Trying to align an old Lathe

Home Model Engine Machinist Forum

Help Support Home Model Engine Machinist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

goldstar31

Well-Known Member
Joined
Feb 20, 2010
Messages
3,422
Reaction score
1,305
Location
Twixt Tyne and Tees
Jon has an old early Myford. To help alignment problems he has sought my help and in addition, bought a test bar etc.

Somewhere I have a copy of test sheet for a Boxford which is similar to a South Bend( I think or was old)

The first thing that I would do is to make an elephant's foot for the dial gauge which gives a wider bearing than the ball bearing affair -which comes in the box.
It saves all this annoying fumbling.
I'd clear the top slide , tool post and whatever to leave a bare boring table--- and put the dial gauge on it.
I'd stick the test bar in the spindle - no chucks , face plates or whatever. THE Spindle and adjust the height of the dial gauge as near as possible to centre height( hence the wider bearing area). Zero the reading and move the the saddle/boring table towards the tail stock end. If it shows a SLIGHT reading towards the operator or is running dead true- all is well. If it points away from the operator it has to carefully brought to zero- from spindle end to the chuck end.
Yes one CAN do a series of checks on lathe centres and whether the spindle is pointing into the heavens but perhaps later- eh>
More or less- things at the spindle are OK.

Now to tackle the tailstock! I assume that like Jon's lathe the bores of spindle are similar and repeating the work as for the spindle will indicate where things are pointing.
Corrections as before.

Now for the final( ?) test to know that the thing is OK___FINALLY.

Measure the tailstock barrel and with the chuck on the lathe -- turn a piece of round to the exact size of the yailstock bareel diameter. Leavevit ther.

Dismount all the Top slide and whatever to leave the boring table clear and put the dial gauge back and clock the newly made cylinder reading to zero and advance the the saddle to the tailstock barrel . Hopefully the barrel test will read Zero but zero but if not the tailstock has to be adjusted until it is.

NO, for the critics- it isn't Schlessingers Limits but the operator should be happier.

HITH

Norman
 
Thank you Norman, I'm afraid you'll have to elaborate on the elephants foot. But I can understand the rest. Do I need to test the accuracy of the test bar first or take it as given, it is ground to 0.0001" concentricity. Or thats what the sales description said, I don't actually know how id confirm this though.

Time to set the old girl up. I managed to get it to hold a 1thou taper over 4" before it was stripped down however when parts were reversed in the chuck. They didn't seem to be concentric. Which lead me to believe the headstock possibly wasn't completely inline. I know there are other reasons for this. Mainly chuck. But different chucks and faceplate were tried with same indication. So this is where I am, having stripped cleaned checked bearings and spindle, and their adjustment with shims, I have the spindle readings down below 0.0005" on endfloat, and vertical and lateral movement are not detectable so must be within 0.0005" or better, and the spindle still rotates. I've a new saddle and cross slide which eliminates the 10thou error over 6" on the cross slide top. From someones previojs attempts to machine it for some reason. Also new mild steel gib strips.

Time to get serious with the old gal.
 
Jon

The 'elephant's foot' is simply a wider area of contact. In our case, perhaps 3/4" diameter flat dick.

I don't think that a thous in in 4" is too bad. It COULD be a from a list of very minor errors like the 'soft' center which hasn't beet freshly cut EACH time it is re-inserted into the spindle taper.

Of course, we all do that----don't we??;)

As for you and me testing a test bar, but Do you have any test gear that will test better then plus or minus half ten thousand of an inch? I can with a ex Boxford grinding shop dial gauge get down to plus or minus a tenth either way but that is not a guarantee of it being to NPL limits- at the right temperature.
Measure as near as you can but don't expect the impossible
 
Hi Norm, sorry for the delay in replying today. Been out with the kids.
What you say is very true, to measure error, you have to have something that can at least measure to half the increment your wanting to measure to. So for a thou accuracy, you need something that'll read tenths. I don't own anything that can measure to half a tenth, however i do have access to calibrated mitotoyo DTI which will measure to half a tenth.

I've just seen your email. So will take your advice and accept it as delivered is within tolerance. I'll be borrowing the DTI from work to ensure I have the best possible alignment. Even if it takes me a day or two of messing around chasing numbers.
 
Last edited:
What you say is very true, to measure error, you have to have something that can at least measure to half the increment your wanting to measure to. So for a thou accuracy, you need something that'll read tenths. I don't own anything that can measure to half a tenth, however i do have access to calibrated mitotoyo DTI which will measure to half a tenth.
It's the other way round I believe - your accuracy is +/- 1/2 the smallest division of your equipment. Let's say you're measuring centimetres with a metre rule with only whole centimetres marked. You measure something that looks to be somewhere between 66 and 67. You can definitely see it's bigger than 66 and shorter than 67 and you make an estimate of 66.7. Now you're accurate to +/- 0.5 cm (you can see it's less than the 67 line so the high figure is fine, and you'd have to be a bad estimator to think your value was more than 66.5 when it was less than 66.2 so your tolerance should be safe).

Measuring in 10ths of a thou is no difference. Your DTI reads individual 10ths so you can see if it is (say) bigger than 22 but less than 23. How close it gets to 23 is where you make your estimate of the final figure and your stated error now becomes +/- 1/2 a 10th.

In both these cases, you are certain about every figure you state, except for the final one - so for centimetres you know to the centimeter that you're correct and to the 10th of a thou you know you're correct. If you get a DTI that can read to 1/2 a 10th then your accuracy will be +/- 1/4 of a 10th. Out of interest, I did the conversion and 1/4 of 10th comes out at 0.00635mm. My 'everyday use' metric DTI, as well as my micrometer both read in 0.01mm increments, so their accuracy is +/- 0.005mm, which is why I prefer the inherent accuracy/precision gained from using metric measuring tools (what a statement to make, let the flame war begin :p).
 
Of course, this is statement is flawed. It USED to be like that but even then it was possible 'to split hairs' and halve the reading. The late GH Thomas in his usual pedantic but informative way asserted that no hair was to fine to fine to split and he went through his lathes using the Pierre Vernier's Methods to get finer results- with what is essentially a ruler. Vernier's ones were and are straight/linear but GHT's were circular.
Had I not been so blind( macular degeneration). I would have saved a considerable sum having just bought a DRO for my own Myford. It's helluva difference between shelling out £439.99 and £30 for GHT's book:).
Splitting hairs for Jon is possible but he hasn't anything but a borrowed 0.001" dial gauge wheras I can go down to a notional 0.0001 and if I tilt that like one does with a top slide- I think that it is 22 degrees, it improves a cut usually half a thous on the dials to remove a thous in a cylinder to the one thous.
Well, this is the way that I can remove steel so fine that it can be removed by rubbing it between the fingers into dust.

Jon, I know is nowhere near this but I continue to try to offer such advice as he improves.
Jon seems to have his name down for my Myford when I have had those final 9 minutes at the local crematorium.

My views, but of course. Now to Gordon and to try to recall what he needs to solve the problem of indexing(?) a theoretical tool holder on a mystical Tinker tool and cutter grinder.

Norman
 
Thank you Norman, I'm afraid you'll have to elaborate on the elephants foot. But I can understand the rest. Do I need to test the accuracy of the test bar first or take it as given, it is ground to 0.0001" concentricity. Or thats what the sales description said, I don't actually know how id confirm this though.

Time to set the old girl up. I managed to get it to hold a 1thou taper over 4" before it was stripped down however when parts were reversed in the chuck. They didn't seem to be concentric. Which lead me to believe the headstock possibly wasn't completely inline. I know there are other reasons for this. Mainly chuck. But different chucks and faceplate were tried with same indication. So this is where I am, having stripped cleaned checked bearings and spindle, and their adjustment with shims, I have the spindle readings down below 0.0005" on endfloat, and vertical and lateral movement are not detectable so must be within 0.0005" or better, and the spindle still rotates. I've a new saddle and cross slide which eliminates the 10thou error over 6" on the cross slide top. From someones previous attempts to machine it for some reason. Also new mild steel Gib strips.

Time to get serious with the old gal.

16-09-2018-001.JPG
16-09-2018-002.JPG

This is an "Elephants Foot" ! It screws into the dial gauge stylus in place of the ball end. Just to give you an idea of size, that is resting on the top of a jaw on a 100 mm three jaw lathe chuck.
 
Some older lathes for leaving just used tapper shims and grout.
If the lathe on floor or on a ship they steel plate and leave to the plate.

Please post photos of the problem it maybe something else.

Dave

Jon has an old early Myford. To help alignment problems he has sought my help and in addition, bought a test bar etc.

Somewhere I have a copy of test sheet for a Boxford which is similar to a South Bend( I think or was old)

The first thing that I would do is to make an elephant's foot for the dial gauge which gives a wider bearing than the ball bearing affair -which comes in the box.
It saves all this annoying fumbling.
I'd clear the top slide , tool post and whatever to leave a bare boring table--- and put the dial gauge on it.
I'd stick the test bar in the spindle - no chucks , face plates or whatever. THE Spindle and adjust the height of the dial gauge as near as possible to centre height( hence the wider bearing area). Zero the reading and move the the saddle/boring table towards the tail stock end. If it shows a SLIGHT reading towards the operator or is running dead true- all is well. If it points away from the operator it has to carefully brought to zero- from spindle end to the chuck end.
Yes one CAN do a series of checks on lathe centres and whether the spindle is pointing into the heavens but perhaps later- eh>
More or less- things at the spindle are OK.

Now to tackle the tailstock! I assume that like Jon's lathe the bores of spindle are similar and repeating the work as for the spindle will indicate where things are pointing.
Corrections as before.

Now for the final( ?) test to know that the thing is OK___FINALLY.

Measure the tailstock barrel and with the chuck on the lathe -- turn a piece of round to the exact size of the yailstock bareel diameter. Leavevit ther.

Dismount all the Top slide and whatever to leave the boring table clear and put the dial gauge back and clock the newly made cylinder reading to zero and advance the the saddle to the tailstock barrel . Hopefully the barrel test will read Zero but zero but if not the tailstock has to be adjusted until it is.

NO, for the critics- it isn't Schlessingers Limits but the operator should be happier.

HITH

Norman
 
Some older lathes for leaving just used tapper shims and grout.
If the lathe on floor or on a ship they steel plate and leave to the plate.

Please post photos of the problem it maybe something else.

Dave

Hi Dave, I don't want to complicate things at the minute, but the lathe was taken apart to give it a paint, and sort some issues such as bearings, new saddle and cross slide, half nuts, along with others, the lathe held a tolerance of 1thou over 4" length unsupported bar. I could never better it regardless of adjustment to level the lathe. That was before I moved house and stripped it to clean and paint, and search for new parts. The headstock was removed which I probably shouldn't have messed with but parts wouldn't be concentric if reversed in the chuck. I know this is partially due to the position been changed and where the chuck holds ect. But I wanted to try and better it. It's not even fully assembled yet. But will be soon enough (I hope).
Jon
 
Jon
Perhaps it would clear a lot of air if You said. 'Myford ML4 and look at 'lathes.co.uk'

Dave- for me - look at the same site but Myford Super7B PXF. 'B' is Gear Box whilst 'PXF' means power cross feed. If you look at RDG Tools site you might get a further shock at the price-- after adding the goodies.

These are investments and not for the faint hearted.
 
Norman probably right there...(again lol).
The serial number of my Myford ML4 dates from 1936. It's had a hard life and wasn't treated so nicely in a former life. With me it gets gentle use and is great for cutting my teeth on pretty much any machining aspect. It's made parts which sold for the price of the lathe, I bought a cheap lot of tools at the end of last year before the house move was decided. There was an ML10 fixed steady and ML7 travelling steady which I sold and got nearly the price I paid for the job lot of tooling. So now I have multiple restoration projects, starting with the lathe. I'm at the opposite end to Norm, less experienced (read not much of) and with a large family my hobby has to be made cheap as possible. Measuring kit was a big investment for me. I was donated a milling attachment from another good friend, who had it sat doing nothing for the best part of 10years. So effectively I am tooled up, just not with as high spec kit as others. So I ask for advice on how to work around errors, of which there's a number of ways. But chiefly I'd like to get this old lathe as precise as possible with as little outlay as I can manage. Norman, Frazer, and John Moore have been very patient with me over the last few years trying to help me and for their help I'm very greatful. There are others that have offered advice and this has been helpful, again appreciate those individuals input. I came into this hobby with zero knowledge of machining, with an old lathe, after a few years I'm now confident I'll be able to get it all set up accurately in its new home. (Hopefully)

I've also a habbit of rambling too much....sorry lol

Here the old gal in the old 6x4' wooden shed

Jon
FB_IMG_1596295248059.jpg
 
Jon, So, you are saying, you turn the surface on a piece of stock, then you turn it around and remount it into the chuck jaws and it is not running true?
Have you checked the jaw surfaces for gualling or deposits of metal sticking to the face of them? Someone may have spun a piece in the jaws and left a deposit of metal . Very common to see. Can you tell if the jaws are sprung?
I see you have a four jaw chuck mounted to the lathe,
What sort of finish are you getting on the part, before you turn the part in the chuck? Is it smooth, is it round?
Are you supporting the piece with a center when you are turning it?
There are lots of factors that can affect your results.
Have you checked the chuck mounting, no specks of metal on the threads, any little specks of imbeded particals of dirt, metal plastic, gibs adjusted properly?
 
Hello,
Thanks for the input, there is a lot of issues that could explain it, but i have been over most of them. Gibs were tight, not loose allowing vibration deflection, or movement. I've fitted a new saddle and cross slide, which fit much better than the previous one, i had to make a new gib strip for it, but a strip of 3/32" thick mild steel, machined, filed, and polished on plate glass, and the reverse drilled for the screws, did that job perfect, and there is minimal gap now between the saddle and dovetail of the saddle at the rear. The bearings needed shims making for the headstock clamp bolts, A job i had never fully got around to. I had previously used feeler gauges so the bolts could clamp down into something and the adjustment could be held, but time was to make some correct thickness shims. Headstock itself before was good, now its better. While the spindle was out i also identified the grub screw thread in the bull gear, and made a brass grub screw to fit i have now gained the missing speeds on my lathe. previously it was just three. now i have all six spindle speeds available. with a different pulley system i can double this, but it does what i need for now.
I have a 3 jaw, and 4 jaw self centering, and 4 jaw independant, the larger of the chucks i have two sets of jaws and had the same issue with all of them in any configuration. Tailstock support would account for a little of the issue, but the part was still only a couple of inches long. finish on the part before it was reversed was as close to polished as you can get with HSS tooling, i had used a slow feed, sharp tool, with a radius on the tip. Two seconds and ill post a flywheel i did which i think shows the finish perfectly.

Jon
FB_IMG_1597149310044.jpg
FB_IMG_1597149314163.jpg
 
Indeed. The four jaw was used, but id still hope to get somewhere close using just the 3 jaw. All chucks were tried, like i said with the reversed jaws where i had them as the theory was if they were worn then the other jaws would not have as much use. But the small shaft i did wasn't right at all. when laid on a flat surface and rolled, you could feel it rise in the middle, not a lot but enough. i know that turning it between centres would have been better, but with a thread each end it was intended to rotate and die cut the thread. Again thread cutting would have been better but on a small BSF thread was easier to try the short cut first. After i moved house i thought the lathe could do with fully sorting to get it as good as it could be before continuing with any builds, I have several projects, and i just need to pick which one first and stick to it.

Jon
 
All with the aid of an oil stone and WD40😉. It was for a Manstrad, (Manstrand?) Roller who's zamak wheel had crumbled away. It only had a tiny bit of wobble when fitted up. Should have used a mandrel.
 
Ok so two hours in the garage, and I have the following readings. Spindle is still able to be turned by grabbing at the nose thread and turning. I think these are workable with a lathe that was made in 1936.
0.05mm ( 0.001968504") at the spindle for both lateral and horizontal play, this was with a taper drill inserted into the spindle and pulled hard in both directions. and 0.02mm (0.0007874016") run out on nose, again pushing and pulling at the spindle with the drill acting as something to hold onto.
Alignment with the test bar is 0.01mm (0.0003937008") over 100mm (or thereabouts).

Not bad for a lathe that was made in 1936. I was using a mag base and verdict DTI on the saddle. At first I was getting wild readings whatever I did, but remembered ridgidity! I adjusted the DTI so the screw adjuster wasn't part of the setup and then my readings began to get a lot more consistent when it wasn't moving around.

Will be confirming this is right with test cuts, and adjusting the bed if required. But as it's a cantilever bed there isn't much adjustment that's possible even with it on jacking points. I could tighten the spindle bearings more to help with the horizontal and lateral play, but that would then be putting extra load on the motor which I don't want to do. Which while I'm on the subject, that got taken into work and the bearings checked with a noise listening device (I forget it's name) the tiniest bit of dust in the bearings would be detectable. Thankfully was given a full bill of health. So the layshaft got treated to new plumb block bearings too 😀
Regards
Jon
 
Last edited:

Latest posts

Back
Top