View Full Version : Scraping lathe compound/cross slide plan

06-21-2009, 03:24 PM
Just looking for a sanity check here, to see how good or bad my plan is. Feel free to blast me if anything isn't quite right here, I'm an amateur just kind of feeling out the idea.

I've got a 93 year old 13" Southbend whose compound and cross slide are loose in the middle and tight at the ends as you'd expect from wear. I'd like to someday clean up the dovetails by scraping them in against a cast iron reference that I'd have to make. For the cross-slide, this is more convenience (I don't tend to face large diameters,) but for the compound it's more important so I can get a decent taper. As it currently is, it's so loose and sloppy that the sideways force from turning the compound handle shows up in the finish. So it might just be that I only want to do the compound now, but maybe do the cross slide if I'm confident enough.

However, I do not want to scrape the ways on which both of those ride, because I would not feel comfortable doing that. Does it make any sense to scrape either compound or cross slide without first scraping the ways? I can't be sure there is no twist remaining in the bed, and I feel like I'd just screw it up worse. Is this a warning flag for asking about scraping the other parts?

For the dovetail reference, I'm planning on getting some durabar cast iron, and milling then scraping it flat and true. One side of it has the dovetail angle, and the other side is 90 degrees vertical, so I can check that the cross slide axis is perpendicular to the spindle axis by fitting the reference in the left-hand dovetail, and checking that fits flat/parallel to the faceplate.

For milling the dovetail reference, I'm planning on drilling holes and countersinking to hold the part on the table. Here's what I'm thinking, three setups after drilling:
1) Light clean up facing cut to flatten rough casting for good clamp down. This will become the bottom of the dovetail reference.
2) Flip over so flat side is on table, face-cut the top flat, cut 90 degree side (for non-dovetail faceplate comparison side) almost to bottom, so I don't need flat riser stock under it. Then angle head and cut dovetail angle side, similarly almost all the way to the bottom of the iron. This should give decent geometry between dovetail-angle side and right-angle side that will be used to compare against faceplate.
3) Retram head, flip part. Face-cut through what remains on the other side from cutting only almost all the way to the bottom.

Then scrape all the sides flat. This sounds tricky because there are three precision sides: flat bottom, dovetail angle side, and 90 degree faceplate checking side. I can check that the faceplate side is parallel to dovetail angle side, and perpendicular to bottom with a DTI held in the height gauge.

The lathe's dovetail is about 1" wide, just over 1/2" tall, and (on cross slide) about 14" long. Would the cast iron reference need to be much larger than this, such as doubling the height or width to 2" just to give it more stiffness on the 14" dimension? I was looking at 1.5" square durabar, which is listed to finish at 1.320", does that seem wide/tall enough to be stiff enough?
Also, should the reference be more than say an inch longer than the dovetails?

As for the scraping, I've made my own scraper out of a steel flat clamps a piece of carbide, which I've sharpened on my gemstone faceting machine (wow, beautiful mirror finish with precise angle and radius, way better than I can do by hand!) This will work great for scraping the reference, but I guess I need something different to get into the dovetail corners. What kind of scraper blade is used for dovetails, any big differences, or something just a little thinner to get in there?

Are there any tips or tricks for maintaining the proper geometry while scraping the lathe dovetails? I'm hoping I can clean it up without moving it too much, such as making it slope up or down, etc. I guess I'd need to measure the width with some precision ground rod to make sure the dovetails are parallel. On the cross slide I'll use the faceplate to make sure it's still perpendicular to the spindle axis. Aside from those ideas I'm not sure what else is required or can be done to make sure the scraping is done in the right directions to keep/correct the geometry.

So whaddaya think, am I missing anything important or have any bad idea here? How would YOU approach this?


P.S. No, I don't have any of the popular machine tool reconditioning books that I know someone is likely to suggest.

06-21-2009, 05:41 PM
Start at post #89 for what's involved.

Greg Q
06-22-2009, 07:01 AM
Why wouldn't you at least get a copy of MTR?

For dovetails it was suggested to me to use a standard carbide tip and grind it to a 60 degree bevel instead of the 85 degrees normally used. The same lapping procedures apply. You want a tip thin enough to get right into the dovetail. You'll also need some kind of scoring tool to open up the relief groove at the apex of the dovetail to allow your spotting tool full access.

I have made a couple of spotting tools from dura bar, using a bandsaw, files and scraper. This was not fun but was a good review of some hand tool skills I didn't want to lose. You might find it easier to dedicate one tool to one angle. ie-one dovetail angle, one for right angles.

It can get complex enough trying to keep two surfaces true to each other let alone three. Any wear or warping will be much easier corrected in two planes than three.

I will again make the point: buy a copy of MTR. It has all the answers you need, and more.

Good luck!