Perplexed

[BaronJ]

Hi Guys,

I've gotten an issue that I can't seem to get my head around.:no:
I started to think about making an ER20 collet holder on the Myford lathe, but decided that it would be a good idea to first make sure that the lathe was turning parallel. So I grabbed a 6" inch length of 25 mm bar and relived the middle leaving a 1/2" wide collar at each end then took a skim across both. I also mounted a .001 dial gauge on a magnetic base on the end of the crosslide so that it would be directly above the test piece. Exactly as described in the Myford handbook.

As I expected the free end was about 2 thou smaller than the chuck end. By adjusting the rear, back jacking nut, I got the lathe to turn both collars to the same size. No measurable difference.

Now this is where my quandary starts. Moving the crosslide so that the dial gauge stylus rode onto the top of the test piece collars, I get a consistent 2 thou smaller at the free end. No matter how I adjust the jacking screws, it never changes.

I don't understand why. :no::no::no:

[Michael G]

Adjusting the lathe so that two collars are turned the same assumes that the head stock axis is in a plane horizontal and parallel to the bed. It is possible that you have adjusted the lathe to take out the combined vertical and horizontal difference but as the tool only cuts in the horizontal the vertical component is still there (but compensated for).

I think.
Michael

[janvanruth]

if the jackingscrews are working in the horizontal they will never change the vertical

[Stustoys]

Did you take another cut after you made the adjustment, because what you've done doesn't make sense to me.You have to measure the collars. the dail gauge is a different test, or maybe I'm misunderstanding you. Do you have a link to the test you're doing?

re the vertical test, vertical misalignment gives much smaller error and as MG says you've corrected for it anyway.

Stuart

[DSEL74]

I assume you have the cast iron blocks with the jacking bolts under the lathe. One question do the jack bolt heads screw down fully into the blocks flush or lower than the surface of the blocks?
If so how do you raise the jack bolt small amounts with the lathe sitting on it?

The reason I ask is I don't have them and I couldn't get the lathe running true by shimming it under the feet. So I am in the process of making the riser block with jack bolts and have only seen photos.



Have you checked the headstock is properly aligned and not running out?
Are any of the beds worn? Mine is badly. Which means things aren't as accurate at the chuck ends as at the tailstock.

[Stustoys]

Hi Baron,

Just done the maths on a .002" error on 1" dia. (assuming I'm right and everything else in the setup was perfect lol)call it 0.05mm on 25.4mm, it works out to .000196mm on dia. Thats ummm not alot :wink:

Hi Dale,

If your lathe is really as bad as you say, I'm not sure you should bother to try and get it to "cut straight". Set it up as level as you can, call it good, then turn between centers if you need to.

Stuart

[simonl]

I reckon Stuart is on the money. I think you may be measuring a difference in tool height. This could be vertical mis-alignment of the headstock or it could be due to bed wear or a combination.

Either way, the 2 thou difference in tool height has a very small (read un-measureable to us) affect on diameter. Your colars are the "same" diameter, cut your losses, stop measuring stuff and get on with it! :U

Simon

[cba_melbourne]

Hi Guys,

I've gotten an issue that I can't seem to get my head around.:no:
I started to think about making an ER20 collet holder on the Myford lathe, but decided that it would be a good idea to first make sure that the lathe was turning parallel. So I grabbed a 6" inch length of 25 mm bar and relived the middle leaving a 1/2" wide collar at each end then took a skim across both. I also mounted a .001 dial gauge on a magnetic base on the end of the crosslide so that it would be directly above the test piece. Exactly as described in the Myford handbook.

As I expected the free end was about 2 thou smaller than the chuck end. By adjusting the rear, back jacking nut, I got the lathe to turn both collars to the same size. No measurable difference.

Now this is where my quandary starts. Moving the crosslide so that the dial gauge stylus rode onto the top of the test piece collars, I get a consistent 2 thou smaller at the free end. No matter how I adjust the jacking screws, it never changes.

I don't understand why. :no::no::no:


Are you by any chance using a lathe with sleeve type spindle bearings? The center line of the spindle rises very slightly when the spindle rotates, because it floats up on the oil film. Its a very small amount, but it could account for 0.01mm. When the spindle stands still, the oil film gets squeezed out and the spindle makes metal to metal contact to the bearing sleeve.

Now, that would explain why the whole spindle sits lower at standstill. But you observe that the spindle points downwards only on the right side. Are you maybe using a lathe that has a sleeve bearing at the right end of the spindle, and a roller bearing at the left end? Like a Myford? That way only the right end of the spindle would drop at standstill. And the 0.01mm drop in the sleeve bearing would be amplified at the end of your test spool to be about 0.02mm.... or about the imperial "thou" that you observe? I am just speculating, hence the question marks. How could this theory be proved? well, what happens to your measurement, if you gently lift the right end of your test spool, is it maybe taking up the slack at the top of the sleeve bearing?

Edit: this sort of up-down spindle movement is normally not detected when doing the spool test, because one measures and compares the diameter of the two spools to untwist the lathe bed. It is rather unusual for someone to measure the top of the spools like you did.

[BaronJ]

Hi Guys,
I've kept out of the workshop today, I don't want to do any more until I understand what is going on.

Dale:
Yes I have the cast iron riser blocks and also the factory metal cabinet stand. The riser blocks have a a pair of threaded holes through them matching the feet on the Myford. There is a long threaded rod going all the way through both holes in the block and through the drip tray and cabinet stand. The riser blocks are secured by nuts and washers inside the stand. There are nuts and washers on each side of all the lathe feet. The nuts on the underside of the lathe are initially positioned so that a spanner can be used to to turn the nut. By turning the nuts you can adjust the twist on the lathe bed. The initial setup procedure is to leave the tailstock nuts loose and level the lathe at the headstock, then nip up the top nuts. After this all adjustment is done by using the tailstock nuts.

Stuart:
Thank you for doing the maths for me. Sums and I don't get on very well. :no: But it's nice to know that the apparent error is very small.

Simon:
Thanks for the encouragement. If I hadn't decided to try a precision job this quandary wouldn't have come up.

CBA:


Are you by any chance using a lathe with sleeve type spindle bearings? The center line of the spindle rises very slightly when the spindle rotates, because it floats up on the oil film. Its a very small amount, but it could account for 0.01mm. When the spindle stands still, the oil film gets squeezed out and the spindle makes metal to metal contact to the bearing sleeve.

Now, that would explain why the whole spindle sits lower at standstill. But you observe that the spindle points downwards only on the right side. Are you maybe using a lathe that has a sleeve bearing at the right end of the spindle, and a roller bearing at the left end? Like a Myford? That way only the right end of the spindle would drop at standstill. And the 0.01mm drop in the sleeve bearing would be amplified at the end of your test spool to be about 0.02mm.... or about the imperial "thou" that you observe? I am just speculating, hence the question marks. How could this theory be proved? well, what happens to your measurement, if you gently lift the right end of your test spool, is it maybe taking up the slack at the top of the sleeve bearing?


Yes my lathe is a Myford, exactly as you describe. Whilst It never occurred to me to try and lift the end of the workpiece, I will try that tomorrow and report back.



Edit: this sort of up-down spindle movement is normally not detected when doing the spool test, because one measures and compares the diameter of the two spools to untwist the lathe bed. It is rather unusual for someone to measure the top of the spools like you did.


Actually, bearing in mind your comments above, trying to belt and braces the measurements of the test piece. My thinking was whilst I can measure each end with the micrometer, if I set a dial gauge up to measure each end from directly above the two readings would concur of the difference.

I'll report back later.

[cba_melbourne]

Hi Baronj,

We may be on tye right track here. Not all Myfords are made the same. But the most common 3.5" center height model have indeed a hand-scraped tapered sleeve bearing at the right and two angular contact ball bearings set against each other on the left. Here some information I just found by googling for the term "spindle centerline rising on oil film":
http://edge.rit.edu/edge/P14453/public/Research/2-_LEADER_-_Understanding_Journal_Bearings.pdf

On page 3 and 4 you can see graphically what I meant by "floating up on an oil film". The motion is actually both forwards and upwards. Confirming that you cannot use a dial indicator, not vertically and not horizontally, to measure the diameters of a test spool. You must indeed use a micrometer screw. You cannot look at a lathe spindle as a static object alone. It has a dynamic life.

PS: for those old fashioned members critical of using the internet to gather information, I may explain how I used it on this example. One first has to use one's own brain to analyze and assess a problem. Then maybe remember from school that shafts in a sleeve bearings float up an oil wedge. Before one can formulate a search string for a search engine suitable to come up with any useful extra information. I still do not claim that this is the right answer or the only answer for BaronJ. But it does explain what he observes. Of course, the amount by which the shaft rises when in motion depends on both the wear state of the sleeve bearing and how exactly it has been adjusted. I would guess that a new and well adjusted Myford spindle only rises by some 0.002mm, but when worn or adjusted with excessive slop that could be much more. I do not own such a lathe, so cannot tell.

[BaronJ]

Hi Guys,

First a few pictures then we all know exactly what I've done and how the results were obtained.
336547 336548

These twp pictures show the dial gauge readings at the chuck and at the end of the test piece. As can be seen there is roughly half a thou between the two. Following on from reading the previous posts, lifting the free end of the test piece results in the half thou disappearing. I also noted that after taking another 1 thou cut, that the half thou increased very slightly. Probably due to the bearings warming up.

336549 336550

These are the micrometer readings after taking the 1 thou cut mentioned above. It can be seen that there is a very slight, maybe the thickness of a division line, difference, which I didn't have the day before. At this point I'm now happy that I can stop chasing rainbows and get on with doing what I started to before I got distracted.

I've learnt a great deal from you guys in the process and must thank CBA for the push in the right direction. I had absolutely no idea about this effect and no inkling that trying to make a measurement the way I was, would cause such head scratching.

Thanks Guys.

[BaronJ]

Hi Dale,
This post is just for you. :):):) I took a couple of photographs of my Myford Lathe riser blocks so you could see exactly how they are set up.

336552 336553

The first picture is the rear/tail stock end, and the other is the headstock end. As you can see there is adequate room to get a spanner in underneath the lathe foot to adjust the nut. When I was setting mine, as per the earlier posts, About 3/4 of a turn was needed.

[DSEL74]

Baron thanks for that mate! I have my block of cast iron ready just need to do the bolts & holes. Thanks to forum members the block was milled and surface ground flat & parallel then cut in half to make a matching pair.


Baron did you ever test your lathe with a ground test bar? It would be interesting to do a comparison of the dial indicator from a ground bar vs a fresh cut bar.



Stu, As per what I showed you with how the saddle moves freely at the chuck end and is really tight at the tailstock end due to wear. What this means is it it going to be difficult to turn long objects the full length of the lathe bed. What I am hoping is if I am turning small short objects I can adjust the gib strip for the chuck end to get the any slack taken out and get some accuracy there. Of course this mean it will be too tight to wind further down the bed.

[BaronJ]

Baron thanks for that mate! I have my block of cast iron ready just need to do the bolts & holes. Thanks to forum members the block was milled and surface ground flat & parallel then cut in half to make a matching pair.


Baron did you ever test your lathe with a ground test bar? It would be interesting to do a comparison of the dial indicator from a ground bar vs a fresh cut bar.



Stu, As per what I showed you with how the saddle moves freely at the chuck end and is really tight at the tailstock end due to wear. What this means is it it going to be difficult to turn long objects the full length of the lathe bed. What I am hoping is if I am turning small short objects I can adjust the gib strip for the chuck end to get the any slack taken out and get some accuracy there. Of course this means it will be too tight to wind further down the bed.

Dale maybe this rough drawing may help. I don't know off the top of my head what the actual thread sizes are. If I can find out I'll post. Re test bar. No I've never had one.

[DSEL74]

That is how I was planning to make them. From photos I had seen of the originals it looked like the lower nut could go down into the blocks in a counter bore and i couldn't see how that would work.
336567336568



For a ground test bar, just go to your local truck mechanic and ask for something suitable. I think the piston pins (don't know their technical name) are one option. But any ground shaft out of the motor etc will be suitable. My local guy lets me dumpster dive for all kinds of bits, most of which I use blacksmithing. I picked up a 1200mm apx Dia., ring bearing type thing the other day, think giant lazy susan might make a rotating bench out of it.

[BaronJ]

That is how I was planning to make them. From photos I had seen of the originals it looked like the lower nut could go down into the blocks in a counter bore and i couldn't see how that would work.
336567336568


Yes there is a small counterbore, but its only a couple of millimeters, nowhere near enough to hide the hex part of the nut. The hex thickness on mine is about 8 mm.



For a ground test bar, just go to your local truck mechanic and ask for something suitable. I think the piston pins (don't know their technical name) are one option. But any ground shaft out of the motor etc will be suitable. My local guy lets me dumpster dive for all kinds of bits, most of which I use blacksmithing. I picked up a 1200mm apx Dia., ring bearing type thing the other day, think giant lazy susan might make a rotating bench out of it.

Ah I see what you mean. I've loads of ground silver steel bar from 1 mm to 30 mm diameter. I can use one of those as a test bar.