or so outstanding. Turn the centre portion down leaving a ring about 1/2" wide at the headstock end and a second ring at the tailstock end. Now, using a *very* sharp knife tool (plenty of top rake with a small flat on the end to produce a good finish) take a thin cut, no more than a couple of thou, across the ring near the headstock, then wind the carriage down the bed and take a cut across the second ring without changing the setting. Take both cuts using the same direction of carriage travel. Repeat if the first two cuts fail to produce a clean turned surface on both rings. You must not use tailstock support for this job so the cut needs to be very fine to avoid the job 'springing'. Now, carefully measure the diameters of the two rings. If the two match then your mandrel is in good alignment with the bed and you can move onto the next step. More likely, there will be a difference between them of a few thou, and this difference is a measure of the misalignment lathe bed and spindle axis. All things being equal, the most likely source (in a good quality lathe anyway) is going to be the result of twist in the bed caused by the uneven seating.

2 ring method for correcting bed twist (winding)

[NOTE: I should add at this point that the errors one sees in cheap imported lathes are potentially going to be different than those seen in good quality lathes. In the Myford (for example) you can garauntee that from new the mandrel and bed are in perfect alignment, that's what you pay for, and any divergence from this alignment is going to be due to lathe bed twist. In cheaper lathes anything *might* be misaligned - including the spindle to lathe bed alignment - but you will just have to assume that any errors can be corrected by adjusting the bed seating. However, it must be said that a poorly constructed lathe will be *impossible* to set up accurately so you may have to live with a compromise setting.]

To correct twist in the bed is a simple matter of adjusting the jacking screws on the raising blocks, or by placing shims beneath the feet of the lathe - such adjustments being carried out at the tailstock end. If the diameter of the ring nearest the tailstock is greater than that nearest the headstock then you will need to ADD shims to the FRONT foot at the tailstock end. If the diameter was smaller you will need to ADD shims to the REAR foot at the tailstock end. Slacken off *all* the bolts and use a lever to raise the tailstock end slightly in order to slip the shim in place. Use a 10 thou shim to start with and work up or down from there. Re-tighten all the bolts to the same torque and take another fine cut across the two rings. If the difference has halved you will know you need to add another piece of shim the same size, and proportionally more or less depending on the changes you measure. I stopped when I got within 0.0003" parallel over a 6" length (about 6 pieces of shim later...)

Next job, while you are set up for it, is to align the tailstock with the headstock. First job - carefully centre the end of the test bar while still held in the chuck. Wind the tailstock barrel back into it's casting as far as it will go and knock in a dead centre. Support the end of the bar with the centre and take a fine cut across both rings once more. If both measure the same diameter then the tailstock is set correctly at zero set-over. If not, you will need to use the adjusting screws to move the tailstock back into alignment. If the diameter of the ring furthest from the chuck is now larger you will need to move the tailstock towards the front of the lathe, and vice-versa if it's smaller. The previous procedure does not tell the whole story though, and you will need to check whether the tailstock barrel is actually parallel with the headstock spindle. To do this wind the tailstock barrel out as far as it will go and test again. Only if both rings *still* measure the same can you be sure your tailstock barrel is correctly aligned. Again, if this is not the case then adjustments will have to be made. Depending on make this may or may not be easy. It may be a simple matter of adjusting gib strips to twist the body around back into alignment.

2 ring method for aligning tailstock.

As I said earlier, in the professional machine shop the lathe bed would have been set truly level first by means of a sensitive level. A parallel test bar would then be inserted in the headstock bore and this would immediately show whether the headstock was misaligned or not. The test bar is of less value to us because we have already set the lathe bed according to whether or not a workpiece is turned parallel. In our case, if the headstock *were* slightly misaligned (say it was pointing to one side by a small amount), and the lathe bed was adjusted to turn truly parallel using the fine cuts on the test workpiece, then what you have done is to inadvertently impart a twist *into* the lathe bed to correct for this. This is not a good situation but is perhaps a reasonable compromise. Consider: In a good quality lathe we have to assume the manufacturers know what they doing and that the headstock alignment is accurate in the first instance. This can be taken for granted in a Myford Super 7 short of the thing being run over by a steamroller (not my quote...) In a poor quality lathe we cannot take this for granted but there is likely little the amateur can do to accurately reset the headstock anyway (which may be impossible in some lathes with the headstock and bed forming parts of a single casting). If I found a headstock alignment fault in a new imported lathe it would go straight back to dealers. In good quality but worn lathes any inaccuracy of the headstock is more likely the result of worn spindle bearings and these should be checked, correction will consist either of replacement, adjustment and scraping, or re-boring on a jig boring machine. So, for the amateur with limited facilities, the above procedure will leave the lathe setup in the best compromise alignment.

Having gone through the basic method of installing the lathe, we can now turn our attention to those things that will help it to perform at it's best. Vibration and noise are not only annoying to the operator but have a detrimental effect on the workpiece. If vibration gets really bad regular chatter marks will appear on the finish of the turned or faced work. Easiest way to check is to simply rest your hand on the bed with the lathe running - you can *feel* both sharp vibrations (metal-metal contact somewhere) and lower frequency out-of-balance vibrations (pulleys out of balance or, more likely, worn or over-tight drive belts). The source of the vibration is either the motor mounting (single phase motors are worse than 3-phase for being vibration prone) or somewhere within the drive train. Modern new lathes have motors mounted in cradles with rubber shock-absorbing mounts, and these should be particularly checked in older lathes. Look for worn rubber mounts resulting in metal-metal contact and replace if worn. Check the belt guards are not knocking against something. Check the alignment of the pulleys with a straight edge such that you are satisfied they are not eccentric or wobbling side to side, and check also that the drive belts line up correctly with each pair of pulleys. Examine the condition of the drive belts - over-tight or worn belts can both cause and transmit vibration - particularly at high speed, and make sure the tensioning is correct - a problem with short radius V-belts which require high tension if they are not to slip under load. Belts left for years under tension in one position can take a 'set' resulting in vibration, so if you buy an old machine that has sat around for a considerable period consider fitting new belts. If your belts do need changing consider a poly-V belt conversion ('flat' belts with multiple small 'V's). Whilst my new Super 7 starts with a bit of a clunk it runs whisper quiet and vibration-free. When setting up a fine-feed drive train with change wheels make sure there is sufficient backlash between each pair of gears, bottoming the teeth of a pair of gears is going to cause vibration and may even damage the gears. You don't need to coat the gears in grease, a thin coat of machine oil is all that is required.

Return to top

(c) Chris Heapy 1996.


Back to previous page