-
Welcome back Guest! Did you know you can mentor other members here at H-M? If not, please check out our Relaunch of Hobby Machinist Mentoring Program!
Another lathe leveling question
- Thread starter Clunker1
- Start date
- Joined
- Oct 11, 2016
- Messages
- 3,897
Thanks for thoughtful responses. Headstock alignment with a morse taper test bar would be a good way to go. My searches on this approach found that this depends on the quality of the test bar. A reliably accurate morse taper test bar is really pricey. A lot of posts I read said the asian test bars were not reliable and were not worth the purchase. The approach someone had to face a large diameter disc seemed to have a lot of merits.
Dabbler, your post said you indicate in the headstock and you shoot for .0002 in 4 inches. I'm not sure I understand what you mean. Can you elaborate and explain how you indicate the headstock?
Dabbler, your post said you indicate in the headstock and you shoot for .0002 in 4 inches. I'm not sure I understand what you mean. Can you elaborate and explain how you indicate the headstock?
Firstly, I hope you are feeling better.
A more accurate level doesn't just show what is level, it shows how much out of level it is. If your bed is out of level the same amount up and down the bed, then bed twist isn't an issue.
Adjust the bed using a precision level to minimize twist, then align the headstock so that the spindle is at right angles to the bed. You check that by chucking up a large flat workpiece and facing it off. Run a spring pass to make sure flex from tool pressure didn't undermine the test. Check the face for flat using a straightedge and feeler gauges, but that is accurate only to the thousandth. Is the faced-off part dished? High in the center? If you have a good surface plate, sweep the part after facing--even if one side of the test part is higher than the other when resting on the plate, the center should be exactly in between the two edges as you sweep it. Use a tenths-reading dial test indicator. But you can save that for the final adjustment--you can get it to the thousandth using a straightedge and feeler gauges without removing the test piece from the chuck.
Adjust the headstock until a facing operation produces a truly flat surface when facing--that's the only job of headstock adjustment. That ensures that the cross slide is at right angles to the spindle. After doing that, check that the tailstock is truly aligned with the spindle by pinching a razor blade between centers. If it pinches at right angles to the spindle and not cocked at an angle, the tailstock is aligned. Adjust the tailstock until that happens. If it's off a little vertically but not horizontally, that's probably the best you can do.
Then, check to see if it's turning tapers using the two-collar method (or use a certified test bar like the one from Edge Technologies). If it is, adjust bed tilt slightly, realign the tailstock, and check it again. After all that, face off something large again just to make sure you didn't undo anything related to the head alignment.
The head alignment is only for making sure that the cross slide moves at a perfect right angle to the spindle, not for curing tapered turning. It will cause tapered turning if out of alignment, but fixing the tapered turning with the headstock alone will likely leave it out of square with the bed.
TL;DR: Level the bed with the best level you have. Align the headstock so that the cross slide is square with the spindle and prove that by the flatness of a faced part. Align the tailstock to the spindle using the razor-blade test. Check for tapers using the two-collar test. Adjust tilt and realign the tailstock and repeat until the lathe doesn't turn a taper. Check again for the flatness of a faced part--if that was lost, do it all again, but you'll be a lot closer this time.
Rick "be glad you have an adjustable headstock--most of us don't" Denney
Last edited:
Need following this discussion and you make a ton of sense. So there’s no way to actually indicate the headstock in? You have to turn a test piece?
- Joined
- Feb 1, 2015
- Messages
- 9,724
A test bar will not determine headstock misalignment. To use one, you would determine the spindle centerline distance relative to the bed and if there is bed twist, it will also affect the measurement. I determine headstock misalignment by measuring the perpendicularity of the cross slide travel to the spindle axis with a dial test indicator. I do so by mounting a boss on my faceplate. The boss allows the measurement point to stand above the faceplate surface, eliminating the need to move the carriage or compound for clearance. I set the boss up as far from the center as is consistent with cross slide travel. Then I rotate the faceplate so the boss is horizontal with the spindle axis to the front and position the test indicator so it is centered on the boss and zero it. Next, I rotate the faceplate 180º and center the test indicator on the boss, rotating the faceplate slightly if necessary.
If the headstock is correctly aligned, the test indicator will again read zero. A less than zero reading will indicate the headstock is rotated counterclockwise looking from above and a greater than zero reading will indicate clockwise rotation of the headstock.
Because the cross slide is not moved during this test, any bed twist doesn't come to play. There is a possibility that the cross slide has some irregularity or that the cross slide ways aren't perpendicular to the bed ways. However, aligning the headstock in this manner will permit cutting a flat face with the lathe in any event. My next step is to level the lathe using my precision (.0005"/10") level. Following leveling, I use the RDM method for determining taper, correcting and taper issues using bed twist. The final test is to use the two collar test to verify my alignment/
Turning a test piece is always the final arbiter--that's what lathes are for.
Every fixture plate I've ever bought had runout and needed to be turned square, so I'm not sure I would trust a fixture plate to be square enough out of the box to indicate using a test indicator mounted on the cross slide. The surface of a good chuck is probably better, but you're still making an assumption that it's really dead-nuts right, and the whole point of indicating in a lathe is to avoid assumptions like that.
Here's a photo of a large fixture plate that I had just skimmed to make flat and true. A straightedge across that with a light behind it shows only the slightest light leakage in the middle--my cross slide is just a smidge out of square--a few tenths in about six inches. That is more than good enough for me. The second photo is a recent project where I faced off an 11-1/2" aluminum disk, and it also measured workably flat using a straightedge. Being able to face parts to flat is the whole point of headstock alignment--if the cross slide motion isn't square to the spindle it won't face a part flat. Mine isn't absolutely perfect, but it's close enough (and not adjustable).
Avoiding turning tapers is about tailstock alignment primarily and carriage alignment to a lesser extent. A twist in the bed causes the tailstock alignment to change laterally as you move the tailstock, and that's what causes tapers. A twist in the bed raises and lowers the tool on the carriage as you turn, but that has a much smaller tapering effect because the tool is moving along the tangent to the diameter you are turning. The twist does also rotate the whole carriage so you'll get a bit of lateral motion as well. But the effect of the twist on tailstock alignment is the bigger effect.
But the cross slide motion has to be square with the spindle first, or you won't know what the tailstock is supposed to align to.
Rick "needs facing accuracy more often than cylindrical turning accuracy" Denney
Last edited:
This is great info. Exactly on point. I will follow the order of checks and adjustments you suggest.
Any advice on the recommended diameter to face off for a 13x40? The post I read about facing a plate for headstock alignment said that the test indicator traversing across with the cross slide will read zero from near side to center and will indicate increasing or decreasing from center to far side depending on the direction of misalignment. Had to draw a diagram of a twisted headstock to wrap my head around that one. In essence, you will have a cone type of taper with the side nearest you square with the ways. Draw out a sketch and you'll see what I mean.
Any advice on the recommended diameter to face off for a 13x40? The post I read about facing a plate for headstock alignment said that the test indicator traversing across with the cross slide will read zero from near side to center and will indicate increasing or decreasing from center to far side depending on the direction of misalignment. Had to draw a diagram of a twisted headstock to wrap my head around that one. In essence, you will have a cone type of taper with the side nearest you square with the ways. Draw out a sketch and you'll see what I mean.
Right. Whatever the lathe machines will be square with the tool, because the tool is what makes it so. If the facing operation creates a high center (gaps under a straightedge at the edges), the headstock is pointing the spindle a bit towards the far side of the lathe bed from you. If it creates a dish, the headstock is point the spindle a little towards you. If you measure dish at the center and it's, say, 0.004, then put a dial test indicator on that center spot and adjust the headstock away from you enough to move it 0.002. If a straightedge touching all along the near half is leaving a gap on the far edge of 0.004, then adjust the headstock towards you enough to move a dial test indicator at the far edge back 0.002. Then, make another skim cut and measure again. When whatever gap is under a thou, switch to your tenths-reading indicator (or declare it good), but at some point you won't be able to move the headstock in small enough increments to make an improvement, and you'll overshoot. That will tell you where good enough needs to be on your equipment.
The diameter you machine is up to you, but the bigger it is, the easier it will be to make measurements. My lathe has a 14-1/2" swing, and that piece of aluminum I was facing was 11-1/2" in diameter. I'm not sure being any bigger would have been necessary. Bigger stuff on that lathe fits, but requires playing games with the compound to get the tool against the work when using my Aloris-style toolpost. But I wouldn't have wanted it any smaller. The error I would have measured on a 5" disk would have been half as much, requiring twice the precision to measure and correct.
The test with precision bosses on either side of a precision fixture plate may lead you astray if the plate has runout. But if you take a single boss mounted on the near side of a plate, measure its position with a test indicator on the cross slide, then rotate the chuck to move that single boss around to the other side, and then run the cross slide back to that point and measure again, you'll see true misalignment. But even then I like the facing test better. My lathe is ancient, and even though the cross slide screw is in very good condition (after I found a replacement for the one I destroyed), it's still tighter on the far side of the swing compared to the near side. Wear in the screw and in the cross slide dovetails will always be a bit worse on the near side, because that's where the cross slide spends 98% of its time. A facing test measures stuff where the lathe will be used.
Rick "breathed a massive sigh of relief after the first facing test turned out acceptably good" Denney