What's the best tool to use for squaring cross slide to lathe spindle axis?

marka12161 · Jan 27, 2021

In the not-too-distant future i'm going to need to scrape in the the saddle such that the cross slide ways are perpendicular to the spindle axis. From the vids i've seen on line, a granite square of sufficient size and heft looks like to tool of choice. The process looks to be to align the short leg of the square parallel to spindle axis by moving the saddle along the ways with an indicator along the short leg of the square. Once complete, the long leg of the square will provide the required perpendicular reference surface. From what i see, the heft of the granite square is a benefit as it keeps the square in place with good old gravity.

My questions are:

Is this the preferred method or are there any other methods i should consider?

Does anyone have any experience with import granite squares? Those with the Shars brand name seem to be reasonably priced.

Thanks
Mark

TGTool · Jan 27, 2021

Maybe a technicality, but you want the crosslide perpendicular to the ways. Theoretically that would also be the same as the spindle axis, but other things can enter in. As you describe the setup, you are indeed using the ways for alignment.

If you have a large machine square, metal, not granite, you can also clamp the beam to dowels in the front dovetail of the crosslide and support the blade sticking out over the ways. Then you can check alignment the same way with an indicator on the bed and the same carriage traverse. Just a different arrangement.

ballen · Jan 27, 2021

There is any easy way to check if the cross slide travel is perpendicular to the spindle axis, that does not require a precision square.

Mount an indicator on the spindle and clamp a small gage block to the edge of the cross slide, with the cross-slide at one extreme of travel. Align the gage block so that it's face is parallel to the cross-slide travel. Set the indicator to zero on the gage block. Then run the cross-slide to the other extreme of travel, and rotate the spindle 180 degrees. Touch off again on the gage block. If the cross slide travel is perpendicular to the spindle axis, you'll find zero. Otherwise you'll see how the cross-slide is angled relative to the spindle rotation axis.

eKretz · Jan 28, 2021

That would probably be better done with a round in the dovetail unless you're absolutely certain that the outside of the cross slide is parallel with the ways.

ballen · Jan 28, 2021

eKretz said:
That would probably be better done with a round in the dovetail unless you're absolutely certain that the outside of the cross slide is parallel with the ways.

If you are talking about the method that I suggested, it is independent of having the outside of the cross slide parallel to the ways. You first mount a gage block on the cross slide, and adjust it to be parallel to the motion of the cross slide ways. Then you move the cross slide to the opposite extreme of travel, and use an indicator to see if the position of the gage block has shifted with respect to a plane perpendicular to the spindle axis. This relies on the spindle rotation defining a single axis. If the spindle bearings are shot, it won't work. But most lathe spindle bearings define a rotation axis that is accurate to a few microns at distances of a few inches, because if they had more play or slop than that, they would have terrible chatter problems.

eKretz · Jan 28, 2021

ballen said:
If you are talking about the method that I suggested, it is independent of having the outside of the cross slide parallel to the ways. You first mount a gage block on the cross slide, and adjust it to be parallel to the motion of the cross slide ways. Then you move the cross slide to the opposite extreme of travel, and use an indicator to see if the position of the gage block has shifted with respect to a plane perpendicular to the spindle axis. This relies on the spindle rotation defining a single axis. If the spindle bearings are shot, it won't work. But most lathe spindle bearings define a rotation axis that is accurate to a few microns at distances of a few inches, because if they had more play or slop than that, they would have terrible chatter problems.

Well, to be fair, you edited that last bit in after I posted... Originally the moving of the cross slide was not mentioned.

And yes that would work better than doing it without moving the cross slide but I'd still rather run to the dovetail in case of any lost motion on the cross slide that might give a false reading. And during scraping for instance, (O.P. specifically mentioned doing this) you don't want to reassemble the cross slide to check this every time for testing...

jwearing · Jan 28, 2021

The Connelly book discusses another method.

Mount a parallel in a 4-jaw chuck. Indicate on one end of the parallel and rotate the chuck. Adjust until both ends read the same, then it is exactly perpendicular to the spindle axis. Then mount the indicator on the cross slide and indicate along the length of the parallel.

I haven't done it, I prefer the big square method, but those big squares are expensive.

marka12161 · Jan 28, 2021

jwearing said:
The Connelly book discusses another method.

Mount a parallel in a 4-jaw chuck. Indicate on one end of the parallel and rotate the chuck. Adjust until both ends read the same, then it is exactly perpendicular to the spindle axis. Then mount the indicator on the cross slide and indicate along the length of the parallel.

I haven't done it, I prefer the big square method, but those big squares are expensive.

Like one of my mentors used to say, "all i've got yo do is read". I should have consulted Connelly before posting.

TGTool · Jan 28, 2021

Raises an interesting question. Depending on the lathe design, the headstock on some machines can be aligned (or misaligned) with the ways. In theory the headstock alignment should be checked before using the spindle axis to set the crosslide alignment. On the other hand, it's the spindle axis/crosslide alignment you really care about in ensuring the machine cuts that tiny bit concave.

SO, depending, you could have alignment with the ways that would be functionally improper with the spindle or vice versa. Ignoring the possible problems of machining a shaft parallel, of course. Options, options.

thermite · Jan 28, 2021

TGTool said:
Raises an interesting question. Depending on the lathe design, the headstock on some machines can be aligned (or misaligned) with the ways. In theory the headstock alignment should be checked before using the spindle axis to set the crosslide alignment. On the other hand, it's the spindle axis/crosslide alignment you really care about in ensuring the machine cuts that tiny bit concave.

SO, depending, you could have alignment with the ways that would be functionally improper with the spindle or vice versa. Ignoring the possible problems of machining a shaft parallel, of course. Options, options.

Meahh.. it's only a nuisance when it is time for new rubber-bands, fresh duct tape, and renewal of the bamboo-splint gibs.

The bed is the master. HS & Spindle, TS, carriage ALL must respect the planes & lines it establishes.

Can't have more than one "boss". Try it. Just as much a nuisance as more than one wife or GF.

All you do is waste time making excuses for your performance because THEY cannot agree where you should be, what direction you should be pointing, and what it is you should be doing "along the way"!

Machine tools are much the same.

Richard King · Jan 29, 2021

If you are rebuilding the machine, The best way is to square the cross-slide to the bed. The way the Connelly does it by sweeping off the spindle is in-correct. The spindle bearings could be bad or if the machine had a crash and the head was knocked out of alignment, you will screw it up. Check out minute 3 and then 16 - Keith and Lance are doing it the way I showed them to do it. Monarch Lathe Restoration - Part 15: Scraping the Saddle Square - YouTube

ballen · Jan 29, 2021

eKretz said:
Well, to be fair, you edited that last bit in after I posted... Originally the moving of the cross slide was not mentioned.

That's not correct. I only edited my post to clarify the role of the spindle rotation axis, by changing the wording to "shifted with respect to a plane perpendicular to the spindle axis" and what followed.

eKretz · Jan 29, 2021

That's not how I read it. But the point is made regardless. Maybe I'm beginning to suffer CRS, who knows.

jwearing · Jan 29, 2021

TGTool said:
Raises an interesting question. Depending on the lathe design, the headstock on some machines can be aligned (or misaligned) with the ways. In theory the headstock alignment should be checked before using the spindle axis to set the crosslide alignment. On the other hand, it's the spindle axis/crosslide alignment you really care about in ensuring the machine cuts that tiny bit concave.

SO, depending, you could have alignment with the ways that would be functionally improper with the spindle or vice versa. Ignoring the possible problems of machining a shaft parallel, of course. Options, options.

Richard King said:
The way the Connelly does it by sweeping off the spindle is in-correct. The spindle bearings could be bad or if the machine had a crash and the head was knocked out of alignment, you will screw it up.

My problem with the parallel-in-the-chuck solution is that throughout the whole scraping process the bed ways are the datum point. Everything else is scraped to match the bed ways. Squaring the cross slide to the spindle would deviate from that, it just seems wrong. Even with perfect spindle bearings, errors could accumulate.

To be fair, Connelly first suggests squaring the cross-slide to the bed with a precision square. If I remember correctly, the parallel-in-the-chuck thing is more of a final verification.

jim rozen · Jan 29, 2021

So any truth in the concept that a machine should be set up to face slightly concave?

marka12161 · Jan 29, 2021

Richard King said:
If you are rebuilding the machine, The best way is to square the cross-slide to the bed. The way the Connelly does it by sweeping off the spindle is in-correct. The spindle bearings could be bad or if the machine had a crash and the head was knocked out of alignment, you will screw it up. Check out minute 3 and then 16 - Keith and Lance are doing it the way I showed them to do it. Monarch Lathe Restoration - Part 15: Scraping the Saddle Square - YouTube

Thanks Richard, I'm familiar with that video.

TGTool · Jan 29, 2021

jwearing said:
My problem with the parallel-in-the-chuck solution is that throughout the whole scraping process the bed ways are the datum point. Everything else is scraped to match the bed ways. Squaring the cross slide to the spindle would deviate from that, it just seems wrong. Even with perfect spindle bearings, errors could accumulate.

To be fair, Connelly first suggests squaring the cross-slide to the bed with a precision square. If I remember correctly, the parallel-in-the-chuck thing is more of a final verification.

Yes, exactly the point. The parallel in the chuck is a proxy for the bed ways and introduce several possibilities for error. A little like the principle of doing all the possible operations in one setup. Every time you re-set the part is another opportunity for error to creep in. Why not just go back to first principles, the relationship to the ways? It's not that hard.

The counter argument is that in the real world you always work with imperfections. But you can do good work in spite of them if you know what they are and can compensate. If I know the car speedometer always reads low compared to actual ground speed (larger tires maybe) I can still avoid driving over the speed limit by compensating.

This gets so hypothetical as to be ridiculous, but IF you knew the spindle alignment was off AND you had to make a flat facing cut, AND you were scraping the crosslide to achieve what you needed, you could use the parallel in the chuck regardless of other factors. Using it as a first check, or maybe even a final check as you say Connelly suggests could be perfectly valid in that case.

marka12161 · Jan 29, 2021

At this point, the parallel in the 4 jaw isn't an option since i need to machine a new back plate before i use the chuck.

TGTool · Jan 29, 2021

to the OP

It looks like you've gotten a good rundown on process from several quarters but no one has commented on Shars granite. I've not heard complaints about their granite products other than the usual whining about imports. They're probably more than good enough for what you're doing.

Unless you anticipate using it for other things it might be hard to justify unless you have the money and just like tool collecting as some of us do but may not admit. In practical terms you could probably do it with a combination square if you qualify it beforehand. Take several measurements with slightly different setups so you have confidence in what you're reading. I noted watching the Keith Rucker video that he uses a granite square lying on the ways, adjusts one side to be parallel with carriage travel, then uses the other with a KingWay sliding in the dovetail to check squareness. In setting up the granite he gets down to two or three tenths parallel with the ways and calls that good enough. That's well within tolerance for tooling and setup you can probably arrange in your shop.

marka12161 · Jan 30, 2021

Thanks. Yes, i saw the same video. I'll likely keep an eye open for a decent granite square and may indeed buy an import if nothing domestic at a decent price comes along.

What's the best tool to use for squaring cross slide to lathe spindle axis?

Stainless

Titanium

Diamond

Diamond; Mod Squad

Diamond

Diamond; Mod Squad

Hot Rolled

Stainless

Titanium

Diamond

Diamond

Diamond

Diamond; Mod Squad

Hot Rolled

Diamond

Stainless

Titanium

Stainless

Titanium

Stainless

Similar threads