Myford ML4 Restoration: Headstock bearings and spindle removal

[Luke MitchellParticipant @lukemitchell30627]

Hi again all,

Last night I reassembled my headstock, complete with new thrust bearing (sadly of much worse quality than the SKF one which did not fit – but it was the best I could find) and some new power-twist belts for ease of adjustment.

When I first turned on the lathe there was a bit of a knocking sound coming from the headstock, once per revolution. I double-checked the bearing tightness, which was snug but not over-tight, and decided to run the motor for a little bit to see if it resulted in any heat. After 20 minutes the sound lessened a bit and the bearings were only very slightly warm to the touch.

I then took some light cuts on aluminium, brass and then steel. Trying to part the aluminium/steel was unpleasant, although less chattery than previously. I was able to part the brass very cleanly.

There is a bit of slop in the handwheels and the backgear assembly that I am going to take out using fiber washers. I think this will help dampen the vibrations and make using the lathe a little more pleasant. I also need to thoroughly examine the headstock to determine where the noise is coming from – it seems as though it's coming from the chuck-side bearing but there is a chance it's the new belt.

I've attached a video and I wonder if someone is able to give their opinion on the noise?

Kind regards

[Brian WoodParticipant @brianwood45127]

Luke,

There is an old trick you can use with a long handled screwdriver to use as a sounding stick in trying to isolate an area for noise diagnosis.

Your short video clip sounds like belt noise. It is regular and maybe coincides with the revolution rate. I would try tightening up the pulley block grub screw, there could be some slop in the fit between grub screw and dimple in the mandrel, I seem to recall Dad's lathe often needed that taking up..

Is it any less when you ease the tension on the belt? The lathe certainly sounds smooth enough when you ignore the noise.

Kind regards Brian

[Jon CameronParticipant @joncameron26580]

Hi Luke,

I believe that it may be the bull gear on the back of the headstock bearing, try moving the pulley over slightly, then lock the bull gear to the spindle DO NOT ENGAGE THE BACK GEAR CLUSTER, lock the nut on the top of the headstock so there is no danger of it falling into mesh with the bull gear. Run the lathe again and see if the noise persists. If it does and that isn't the issue try the following below. Which I wrote before I realised what noise I was listening to. (Mine makes the same noise with the change gears disengaged). The bull wheel is normally free to rotate on the spindle as it will have varying speeds to the spindle any slack is magnified.

If that doesnt work try slowing the spindle speed down and see if you still hear a noise, if the spindle turns over freely it could be a variety of issues that will need working through systematically.

1. remove the belt to the lathe from the layshaft, run the motor see if its actually somewhere else first. it may be that one of the pulleys is loose on the shaft or you have wear in the motor bearing and its transferring vibration to the lathe via the belts. So if there is still the noise with only the layshaft belt connected up disconnect the one to the motor and see if you can hear anything there, it may be as simple as the belts are too tight, but check it all before fettling with the lathe headstock again. Check the layshaft by hand and check all pulleys are correct on the shaft and not slightly to one side of the flat on the shaft which could also cause issues.

2. remove the chuck and turn the spindle by hand, slowly, if it is catching then you'll feel it. Are you sure the headstock was completely clear of any swarf or debris that could be catching in the new bearing, There should be a washer plate front and rear of the thrust bearing. It should not run direct onto the headstock casting.

3. Put a DTI on the register of the spindle, push it forward and back to see if the needle moves, you should have zero end float in an idea world, though if the lathe is slightly warn you wont be able to adjust out all the end float with the threaded washer at the change gear end. Check on the nose thread that there is no up or down movement in the bearings. then repeat the DTI test by turning the spindle by hand.

4. if all these check out and don't present any issues or you don't feel any tight spots then check the pulley and bull gear on the spindle are clear of the headstock and casting, and not knocking each other, also check that the grub screws are loose enough to permit the spindle to turn without the bull gear grub screw catching the spindle. (after re-watching the video I think this is the most likely)

5 check the tool your using is sharp and on centre, also that its 90 degrees to the axis of the lathe, I use the chuck body to align it on smaller work.

6. You may have issues with parting or surface finish if the gib screws are loose. Check these are tight enough again with a DTI so you know that the saddle and cross slide don't move in any axis, any movement here will be amplified at the tool tip.

From the video i'm afraid its difficult to hear anything, but going slower to determine the issue will be far more beneficial than thrashing the lathe at full speed when you know there's something not right.

Jon

[Robert Atkinson 2Participant @robertatkinson2]

Posted by Jon Cameron on 12/05/2020 10:33:15:

Hi Brian, Luke

Ive found the article by Martin Cleeves. It is indeed for an ML4, my spindle has the earlier spindle and nose thread of 7/8" 9TPI, therefore it only accommodates an MT1 taper and passes 3/8" stock. Below youll see a scan of the article, which also shows a grinder attachment for the rear of the lathe spindle. Something handy if you only have the one motor, though not recommended as the dust and grit would not do your precious lathe much good. Though the rest would be handy for setting up in front of a bench grinder, to assist with grinding HSS, or Carbide tooling. Something that Luke might like to look into for his tool grinding, it is certainly something ill be doing.

Jon,

what a tease! You can't read the dimensions! Any chance of putting a higher resolution scan of the spindle drawing in your album or giving us the magzine issue information?

Robert G8RPI.

[Brian WoodParticipant @brianwood45127]

Jon,

I have to take issue with your statement about the bullgear being free to rotate on the spindle.

It is the pulley cluster that is normally locked to the spindle in direct drive. It is unlocked to be free to rotate when back gear is engaged. The bullwheel is normally locked by it's grub screw to the spindle at all times. It would otherwise tend to rattle and wear both the bore and possibly even the spindle diameter at it's location.

Sorry

Kind regards Brian.

[Luke MitchellParticipant @lukemitchell30627]

Hi again all, and thanks for the responses.

Posted by Brian Wood on 15/05/2020 16:54:16:

It is the pulley cluster that is normally locked to the spindle in direct drive. It is unlocked to be free to rotate when back gear is engaged. The bullwheel is normally locked by it's grub screw to the spindle at all times. It would otherwise tend to rattle and wear both the bore and possibly even the spindle diameter at it's location.

I'd tightened the bullwheel for exactly this reason. I will double-check later (just finishing at work for the week) but I'm quite sure I have both the bullwheel and the pulley assembly snugged down securely – I was careful to seat the grub screws in the divots and not to further burr the spindle (as obviously these had been loose on another occasion).

Jon,

I'll do an end-float test and let you know what happens. I'm also going to indicate the amount of axial play present. I don't think there is much but I may be able to tighten up the thrust collar a little more on the change gear side. The trust bearing does indeed have a washer on either side of the ball bearings – thanks for your concern.

Currently I've got 0.015mm of radial run-out as I turn the spindle. It seems to occur in the same place at each rotation which I find a little alarming. What are the chances that the spindle is worn oval?

I'm also going to try the other tests to see if the problem is originating elsewhere. The motor has a lot of axial play in its shaft so that's a possible culprit. I've got another one that I'm in the process of wiring up for use with a VFD that will soon be replacing it.

Kind regards,

Luke

[Brian WoodParticipant @brianwood45127]

Luke,

I think it unlikely the spindle is worn oval, there could be other causes that can mimic that. The way to test for it is to support the spindle in two vee blocks and rotate is slowly with a DTI bearing on the surface. There should be no fluctuation. Measuring across a diameter with a micrometer or caliper can be fooled by the effects of lobing which gives an apparently true diameter. The vee block test will find that too. I agree that 0.015 mm run out is rather more than tolerable

However, for you to test for those conditions means pulling the spindle!

Is the radial run out measured on the plain diameter of the nose or within the MT bore? Is the vertical chuck register absolutely true?

If the MT bore is running true along with the vertical register then you could skim a few thou off the horizontal register and apply David Bennett's findings that I supported very recently and locate the chuck on the vertical register alone.

Kind regards Brian

Edited By Brian Wood on 15/05/2020 18:06:32

[Luke MitchellParticipant @lukemitchell30627]

Posted by Brian Wood on 15/05/2020 18:03:59:

Luke,

I think it unlikely the spindle is worn oval, there could be other causes that can mimic that. The way to test for it is to support the spindle in two vee blocks and rotate is slowly with a DTI bearing on the surface. There should be no fluctuation. Measuring across a diameter with a micrometer or caliper can be fooled by the effects of lobing which gives an apparently true diameter. The vee block test will find that too. I agree that 0.015 mm run out is rather more than tolerable

However, for you to test for those conditions means pulling the spindle!

Is the radial run out measured on the plain diameter of the nose or within the MT bore? Is the vertical chuck register absolutely true?

If the MT bore is running true along with the vertical register then you could skim a few thou off the horizontal register and apply David Bennett's findings that I supported very recently and locate the chuck on the vertical register alone.

Kind regards Brian

Edited By Brian Wood on 15/05/2020 18:06:32

Hi Brian, thanks again for your response.

I've just had a look at the lathe. Running it with slack belts results in a lot of noise from the motor and I think that may be the cause of the sound. I've attached a video to illustrate. It will be interesting to see how quietly it runs with the new motor/VFD.

I've not indicated on the vertical register as, honestly, I didn't think to. It seems obvious now though. I'll do that soon and report back on how true it is. How would one reinstall the spindle to ensure it is straight? the fit was so snug that I can't imagine how it would have been pulled off true.

I also took the time to tighten up the nut securing the backgear assembly. This has a bit of play in it and would rattle under vibration. There is a bolt in the casting that would prevent accidental engagement but, to be on the safe side, and to keep things quiet, I snugged the nut to lock it in place. Thanks Jon, for that.

I also re-tightened the pulley assembly and the bull wheel via their grub screws. As you said, they had already come a bit loose. I'll have to watch that I adjust them regularly.

Another benefit of having locked the back gear via the nut is that it will force me to think before I engage it. Hopefully this will prevent any unfortunate accidents such as forgetting to remove the grub screw from the pulley assembly!

Here is a photo of the clean, adjusted headstock for your enjoyment.

As an aside, I am enjoying the new belts. They were a bit fiddly to install but it's nice to know I can remove them without having to touch the spindle.

I may end up removing the spindle to perform the test that you mention Brian (although it will require finding/borrowing/buying some vee-blocks), but until then I'm happy to leave the spindle in situ. I'm still hopeful that a rear tool post and some more dampening of vibrations will make using the lathe even more pleasant.

I've ordered a pair of 20T wheels from Myford, as these were missing from my set, and I'll try to set up the fine-feed that you mentioned earlier in this thread too. To clarify, does 20 : 60/20 : 65/20 : 60 mean the 20T on the spindle, 60T with a 20T compound on the 1st stud, 65T with a 20T compound on the 2nd stud, then 60T on the leadscrew? I apologise if it's a basic question but I've not had to use gearing/drive chain notation before.

Kind regards all, and have a nice weekend.

Luke

[Robert Atkinson 2Participant @robertatkinson2]

I've got to agree with Brian on the pulley / Bull gear locking. The Myford instructions clearly say to unscrew the grubscrew in the pulley. Some ML1/2/34 have had a groove cut in the front headstock bearing casting to give access o the Bull gear locking screw. I wonder it this was to counter a seized pulley.

Robert G8RPI.

[Jon CameronParticipant @joncameron26580]

Brian, Robert, my instruction above with the bull wheel was to rule out that it wasnt rattling around on the spindle. I don't see where the issue is in my statement to rule out a possible cause of vibration and thus noise. Luke has said that they had already become a little loose on the spindle

Luke, With regards to the spindle if it's moving 0.015mm, and only in one place, then i would be worrying about a Burr either on the spindle or the bearing. although with the link belts, double check they are all connected correctly, as one loose one would also pass vibration, although wouldn't explain the movement you've measured.

Do the DTI test without any belts attached, as I'm sure your already aware.

[Jon CameronParticipant @joncameron26580]

Sorry forgot to add, yes the way you describe the gear train is correct. Work small to large gears.

You will find the new myford gears are keyed, (I'm not sure if this has already been mentioned) but you will need to drill for the pin to compound the gears. Hole is 3/32", and be a reasonably snug fit for the pin but not so tight that it is impossible to remove.

[david bennett 8Participant @davidbennett8]

Luke, I agree with Brian Wood's comments about where to measure the run-out . If the vertical register is true, then too snug a fit on the horizontal register can interfere with the self-centering action of the spindle threads. ( if carefully done, you might be able to see this with the aid of a little marking blue on the register)

Dave.

[david bennett 8Participant @davidbennett8]

Luke, as a matter of interest, would you measure the register clearance on your old ml4 backplate? ( mine has a 20 thou' clearance.)

Dave.

'

[Luke MitchellParticipant @lukemitchell30627]

Dave, I would, gladly, but I’m afraid I don’t know what the backplate is. If you could describe it for me then I’ll measure it for you.

i have 0.005mm or less run out on the vertical register so I believe that is positioned correctly.

Last night I did a little turning on a small piece brass. As it was so small I had to turn it around during turning and needed to use the 4-jaw to ensure concentricity. I was able to dial in the workpiece to within 0.015 on all three occasions so I think the runout is genuine. Whether or not it exists on the horizontal register or somewhere else remains to be determined.

I’ll apply some engineers blue and see if the chucks are being seated evenly on the register as a next port-of-call. Actually, it may have to be sharpie 😂

Thanks all,

Luke

[Brian WoodParticipant @brianwood45127]

Robert,

I remember the groove now you mention it

Jon,

I quote " The bull wheel is normally free to rotate on the spindle as it will have varying speeds to the spindle any slack is magnified" That is where I got the information

Luke,

Your gearing set up as described will give you a truly fine feed off 0.004 inches per revolution of the workpiece. Did you measure your runout using the spindle MT bore or the plain external register? At least the vertical register is as good as you can expect.

David,

It maybe needs a leap of faith for Luke at this stage to take metal off his plain diameter. All is not lost though if he feels that has been a mistake, a sleeve will restore the status quo.

Kind regards to all Brian

[Luke MitchellParticipant @lukemitchell30627]

Brian,

Excellent, I look forward to establishing that fine feed.

I indicated on the plain register of the spindle itself. I’ve not indicated from the tapered bore as I don’t have anything I deem reliable enough for such a measurement. I’d like to purchase a test bar with an #2 taper on one end for this purpose, and to assist with adjusting the tailstock, but I’ve not done so yet as they seem to be about £35 and I have already spent a lot of money on this project.

I am hesitant to take any material from off of the spindle unless I absolutely have to. At present I think I am flappy enough. Having just converted the run out into inches, 0.015mm is a little under 6 tenths. I believe that 5 tenths was the specification for the original spindle so it’s not far from this anyway.

On another note, the modified hand wheel for the cross slide has 42 divisions. One of my relatives, an earlier owner of the lathe, must have known about the inaccuracies of having 80 divisions on the original. I believe the current wheel now represents approximately 0.002” per division.

[Brian WoodParticipant @brianwood45127]

Hello Luke,

You are quite right, 6 tenths of a thou, I got a decimal place wrong so made that rather more at 6 thou!

Leave well alone—sorry to be alarmist.

A test bar would be a good accessory but it isn't really necessary at this stage.

A first setting for your tailstock is easy enough but it does depend on having a 3 jaw chuck you trust. It works as follows. Wind out the barrel, mark the outside with a black marker pen and offer it up to the chuck with the jaws just touching, Pull over the spindle by hand for a rotation or two and see where the marking has rubbed off. Offset the tailstock accordingly a little and repeat the sequence, chasing the "grip" of the jaws down as you do until the marking is pretty evenly wiped off all round. Lock the lailstock offset at that point.

Now you can refine that setting, if it needs it, with fine adjustments to the tailstock offset on a turning test between centres

The cross slide handwheel of 42 divisions is a close approximation to 0.002" per division. The 80 divisions that Myford offered gave 1.042 thou movement per division.

Kind regards Brian

Edited By Brian Wood on 16/05/2020 11:55:01

[Howard LewisParticipant @howardlewis46836]

Based on my experience, you may need to do a little filing at the mandrel end of the banjo, to be able to mesh all the changegears correctly for the 20:60 / 20:65 / 20:60 fine feed train.

Or it may have been the studs that I had to make for that particular ML4?

The Top and Cross Slide feedscrews are 12 tpi, and the original 80 division dials gave an increment of 1.042 /1000" per division.

A closer approximation would have been a dial with 83 divisions, to give 1.004 /1000 per division.

42 Divisions gives 1.984/ 1000"

HTH

Howard

[david bennett 8Participant @davidbennett8]

Luke, I am getting cofused here. I was under the impression that you had added a collar to your spindle to fit a new chuck. Do not cut anything from your original spindle. The horizontal register should be absolutely true. If it is indeed a collar that you have added, that too should be absolutely true (or within spec.) The chuck backplate is the loose plate that screws to the lathe nose, which should be trued before the chuck is bolted on. The register is simply the plain bore of the backplate. Can you measure the old chuck backplate bore that fitted to the original 1.125" horizontal spindle register.

Dave.

Edited By david bennett 8 on 16/05/2020 13:18:54

[david bennett 8Participant @davidbennett8]

Luke sorry, I put my edit (or within spec) in the wrong place Any slight error in the spindle vertical register will be eliminated when the backplate is trued on the lathe. The backplate should be absolutely true and repeatable,if nothing interferes with the fitting.

Dave.

Edited By david bennett 8 on 16/05/2020 14:29:48

[Luke MitchellParticipant @lukemitchell30627]

Posted by david bennett 8 on 16/05/2020 13:18:22:

Luke, I am getting cofused here. I was under the impression that you had added a collar to your spindle to fit a new chuck. Do not cut anything from your original spindle. The horizontal register should be absolutely true. If it is indeed a collar that you have added, that too should be absolutely true (or within spec.) The chuck backplate is the loose plate that screws to the lathe nose, which should be trued before the chuck is bolted on. The register is simply the plain bore of the backplate. Can you measure the old chuck backplate bore that fitted to the original 1.125" horizontal spindle register.

Dave.

Edited By david bennett 8 on 16/05/2020 13:18:54

Hi Dave,

I've not yet added such a sleeve – all of the measurements are taken from the spindle itself: the plain, unthreaded section (which I referred to above as the "plain register", perhaps confusingly) and the vertical register. To summarise, these are the run-out measurements:

Radial run-out, as measured from the unthreaded horizontal section: 0.015mm / 0.0006"

Axial run-out, as measured from the vertical register: 0.005mm / 0.0002"

On re-reading, it's obvious that you meant the chuck backplate. I'm not sure what confused me the first time around – perhaps I'd drunk my Friday night beer a little too quickly!

I confess that I'd not thought to measure the concentricity of the backplate, or indeed to remove it from the chuck. Given that the chuck is not running very true, nor very repeatable, I suspect that I won't worsen things by removing it – as long as I clearly mark the original position to ensure I re-fit using the same bolt holes.

Kind regards,

Luke

[david bennett 8Participant @davidbennett8]

Luke, thanks for clarifying that. All lathe noses have different errors. The whole purpose of the chuck backplate is to allow you to eliminate them . (Except the axial play – that is adjusted by the threaded collar at the rear end of the headstock ) For a full description of truing the backplate, see Tony's famous website "lathes.co.uk" the article "fit a new chuck"

Regards, Dave.

[david bennett 8Participant @davidbennett8]

Luke, something you said there made me think the term "repeatable" may be being misunderstood. Do not expect that you can remove a piece of work from the chuck and then replace it and have it run true. Turn a piece of work true. Do not remove it from the chuck. Remove/refit the chuck with the work still clamped in the jaws, if it then runs true it is repeatable (apologies if you knew that)

Dave.

[Author]

Posts