PDA

View Full Version : Headstock alignment



koda2
11-17-2009, 09:23 PM
In reference to trying to align the lathe portion of my Smithy,

I spent hours,...no days, trying the RDM with a 1" diameter rod. The best I could do was a .001 -.003 taper in 5 inches. Finally gave up and inserted .0002 shims randomly until a fine cut on the lathe was a uniform dimension on a 1.5" cylinder 5" long.

Part of the problem is the puny and tedious way the headstock is locked down.

My progression of alignment attempts:
-Build a rock-solid table with threaded feet.
-Level the table.
-Rebuild the spindle with better bearings and preload as best I could (result=TIR 3 tenths or less on the spindle taper.)
-Did not scrape the ways; they were new and ground by the factory. There was no appreciable twist and the carriage slides uniformly easily down the ways with the gibs fairly snug.
-Level the lathe bed with Starett precision level fore and aft and side to side.
-Remove the headstock, clean all grit, paste, paint from joining surfaces and reinstall. Then align with RDM and 20 inch aluminum cylinder. No joy.
-I believe the crosslide is perpendicular to the ways but don't know how to check it.
-The tailstock is still a mess. The ram is sloppy in the tailstock bore. I will have to figure out how to bore out and recylinder the tailstock eventually but the headstock alignment (I believe) takes precedence and is more important.
My next attempt when time permits will be to try to rig up a laser pointer down the center of the spindle bore and check alignment on a tiny centered hole at the other end, but I don't think it will give precise enough readings by itself.
Dave A.

The Artful Bodger
11-17-2009, 09:40 PM
I suppose one way to check that the spindle is parallel to the ways would be to make a 'cylindrical square' and test that it really is square. One way to test your cyclidrical square would be to make a second one the same size then stand them side by side on a plane surface where there should be no visible gap, tip one square end for end and repeat.


A cylindrical square is just a cylinder cut on the lathe being careful not to move the cross slide on the final pass. Both ends are parted off on the lathe.

dp
11-17-2009, 09:44 PM
I suppose one way to check that the spindle is parallel to the ways would be to make a 'cylindrical square' and test that it really is square. One way to test your cyclidrical square would be to make a second one the same size then stand them side by side on a plane surface where there should be no visible gap, tip one square end for end and repeat.


A cylindrical square is just a cylinder cut on the lathe being careful not to move the cross slide on the final pass. Both ends are parted off on the lathe.

Large piston wrist pins have been suggested for use as cylinder squares. Worth a try.

Evan
11-17-2009, 09:56 PM
My next attempt when time permits will be to try to rig up a laser pointer down the center of the spindle bore and check alignment on a tiny centered hole at the other end, but I don't think it will give precise enough readings by itself.

That is precisely what I am working on but it includes a way to multiply any error by a factor of perhaps 100 to 1000. This will make it possible to detect an off axis condition of as little as .0001 per foot.

I have already developed a simple method of bed alighment that uses a laser and two 90 degree prisims but have never shown it here because most people can't make use of it. I happen to have about 30 feet of open hallway that I can shine a laser beam down and back which gives a 60 foot angular multiplication of a six inch spacing of the bed ways. A 1 thou twist in one foot becomes a 1/4 inch movement of the laser beam once it has gone down and back.

The Artful Bodger
11-17-2009, 09:57 PM
Large piston wrist pins have been suggested for use as cylinder squares. Worth a try.


....Ummmm, yes they might be good cylindrical squares but that would not test the lathe. Of course making one on the lathe and comparing with the wrist pin would be a test but by making two cyclidrical squares on the lathe the error would be double and presumably easier to detect.

darryl
11-17-2009, 10:02 PM
'the puny and tedious way the headstock is locked down'- what's going on, is the headstock trying to settle into its old location whenever you tighten it down?

Is there a method given to adjust the headstock, or is it done by shimming-

Can you improve on the method of securing it?

It might help to chuck a solid bar and perform 'the method', then use the crosslide to 'adjust' the outboard end of the bar by the needed amount while using the tedious securing method. Put something solid in place of the cutting tool that you can use to butt up against the side of the bar to nudge it with. Depending on which way you have to rotate the headstock, you may have to nudge the bar from the rear side. It would be from the rear if the outboard end is too large, and from the front if the outboard end is too small. Nudge by half the amount of the difference in diameters between the two test collars.

Maybe you need to raise the headstock by the thickness of some spring steel strips to get rid of mating ridges that might be locking the headstock in just the wrong position. You can usually make a larger movement and re-secure it at that spot, but sometimes you just can't make a very minor adjustment.

Just some ideas.

The Artful Bodger
11-17-2009, 10:11 PM
That is precisely what I am working on but it includes a way to multiply any error by a factor of perhaps 100 to 1000. This will make it possible to detect an off axis condition of as little as .0001 per foot.

I have already developed a simple method of bed alighment that uses a laser and two 90 degree prisims but have never shown it here because most people can't make use of it. I happen to have about 30 feet of open hallway that I can shine a laser beam down and back which gives a 60 foot angular multiplication of a six inch spacing of the bed ways. A 1 thou twist in one foot becomes a 1/4 inch movement of the laser beam once it has gone down and back.


Evan, I did a simple test with a laser for 'bed twist'. I first 'levelled' the lathe end for end using a 36" builders level then I mounted a laser pointer on the carriage shining at about 90 degrees from the lathe axis to a wall 20 feet or so away. I marked the point of the laser dot on the wall then ran the carriage to the other end and marked the new position of the dot. A simple test with the builders level showed the two marks to be level, within the accuracy of the level.

If I am not mistaken this rig amplified any error to the ratio of carriage travel (about 36") to the distance to the wall, about 20', lets call that 6:1; A better level would have been better and a longer path for the laser would also have been better.

Ken_Shea
11-17-2009, 10:11 PM
I had success to nudge the head stock by just barely snugging the head down then using a dead blow hammer to accomplish the "nudge"

oldtiffie
11-17-2009, 10:17 PM
No need for a "cylinder square" as a large - say 1" x 4>6" long - wrist pin (new) will do all that is necessary for a pretty quick and accurate test of the head-stock spindle axis to the bed.

If the wrist pin ("gudgeon" pin here and the UK) has a centre in each end that it was ground on - so much the better. That may not be the case as many pins are "centre-less" ground. But there is a good easy cheap work-around for that as well - I will cover that later.

Lining up the tail-stock is not difficult either so far as the "forward/back" alignment is concerned. Testing thhetail-stock for correct centre height can alos be done as partm of the process.

I won't address the morse tapers in the head and tail-stocks until later.

"Flatness" of a facing cut only requires a facing cut and a reasonably good dial indicator.

A bit of precision-ground round stock (mild steel will do just fine) that is say 20mm (say 1") x say 8" to 12" long that is not bowed or "bent" more than say "20 thou" (0.020") will be fine.

A good carpenter's level that is metal/aluminium, machined and with an accuracy of 0.5mm/metre (1":2,000 = 0.006"/foot = 0.0286 degrees = 1.72 arc minutes) will help a lot too.

lts all pretty effective, easy, simple, quick and cheap to do.

I will cover it in more detail later in the day or tomorrow.

The Artful Bodger
11-18-2009, 12:03 AM
I will be looking forward to what you have to post Oldtiffie though you will have difficulty in convincing me that going out to search for a wrist pin is worth while when I can make a cylindrical square from a bit of scrap under the bench.

J Tiers
11-18-2009, 12:20 AM
In reference to trying to align the lathe portion of my Smithy,

I spent hours,...no days, trying the RDM with a 1" diameter rod. The best I could do was a .001 -.003 taper in 5 inches. Finally gave up and inserted .0002 shims randomly until a fine cut on the lathe was a uniform dimension on a 1.5" cylinder 5" long.


I think you hit on the best possible way of checking the alinement there.

What I would suggest in general is to use a level, NOT the fanciest type, which are too sensitive, but one of about 0.005 per foot sensitivity, and get the bed to the point that level doesn't show a twist.

Then do what you did, although if you rough turn a couple collars on a large diameter piece of pipe it may be easier.

Just take a real slow cut of minimal depth, and measure the two collars. Adjust to correct the discrepancy, if any. When you are happy, you are done.

You can use RDM, so long as you level first. ALL the "self-referential" tests need an "outside arbitrator" to keep them from running off into the weeds. I think RDM is a big pain, compared to a simple two collars test, but if you like the hassle, no problem;)

oldtiffie
11-18-2009, 02:27 AM
I will be looking forward to what you have to post Oldtiffie though you will have difficulty in convincing me that going out to search for a wrist pin is worth while when I can make a cylindrical square from a bit of scrap under the bench.

Just be patient AB and all will be revealed.

A good -preferably new - wrist/gudgeon pin will be very useful.

The faced/squared ends of a good cylinder (that's all a cylindrical square is after all) is only needed if the squared-off end/s is/are required as a reference or datum face or surface. They are not needed here in the method I propose to use.

The cylinder I intend to use needs to be accurately round, bent no more than 20 thou (0.020") and with band-sawn ends that only need to be faced off sufficiently well for the tail-stock to drill a centre in each end. That's all.

The OP will also need a band-sawn bit of mild steel - aluminium and brass will do as well - that is what-ever the Smithy 4-jawed chuck is happy carrying.

I don't know - or care much - how well the lathe is fixed to its stand or how sturdy or stiff the bench or stand is. Its the OP's choice as to what he uses.

Finer/better levels and Rodney's Dad's Method et al may be used as well - later - if required to prove any points. A good set of feeler guages and a good dial indicator will come in handy as well.

I don't know how good/well the OP really needs (as opposed to wants) his lathe to be aligned. That is for him to decide and tell us - it he wants to.

koda2
11-18-2009, 09:33 PM
Okay gents,
Here is what I did after the comments about cylinder squares. I am not sure its what you wanted but maybe you will have some idea.

I chucked up 1" OD Al 2024 stock with ~3 inches protruding from the 3 jaw chuck and faced the end. Then I turned off the rough surface and made a final cut about 4 thou thick with powerfeed.

I parted off the cut area in the middle and at the headstock end creating two cylinders, rotated one and put them on a surface plate to see if they are parallel. There is a sliver of light showing but may be a slight burr from the parting blade, its hard to tell and I will have to chamfer the ends.

http://pages.suddenlink.net/tismuoi9/squares.jpg

The ends measure as shown, .9725 on the headstock end and .9740 on the right hand side. This is basically what I got after randomly placing shims in various places between the bottom of the headstock casting and the ways; about .001 difference in 3" of travel, under optimum conditions: soft metal, light cut, and little pressure (flex?) against the headstock.

Is it good enough for what I am doing now? Yeah it is, but I have given up for now on trying to make really good stuff. I originally wanted to make the headstock alignment as true as possible because a lot of other stuff, i.e., making back plates, collet holders, etc. seemed to me to require that first.

The aircraft builders around here have a saying, "The quality stops the minute you start saying ""its good enough""." And they make some Oshkosh winners.

The reality is that it may be the best the machine and the operator can do at least for now.
Dave A.

darryl
11-19-2009, 01:38 AM
If you're getting .001 difference in diameter over 3 inches, that isn't so so bad. You can do lots of precision work with it. If you want to be able to turn the full length of a longish piece and have both ends be exactly the same diameter, there are many other factors that could come into play anyway. .004 for a finishing cut is easily enough to deflect the outboard end of the workpiece, and if you change the geometry of the cutting tool you could either suck the workpiece towards you, or push it away.

What you're showing is a very slightly larger diameter on the outboard end of the workpiece, so possibly it's pushing away just slightly. That would not be abnormal.

I'd be inclined to repeat the test several times with another piece of round bar. Change one parameter at a time with the cutter, and measure the results.

What I would do is first remove the roughness over the entire length that sticks out, then remove say 50 thou or so of material from two spots, leaving a larger diameter at both ends and one in the middle. Then clean up the cutter and take a very light pass across all three spots, like 2 thou or so. Take another pass without changing anything and see it there's more cutting going on. If all is well, and play is accounted for, you should probably get a slight re-cutting with one 'spring pass', and that should be pretty much it for material removal. If you get no slight re-cutting at all, the tool is probably being drawn in towards the workpiece by the cutting action, and you won't be able to get precise results easily. Largely, this is a function of the angles ground into the cutting tool, it's sharpness, and height with regard to the spindles center of rotation.

So assuming that you are only getting a very slight re-cutting with the spring pass on all three collars, now you can take measurements again.
That would be my first step- to know that I'm not forcing the workpiece in any way as the finishing is being done, and that the finished surface is able to just touch the cutter, proving that the finished surface is right where you are expecting it to be.

I hope this is understandable and not boring. I do have a tendency to go on and on-

Anyway, if your lathe is cutting a taper, it will show in the measurements. If that taper has a curve, that will show also.

There's so many things that can affect this testing, and subsequent finished work also. Bed wear, play in slides, rocking motions anywhere on the carriage and all the stuff that's mounted to it- frictional and torsional effects from the leadscrew-

J Tiers
11-19-2009, 08:46 AM
Anyway, if your lathe is cutting a taper, it will show in the measurements. If that taper has a curve, that will show also.

There's so many things that can affect this testing, and subsequent finished work also. Bed wear, play in slides, rocking motions anywhere on the carriage and all the stuff that's mounted to it- frictional and torsional effects from the leadscrew-

And height of the cutter relative to center. It wants to be on-center within a small error (in percentage of the work diameter).

dockterj
11-19-2009, 08:51 AM
Seems to me (and I'll start out by saying that I am not an expert here) that if you knew that the cross slide was perpendicular to the ways things would be a lot easier. In that case you could put a DTI on the cross slide, pick a spot on the spindle (or the face of the chuck - doesn't matter) and see how much difference there is when the spot is at 9 oclock vs. 3 oclock.

I struggled with this on my smithy too. The way the headstock rides on the back ways was pretty poor on mine, seems to rock. I ended up putting a shim in and calling it good enough.

Is there an easy way to tell if the cross slide is perpendicular? I thought of clamping a 123 block parallel to the ways (using a dti and running the sadlle back and forth) then using the dti while running the cross slide in and out. I think if I repeated that with the 123 block flipped over (using the same edges) any angular error would be shown. My whole shop is in storage right now (in the process of moving) so I won't be able to look at it again for a few months.

aboard_epsilon
11-19-2009, 08:54 AM
That is precisely what I am working on but it includes a way to multiply any error by a factor of perhaps 100 to 1000. This will make it possible to detect an off axis condition of as little as .0001 per foot.

I have already developed a simple method of bed alighment that uses a laser and two 90 degree prisims but have never shown it here because most people can't make use of it. I happen to have about 30 feet of open hallway that I can shine a laser beam down and back which gives a 60 foot angular multiplication of a six inch spacing of the bed ways. A 1 thou twist in one foot becomes a 1/4 inch movement of the laser beam once it has gone down and back.

i tried to use a laser a few months back on my southbend ..
i found that the red line was too thick ..so unable to judge propoperly .

all the best.markj

koda2
11-19-2009, 10:47 AM
The headstock is mounted on the flat ways with a dovetail in the back and two set screws in front. Additionally there are two bolts extending upwards from the bottom of the bed casting, one on each side.
The set screws are small and might work better if they were enlarged and were pressing on a gib.
The two bolts are diagonally placed from one another and I think have to be tightened slowly and in unison lest the alignment be disturbed.

I am pretty sure the cutter height is correct; I used a technique described here and elsewhere to construct a machined height gage that made a huge difference in my ability to center cutting tools.

I will have to do more fiddling and testing; unfortunately its end of year and deadlines loom so the Smithy gets pushed to the rear. :)
Dave A.

Evan
11-19-2009, 11:45 AM
I didn't get a chance to play with it yesterday. When I went down to the shop the power cut out and stayed out for several hours.

The laser dot is adjustable on nearly all laser pointers. They have a collimating lens on the front that usually is a small plastic holder with a serrated edge. It may have a dab of glue to hold it in place but that can usually be cracked just by turning the lens. The beam can usually be adjusted to a very small spot.

It's very hard to judge the size of a dot of laser light because of the interference pattern it makes on any surface that isn't a specular reflector. This can be reduced a lot by wearing a pair of sunglasses that are the complemetary colour to the laser. If you have a red laser then wear green sunglasses and if you have a green laser wear pink glasses. If you have a blue laser then just send it to me for testing and hire sombody else to align your lathe. :D

lazlo
11-19-2009, 11:54 AM
If you have a blue laser then just send it to me for testing and hire sombody else to align your lathe. :D

Evan, if you want a cheap blue laser, HD-DVD's are selling for a song these days (because they lost the VHS versus Beta war).

Kooky, but amusing, video:

http://www.youtube.com/watch?v=vIwaMNab5pw

aboard_epsilon
11-19-2009, 11:57 AM
I didn't get a chance to play with it yesterday. When I went down to the shop the power cut out and stayed out for several hours.

The laser dot is adjustable on nearly all laser pointers. They have a collimating lens on the front that usually is a small plastic holder with a serrated edge. It may have a dab of glue to hold it in place but that can usually be cracked just by turning the lens. The beam can usually be adjusted to a very small spot.

It's very hard to judge the size of a dot of laser light because of the interference pattern it makes on any surface that isn't a specular reflector. This can be reduced a lot by wearing a pair of sunglasses that are the complemetary colour to the laser. If you have a red laser then wear green sunglasses and if you have a green laser wear pink glasses. If you have a blue laser then just send it to me for testing and hire sombody else to align your lathe. :D

had it spread ot to form a line ...you know where the line projects every-where at 180 degrees .
the idea was not to trust the bubble on the laser.........but have a clear plastic pipe with water in .......each end 10 foot apart .......project the laser onto these .......and then along the lathe bed at the same time.

line about 2mm thick .so no good .

may try the red gogles idea...then perhaps it will be less than 1mm thick and more focused.

all the best.markj

Evan
11-19-2009, 01:50 PM
Water won't work because of the high surface tension. Use alcohol instead.

J Tiers
11-19-2009, 10:53 PM
or put a bit of detergent in the water, that also breaks the surface tension and eliminates the problem to a decent degree. I've used that in water manometers for cooling fan pressure drop, etc.


Are we maybe overcomplicating this? Started as an alignment problem, which should be fairly easy to solve with the "two collars" test, or equivalent.

Now we are on laser alignment etc. I just can't suppose the real problem is that complex. Problems like this were solved years ago without lasers etc........

Evan
11-19-2009, 11:32 PM
Now we are on laser alignment etc. I just can't suppose the real problem is that complex.

Whay does "laser" somehow infer "complex"? As has been said, the Laser is a solution looking for a problem. Alignment of the lathe is a problem and a laser can simply provide a solution.

oldtiffie
11-20-2009, 12:53 AM
or put a bit of detergent in the water, that also breaks the surface tension and eliminates the problem to a decent degree. I've used that in water manometers for cooling fan pressure drop, etc.


Are we maybe overcomplicating this? Started as an alignment problem, which should be fairly easy to solve with the "two collars" test, or equivalent.

Now we are on laser alignment etc. I just can't suppose the real problem is that complex. Problems like this were solved years ago without lasers etc........

Right on JT.

I think that you are keeping it simple and effective and at a skill and tool level that most here can handle.

Back to basics and fundamentals.

I haven't been able to get back and get to grips with this thread and its topic yet - but I will.

It IS basic -as you say.

It SHOULD be able to be at least identified and, to a large extent, be rectified to an acceptable level by just about anyone here with tools that are either in the shop or able to be be got easily and cheaply and used in the shop for other stuff as well.

The title of the original post (OP) was -and still is:
Headstock alignment.

The body of the OP (Dave) was - and still is:

In reference to trying to align the lathe portion of my Smithy,

I spent hours,...no days, trying the RDM with a 1" diameter rod. The best I could do was a .001 -.003 taper in 5 inches. Finally gave up and inserted .0002 shims randomly until a fine cut on the lathe was a uniform dimension on a 1.5" cylinder 5" long.

Part of the problem is the puny and tedious way the headstock is locked down.

My progression of alignment attempts:
-Build a rock-solid table with threaded feet.
-Level the table.
-Rebuild the spindle with better bearings and preload as best I could (result=TIR 3 tenths or less on the spindle taper.)
-Did not scrape the ways; they were new and ground by the factory. There was no appreciable twist and the carriage slides uniformly easily down the ways with the gibs fairly snug.
-Level the lathe bed with Starett precision level fore and aft and side to side.
-Remove the headstock, clean all grit, paste, paint from joining surfaces and reinstall. Then align with RDM and 20 inch aluminum cylinder. No joy.
-I believe the crosslide is perpendicular to the ways but don't know how to check it.
-The tailstock is still a mess. The ram is sloppy in the tailstock bore. I will have to figure out how to bore out and recylinder the tailstock eventually but the headstock alignment (I believe) takes precedence and is more important.
My next attempt when time permits will be to try to rig up a laser pointer down the center of the spindle bore and check alignment on a tiny centered hole at the other end, but I don't think it will give precise enough readings by itself.
Dave A.

I rather got the impression here that the OP had pretty well given it away as a bad cause/job - and I don't blame him:
http://bbs.homeshopmachinist.net/showpost.php?p=488580&postcount=18


The headstock is mounted on the flat ways with a dovetail in the back and two set screws in front. Additionally there are two bolts extending upwards from the bottom of the bed casting, one on each side.
The set screws are small and might work better if they were enlarged and were pressing on a gib.
The two bolts are diagonally placed from one another and I think have to be tightened slowly and in unison lest the alignment be disturbed.

I am pretty sure the cutter height is correct; I used a technique described here and elsewhere to construct a machined height gage that made a huge difference in my ability to center cutting tools.

I will have to do more fiddling and testing; unfortunately its end of year and deadlines loom so the Smithy gets pushed to the rear. :)
Dave A.

I will sit this out for a while - "on the sidelines" - and continue to be amazed at just how far the topic has "drifted" without comprehensively addressing the OP in a meaningful way - JT is the stand-out exception amongst a few here in that regard.

Ed P
11-20-2009, 08:47 AM
In reference to trying to align the lathe portion of my Smithy,

-I believe the crosslide is perpendicular to the ways but don't know how to check it.
Dave A.

Place a straight edge in the chuck so that the edge is parallel to the cross slide travel. Attach a dial indicator to the cross slide. Adjust the straight edge until the indicator reads the same at each end of the straight edge, i.e. check one end of the straight edge then roatate the headstock spindle 180 degrees and check it again. The indicator should read the same. Now run the cross slide in and out so that the dial indicator finger runs along the straight edge, the reading will not vary *if* the cross slide is perpendicular to the headstock spindle.

Ed P

J Tiers
11-20-2009, 08:54 AM
If the machine is adjusted, as many are, to be "at most" parallel, with any errors on the side of facing concave, you *may* find an error off perfectly perpendicular in that test, yet machine might be perfectly within spec.

An error to make facing conVEX would be undesirable, but would obviously also give an error off parallel.

Thanks Tiffie for the kind words....

Evan
11-20-2009, 09:19 AM
Now run the cross slide in and out so that the dial indicator finger runs along the straight edge, the reading will not vary *if* the cross slide is perpendicular to the headstock spindle.


It's not the spindle Dave want to check against. It's the ways.

Evan
11-20-2009, 09:28 AM
Clamp a square in the tool holder so one edge is parallel to the bed and the perpendicular edge is exposed for measuring too. Fasten a dial guage to some convenient location on the bed. Adjust the edge of the square that is parallel to the bedways to be actually parallel with carriage travel using the gauge. Then dial the perpendicular edge to see if the cross travel is square.

Ed P
11-20-2009, 11:06 AM
It's not the spindle Dave want to check against. It's the ways.

Let me clarify, the procedure I provided is to be used after the headstock
is aligned to the ways. Perhaps some do it differently but I would verify the bed is flat then get the headstock parallel then check the cross slide.

Ed P

koda2
11-21-2009, 12:14 AM
Okay, one day for shoveling decorative rock and finishing landscaping before the weather turned and one day in back recovery mode, and I am focused.

I completed the check of the cross slide:
http://pages.suddenlink.net/tismuoi9/crosslidecheck.jpg

Its as close to dead on as it can be; the dti barely flickered. That makes me feel good.

The cross slide check should be largely independent of the first three important requirements: 1) bed flat and untwisted, 2) spindle free of excessive runout, and 3) headstock(or rather spindle) aligned with the ways.

If I have those four, one would think I should have the basis to be able to work accurately at least with the headstock side of the lathe.

Dave A.

oldtiffie
11-21-2009, 01:30 AM
Dave.

You are going well.

Looks good so far.

I will deal with lathe bed, head-stock, cross-slide and tail-stock in pretty well that order. It should be sufficient for you to get a pretty good appraisal of your lathe so that you can refine the processes further if you see a need to.

So that I can focus a bit more directly upon your lathe as well as the general case, could you please post the following details of your lathe:
- head-stock spindle bore;
- capacity of 4-jaw chuck;
- distance between centres;
- distance between end of the lathe bed and the back of the carriage with the carriage up against the head-stock;
- distance between front of the lathe bed (nearest the head-stock) and the front of the carriage with the carriage as far back toward the end of the lathe-bed as it will go;
- morse taper in the head-stock spindle; and
- morse taper in the tail-stock quill;
-the maximum size of the capacity of the drill chuck that you use in your tail-stock.
- an image of your lathe;
-a web page with your lathe details.

oldtiffie
11-21-2009, 02:28 AM
Originally Posted by The Artful Bodger
I will be looking forward to what you have to post Oldtiffie though you will have difficulty in convincing me that going out to search for a wrist pin is worth while when I can make a cylindrical square from a bit of scrap under the bench.


Just be patient AB and all will be revealed.

A good -preferably new - wrist/gudgeon pin will be very useful.

The faced/squared ends of a good cylinder (that's all a cylindrical square is after all) is only needed if the squared-off end/s is/are required as a reference or datum face or surface. They are not needed here in the method I propose to use.

The cylinder I intend to use needs to be accurately round, bent no more than 20 thou (0.020") and with band-sawn ends that only need to be faced off sufficiently well for the tail-stock to drill a centre in each end. That's all.

The OP will also need a band-sawn bit of mild steel - aluminium and brass will do as well - that is what-ever the Smithy 4-jawed chuck is happy carrying.

I don't know - or care much - how well the lathe is fixed to its stand or how sturdy or stiff the bench or stand is. Its the OP's choice as to what he uses.

Finer/better levels and Rodney's Dad's Method et al may be used as well - later - if required to prove any points. A good set of feeler guages and a good dial indicator will come in handy as well.

I don't know how good/well the OP really needs (as opposed to wants) his lathe to be aligned. That is for him to decide and tell us - it he wants to.

In the interim, I thought I'd show those that may not realise what a "Cylinder master square" is.

Its is nothing more than a cylinder that is accurately round, straight and the same diameter and distance over the ends.

Mine are made from precision-drawn steel tubing - 3" outside diameter with a nominal wall thickness of 1/4".

The cylinder was very accurately set up on the lathe with minimal Total Indicated Run-out (TIR) at both ends using my Test Dial Indicator - TDI - (0.01m ~0.0004" calibration). It was accurately set up and supported in my lathe fixed steady rest and faced-off for a smooth finish on the lathe. The ends were then very accurately square to the cylinder.

I tested them against my Class 1 machinists square, my "Frame Square" and last but not least, each cylinder was tested against its pair and all other cylinders "end for end" and at the four cardinal (90 degree) positions at each end. They were excellent.

I used them to "recover" and "square up" my cast-iron angle-plates - which also came out very well as regards accuracy and finish too.

I will not need these for checking the OP's lathe - the subject of this thread.

http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Cylinder%20master%20squares/Squaring1.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Cylinder%20master%20squares/Squaring2.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Cylinder%20master%20squares/Squaring3.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Cylinder%20master%20squares/Squaring4.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Cylinder%20master%20squares/Squaring5.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/measuring/Cylinder%20master%20squares/Squaring6.jpg

Evan
11-21-2009, 04:18 AM
I will presume that you have already aligned the bed so that it is parallel and the ways are coplanar.

Make a holder for the laser that allows it to be placed securely in the left end of the spindle. You will need to provide a means to adjust the aiming of the laser. Here is an example of this using two triads of setscrews to hold the laser pointer in a tube. The plastic knob screws in to actuate the on/off switch. This device serves a very similar purpose as it is used to align my telescope.

http://metalshopborealis.ca/pics/coll1.jpg


Align the laser with the spindle axis by projecting the laser to a point as far away as possible. Place a mirror there and angle it slightly to bounce the beam back to the area of the carriage. Place a piece of clean cardstock or similar to use as a screen, The point of reflecting the beam back is to lengthen the baseline and to make it easier to observe.

Rotate the spindle by hand and adjust the aiming of the pointer until the spot stays pointed at the same place during a full rotation of the spindle.

Now replace the tailstock if it was removed. Place a dead centre in the tail stock. The laser beam should fall directly on the point of the centre. Cut a hole in a piece of cardstock just big enough so that it can be placed on the nose of the dead centre. If the beam is pointing to the centre a specular reflection pattern will be seen on the cardstock surrounding the dead centre.

This reflection pattern is extremely sensitive to changes in the alignment of the beam and will easily show a change of .001" in centering over the length of the bed in the aiming of the laser. If the beam in directly centred on the point of the centre the pattern will be approximately equal all around the dead centre with some symmetrical light and dark areas.

If the beam is not falling on the point of the centre then move the tailstock close to the chuck and secure.

If the laser is still not aligned with the centre then adjust the headstock until it is. The spot size of the laser isn't critical since it is the reflected pattern that matters and it will be symmetrical when the beam is on target.

Loosen the tailstock enough that it will slide and use the carriage to rack it to the full right position. If the beam is not on centre then adjust the headstock unti it is. Then move the tailstock back to the left up to the spindle. If the beam is not on centre the adjust the TAILSTOCK until it is.

Repeat this sequence until both positions result in alignment being maintained. This method is sensitive but is also subjective depending on the observer. I am developing a more sensitive and less observer dependent method but don't yet have it working to my satisfaction.

Once you have performed this alignment then use the conventional method of turning test pieces to test the deviation from true.

koda2
11-21-2009, 08:32 PM
Oldtiffie,
Here are the specs and picture of the lathe.
Smithy CB-1220XL mid 90s once the top of their 3-in-1 line, now continued as their entry level Midas 1220

head-stock spindle bore - 1.03"
capacity of 4-jaw chuck - 6" dia
distance between centres - ~20"
distance between end of the lathe bed and the back of the carriage with the carriage up against the head-stock -19"
distance between front of the lathe bed (nearest the head-stock) and the front of the carriage with the carriage as far back toward the end of the lathe-bed as it will go(without removing tailstock) - 17"
- head stock - MT-4
-tail stock - MT-3
-drill - 1/2"
http://pages.suddenlink.net/tismuoi9/smithypix.jpg
Dave A.

oldtiffie
11-21-2009, 09:43 PM
Oldtiffie,
Here are the specs and picture of the lathe.
Smithy CB-1220XL mid 90s once the top of their 3-in-1 line, now continued as their entry level Midas 1220

head-stock spindle bore - 1.03"
capacity of 4-jaw chuck - 6" dia
distance between centres - ~20"
distance between end of the lathe bed and the back of the carriage with the carriage up against the head-stock -19"
distance between front of the lathe bed (nearest the head-stock) and the front of the carriage with the carriage as far back toward the end of the lathe-bed as it will go(without removing tailstock) - 17"
- head stock - MT-4
-tail stock - MT-3
-drill - 1/2"
http://pages.suddenlink.net/tismuoi9/smithypix.jpg
Dave A.


Thanks Dave.

All that info is just what I needed.

I needed the bore size to settle on the test piece maximum diameter (1"/25mm) and the distances to the carriage for the length of the level so as to not have to "bridge" the level on the bed to get over the saddle. A 12" or 15"" (16" max) length level will do fine.

Having a flat bed lathe with the "vee" bevel underneath the bed makes using the level across the bed easy as there is no "vee-guide/s" sticking up that we have to compensate for or to "bridge".

The 4-jaw capacity means that the facing test piece can be - say -anywhere from 4">6"(max) which is fine.

I have a "3-in-1" that find is really good -
http://i200.photobucket.com/albums/aa294/oldtiffie/Lathe_misc/Lathe1.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Lathe_misc/Lathe2.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Lathe_misc/AirSmith09.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Lathe_misc/AirSmith06.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Lathe_misc/AirSmith01-1.jpg

I only recently took the "mill" off to fit a taper-turning attachment to the back of the lathe bed.

I have a 6" "Sieg" lathe that is pretty similar to yours as mine has the "Vee-guides" under the flat bed, the lead screw (no rack) is under the centre of the lathe bed and is fed from the end of the lathe bed. The cross-slide has full-length "Tee-slots" as yours does. I have all the "milling" and other accessories for it and it works pretty well within its capacity.

http://i200.photobucket.com/albums/aa294/oldtiffie/Micro-lathe/Micro-lathe1.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Micro-lathe/MIcro-lathe2.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Micro-lathe/MIcro-lathe3.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/Micro-lathe/Micro-lathe4.jpg

That's not rust - its the "Lanolin" that I use as preservative/rust-preventative.

Don't take too much notice of the anti-"3-in-1" brigade either. There is a lot of capacity within that quite small foot-print. I have read some very good reviews on the "Smithy".

I will get back to this later in the day as we have some shopping and other stuff to do.

oldtiffie
11-22-2009, 07:39 AM
For the OP - koda2 - Dave:

Dave.

I really do apologise for running late with giving this post the attention it not only deserves, but what I under-took to do.

As will be obvious, I got myself side-tracked on other issues and other threads.

I have pretty well collected and collated all I need to address in/for my response and should get it done tomorrow (here in OZ).

This is by way of both a "bump" (so that I will see it) and a kick in the ar$e as a prompt to "get my finger out".

Again, my apologies.

koda2
11-22-2009, 12:46 PM
Oldtiffie,

Certainly no apologies needed. I suspect that most everybody, myself included, has multiple issues demanding time and energy. Throw in the holidays coming up as well. This forum's members give freely of their time and knowledge.

Besides, I have plenty of "pots on the stove that need stirring" in the meantime.

BTW, that 3-in-1 is pretty impressive and looks like it could handle anything you throw at it.

Dave A.

doctor demo
11-22-2009, 01:17 PM
I will presume that you have already aligned the bed so that it is parallel and the ways are coplanar.


Align the laser with the spindle axis by projecting the laser to a point as far away as possible. Place a mirror there and angle it slightly to bounce the beam back to the area of the carriage. Place a piece of clean cardstock or similar to use as a screen, The point of reflecting the beam back is to lengthen the baseline and to make it easier to observe.

Rotate the spindle by hand and adjust the aiming of the pointer until the spot stays pointed at the same place during a full rotation of the spindle.



.


Evan, does the quality of the mirror matter much? ( I don't mean a polished turd) but is there a preference?

Also ,when reflecting the beam would there be any advantage to reflect it past the lathe to a point farther (like the other wall)?

Steve

Evan
11-22-2009, 01:27 PM
The mirror doesn't make much difference. It's easy to tell when the spot is rotating with minimum deviation. It will help to refocus the laser for this particular task and then focus again to a point somewhere along the lathe bed.

You need a baseline at least several times longer than the lathe bed but that also isn't critical.

The Artful Bodger
11-22-2009, 01:48 PM
Evan, I do not see how you ensure the laser is mounted exactly in the centre of the spindle which is surely vital to your method?

aboard_epsilon
11-22-2009, 03:09 PM
Evan, I do not see how you ensure the laser is mounted exactly in the centre of the spindle which is surely vital to your method?

he's said it all

the further away you have the laser projected ...the more it will trace any out of centre wobble on what it is projected onto ..when you turn the spindle by hand .

i think he than goes onto say .by putting a target (piece of white paper) on the tailstock pushed over the dead centre so that the centre only protrudes a couple of mm through the paper ......
you will then be able to see the diffraction of the light ..and sought of work out if you're on centre or not ..

thats how i see it ..and make of it ...the projection bit on a wall will work great .the farther away the better ..but not sure about trying to line up the tailstock with it........weather it will work out in practise or not ..
how about some pics Evan ..of you doing the latter part.

all the best.markj

The Artful Bodger
11-22-2009, 03:26 PM
he's said it all

the further away you have the laser projected ...the more it will trace any out of centre wobble on what it is projected onto ..when you turn the spindle by hand .

i think he than goes onto say .by putting a target (piece of white paper) on the tailstock pushed over the dead centre so that the centre only protrudes a couple of mm through the paper ......
you will then be able to see the diffraction of the light ..and sought of work out if you're on centre or not ..

thats how i see it ..and make of it ...the projection bit on a wall will work great .the farther away the better ..but not sure about trying to line up the tailstock with it........weather it will work out in practise or not ..
how about some pics Evan ..of you doing the latter part.

all the best.markj

Mark, yes I see that, I think. However consider if the laser were mounted off centre in the spindle by adjusting the screws it could still be made to point to a single point on the far away screen. In that case as the spindle is rotated the beam of the laser would follow a cone with its apex at the screen but the point would be off centre at the tip of the tailstock. Or did I miss something?

Evan
11-22-2009, 03:29 PM
I will see what I can do but don't count on it. We have unexpected company, a very good friend from the coast who showed up here on Friday because her mother was gravely ill. Her mother passed just this morning so our schedule is shot completely and I have no idea what I will have time for.


Test the concentricity by making sure it stays on the spot both near and far. That is how I calibrate my telescope alignment tool.

aboard_epsilon
11-22-2009, 03:35 PM
behind that focus point of the cone ..will be a cross over cone ..so you better make sure that ..this does not happen i supose ....by making sure that the wall isnt that exact point.......or moving the point projection area closer and adjusting on that .

or make sure you start off...with a spreading cone and not a converging one ...then make fine adjustments from that.

and it would better if you could adjust the laser without taking the laser out of the spindle each time.

all the best.markj

oldtiffie
11-23-2009, 02:13 AM
Oldtiffie,

Certainly no apologies needed. I suspect that most everybody, myself included, has multiple issues demanding time and energy. Throw in the holidays coming up as well. This forum's members give freely of their time and knowledge.

Besides, I have plenty of "pots on the stove that need stirring" in the meantime.

BTW, that 3-in-1 is pretty impressive and looks like it could handle anything you throw at it.

Dave A.

Thanks Dave -appreciated.

I will handle this progressively with successive posts over what I hope will be a relatively short time.

Level:
If you have access to a mill or machine table that is in reasonable condition, it is probably fair to say - subject to checking - that it will be pretty flat. Failing that - use your cross-slide as it will not have been stressed by any bolting down as your lathe bed may have been.

Now put level on that "flat" and check to see if there is a "bow" in your level. Have the bow touching at the ends and clear of the table at the centre. If the bow does not exceed about a "thou" (0.001") - or so - it will be OK. Turn it over (bow down) and see if it "rocks" by pressing down on the ends.

If still a thou or so (or less) proceed to check the level vial/"bubble" centreing. Put a good parallel strip in your vise on your bench - yep - that's good enough, put the level on the parallel and tap the parallel until bubble is centred. Now turn the level end-for-end. If the bubble is still centred - job done. If not centred, tap the parallel strip until half the error is removed. Mark the vial for one of the bubble ends. Now turn the level end for end again and check to see that the end of the bubble still touches the mark you made. If all OK - job done and the level true centre is known - and marked. If not OK go back and repeat until it is. Please DO NOT remove the mark on the vial as it is the reference for all leveling.

As a short note here, as I said that the level should be machined flat and have an accuracy of 0.5mm (~0.020") per metre (~40") that is a maximum error of 0.020"/40" = 0.0005" per inch - at most, which for this purpose thus far and for a while yet is not too bad at all. The "Starrett" no.???? super-good level will be a real PITA at this point and for some time yet.

The lathe bed:
Free the lathe bed from any clamping-down so that it is unstressed. Leave it for a couple of hours to "ease" any stresses. At the same time, run the lathe at medium speed top get to its normal operating ambient temperature.

Remove the tail-stock from the bed.

Move the saddle toward the tail-stock end of the bed - but leave it say 1" or so clear of the end so that the level can span the bed at its end.

Use the level on the lathe bed in pattern that is a rectangle that has diagonals as well as bisecting opposite sides - similar to the "Brit" "Union Jack" flag:
http://en.wikipedia.org/wiki/File:Flag_of_the_United_Kingdom.svg

http://en.wikipedia.org/wiki/Union_Jack

Checkeach "line" "both ways" ie "end-for-end" with the level.

Jack or adjust the the lathe bed until it is flat as you can get it - use the mark on the level vial for reference.

Use one point or "corner" of the leveling pattern as a reference. Use shims or feeler guages or a good dial indicator to see how much rise or fall there is across each "line" in the leveling pattern. Carefully record these results.

Use the level as a straight edge to see if there is any lateral "bowing" of the lathe bed. Note/record this as well.

You now have a pretty good picture of the "straightness" or "flatness" of the top face of your lathe bed.

The lathe bed is the basis of and reference for everything else that follows.

Now re-fasten the lathe bed to the bench getting it at least as good and preferably better as regards level that it was when measured in its previous free state.

I will be surprised if any later fastening or "corrective??" measures will make much difference as I'd guess that the lathe bed is pretty stiff in terms of resistance to bending (up-down and in-out) or torsion (twisting) and that the bench is stiff enough for that as well as resisting a lot of distortion due to loads on the bench.

The lathe is now as "flat" and as accurately in the horiziontal plane as it is practical to achieve and maintain -for the purposes of these tests and the future use of the lathe.

So far - so good.

I will address the head-stock alignment and accuracy of facing tests in the next item/post.

The rest of the tests will follow in order after that.

I hope this helps.

If you have any queries or comments to or for me, please don't hesitate to post them here or PM or email me. I will be happy to oblige and answer as best I can.

oldtiffie
11-26-2009, 06:27 AM
This is a "bump" so that I am more likely to see it and "get my finger out" and progress it better than I have lately.

oldtiffie
11-28-2009, 01:51 AM
OK.

On we go.

I will try to cover the head-stock spindle and the cross-slide today.

Head-stock:

First of all, I need to make the point that when checking a shaft for alignment in a - in this case - chuck, but the case applies for collets too - to be used in a static (ie non-rotating) state that all that is required is that test shaft be accurately round/circular and straight. The shaft is OK if the Total Indicated Run-out (TIR) as indicated on a good - say 0.0005"/0.01mm (~0.0004") Test Dial Indicator (TDI) is within 0.0002" at the face of the chuck and at the end of the test piece. The TIR maximum and and minimum amounts must be within say 5 degrees (as measured at the test piece and/or the lathe head-stock spindle axis) of eachother. It is pretty well the same as a crank-shaft journal (for a connecting rod) that is off-set from but accurately parallel to the crank-shaft main bearing journals. It is essential that any "conical" misalignment of the outer end of the test piece be as close to zero as possible (ie it should not rotate as if it was "bent").

Mount the test piece in the 4-jaw chuck and adjust and "dial it in" with the TDI to have/show/indicate minimum TIR (see above).

I'd suggest just having about 1/2" (most) gripped by the chuck as it will be easier to "tap" the outer/other end to "true up" the other end. Put small copper wire wire between the jaws and the test peice as this may assist. I am making use of or emulating chuck jaw "bell-mouthing" here as it will assist.

The test piece is now accurately aligned and parallel to the head-stock axis. If it is eccentric (half the TIR) within the limits required (see above) to the HS axis it does not matter here.

Put a good TDI base/holder on the lathe tool-post, cross-slide or saddle/carriage (your call as its more or less all the same). Put the TDI on the test piece as near centre height as you can reasonably get it.

Rotate the test piece (by hand - use the chuck - not the test piece) and stop at the point that is mid-way between the TDI defection limits and leave the spindle and test piece in this position.

Now zero the indicator.

Run the lathe carriage left/right within the length of the test piece.

The mis-alignment of the lathe head-stock spindle axis to the lathe bed axis over the length of the test piece will be indicated directly with/on the TDI.

This can be resolved in terms of say 0.000?"/inch (aka "tenths per inch") or angular of say 0.000? degrees or 0.00d:0.00m:??.??s (degrees:minutes:seconds).

Put the TDI on the top ("top dead centre") of the test piece to test for "tilt" of the head-stock axis in the vertical plane - as for lateral (see previous).

Head-stock alignment testing is now completed.

Do not (re?)adjust the head-stock alignment at this stage -wait until after the cross-slide is tested (next item).

Cross-slide:
I will come back to this later in the day or tomorrow.

Walter
12-08-2009, 02:37 AM
Bump !

OT, you gonna finish your tutorial, I'm interested in the rest of it as I'm sure others are.

--Walter

beanbag
12-08-2009, 03:18 AM
[procedure snipped]


I initially did it very similar to this, and later realized two possible errors:

One is that the indicator exerts a force on the rod sticking out, and will deflect it towards the end.

Two is that you have to make sure the rod is not slightly tapered.

oldtiffie
12-11-2009, 07:47 PM
I will proceed fairly slowly over time with this thread - on an "occasional" basis.

Cross-slide:
The assumption here is that the cross-slide is straight or not too badly worn although this should show those conditions up.

This procedure will identify whether the cross-slide is "square" to the lathe head-stock spindle axis or not.

If the cross-slide is square to the head-stock spindle axis, the surface cut by the cross-slide will be "flat".

If it is not "square" the surface cut by the cross-slide will be either:
a.
a concave ("hollow") cone with its top/"point" pointing toward the head-stock, or:

b.
a convex" ("solid") cone with its top/"point" pointing toward the tail-stock.

The "error" - either angular or in terms of "thous per inch" (0.00?"/") will need to be considered with any error in the alignment of the head-stock spindle axis to the lathe bed longitudinal axis (discussed previously). If there are any errors in the head-stock spindle and cross-slide alignments, any corrective measures will need to be prioritized in terms of the lathe performance and accuracy.

In an ideal case where there are any head-stock spindle or cross-slide alignment errors, any corrective action on the head-stock spindle alignment will also improve, cancel-out or minimize any cross-slide alignment errors - but it may not.

My guess is that most will opt for the best attainable head-stock spindle axial alignment and just "put up with" or "allow for" any cross-slide error/s.

Correcting any cross-slide errors independently of and without "up-setting" a head-stock spindle alignment will be quite difficult and it will or may require corrective action (ie manual "scraping") of either or both the lathe saddle/carriage inverted "v"-way/s and/or the cross-slide dove-tails on the lathe carriage and/or the cross-slide.

Here is a sketch of the diagram of the geometry of the test cut and measurement of the cross-slide cut:
http://i200.photobucket.com/albums/aa294/oldtiffie/Sketches/Face-platecheck1.jpg

Set the test piece up in a 4-jaw chuck, set the lathe tool to centre height and lock/clamp the saddle/carriage.

It will not matter - in fact it may assist - if there is a hole in test piece.

Run the lathe/head-stock until it has "warmed up" to normal "working temperature".

Face-cut the test piece until it is cut from edge to centre, than take a light facing cut with a good very sharp tool (your choice of tool shapes/angles).

Mark the cross-slide for the cross-slide positions at the start and finish of the cut.

Withdraw the cross-slide until it is at the starting point of the cut. Put a good dial indicator on the cross-slide and set it at the inner edge of the faced cut (on the "opposite" or "back" side) of the faced cut (see sketch). "Zero" the indicator.

Wind the cross-slide forward until the indicator stylus/probe is at the "back" of the faced cut.

Read-off the indicator/indicated needle/pointer "deflection".

The indicated deflection is twice/double the actual/real error of the cross-slide error.

I will address the tail-stock and testing for "parallel turning" as well as types of eccentricity in due course - on a "progressive" and "casual" basis - as advised previously.

koda2
12-21-2009, 04:39 PM
You probably thought this thread was dead. It probably should have been for me.

I finally forced myself to dig into the Smithy again. I spent a good deal of time (many days) in the last several weeks using the tips given here with limited success.

First, I tore the machine down to the bed again looking for any discrepancies that I might have missed and I did find a couple. One was a bump in the dovetail groove on the saddle which had not been machined completely and was interfering with the bed way. Here is a picture after I finished carefully filing it down:

http://pages.suddenlink.net/tismuoi9/groovefixed.jpg

The detents on the carriage gib did not match the set screws correctly and I used a dremel tip and a drill press to make some uniform holes:

http://pages.suddenlink.net/tismuoi9/gibrelocate.jpg

I had always wondered why the gib was hard to set and had to be constantly fussed with and fixing these things did help that a noticeable amount.

The bed level was rechecked multiple times and was identical on each end and middle with my Starrett level. The spindle bore also still had less than .0003 TIR and so was the chuck indexing face.

I initally was going to try the laser method but I toasted my laser pointer trying to adjust the collimation and by that time I had located a wrist pin from an aircraft engine that was big enough to run the alignment procedures. Close examination showed some wear on one end of the pin but I tried to compensate for it in doing the measurements:

http://pages.suddenlink.net/tismuoi9/wristpin.jpg

After setting up the pin in a 4-jaw and making some changes to the alignment, I turned another aluminum round. I got it to a thousandth in almost 4 inches but interestingly enough the result was slightly convex in the middle. This was after taking a last light cut and then going back with no advancement of the cross slide:

http://pages.suddenlink.net/tismuoi9/results.jpg

I also did a face cut on a larger round and although it doesn't show well here it blacked out both the square and a granite plate with the faced edge down. Cross slide is good.

http://pages.suddenlink.net/tismuoi9/facecut.jpg

For now I have come to grips with the reality that you can only polish a ....metal... so much. Plus, I still have some other big fish to fry. Thanks to everyone who provided tips on how to proceed.
Dave A.

koda2
02-23-2010, 09:02 PM
I couldn't give up on being able to turn a cylinder on my Smithy so I scrounged up a hundred bucks for a cylinder square(on sale from Enco) and threw another twenty or thirty hours putzing with the machine again. I figured using better testing equipment would give better results.

http://pages.suddenlink.net/tismuoi9/spicylinder.jpg

The cylinder gave mixed results. It was easy to dial in on the chuck side but couldn't get repeatable results on the other end. I was able to convince myself the vertical alignment was good but after repeated tries of shimming the headstock, the result was that I was now turning a bigger taper to the tailstock instead of to the headstock.

The only misalignment that could cause that would be a headstock pointing towards the operator, so I eventually resorted to just adding shims until I got to less than a thousandth in almost 5 inches. Good enough!

http://pages.suddenlink.net/tismuoi9/shim200.jpg
http://pages.suddenlink.net/tismuoi9/largestock.jpg


I can't explain why the cylinder didn't work. I tried again after finishing the alignment procedures and the cylinder indicated that the spindle alignment was still off. The cylinder was solid, and appears to be well made but it weighed 11 lbs. The 6" 4jaw chuck weighed an additional 20 lbs or so, so maybe the weight hanging off the headstock somehow affected things. At any rate, I can now turn a cylinder at least.

Thanks to all who offered tips in alignment procedures.
Dave A.