Three-surface Method For Using Dovetails Against Themselves?

Riaan

Registered
Registered
Joined
Oct 8, 2015
Messages
71
Gentlemen,

I've been reading the theory behind 3-surface truing and leveling an I think I understand how it works, but something occurred to me and I'd appreciate some feedback on my reasoning.

Let's assume you're starting out, you've bought a cheap, second hand import mill with no warranty and you don't own any flat reference planes like a granite plate or straightedge. Let's also assume the dovetails were ground in by dragging the parts behind a dump truck.

Consider the Y axis. It makes sens to me that you'd be able to slide the bed out, turn it through 180 degrees and slide it back on.

And I think you already know where I'm going with this. Flipping the bed gives you your three surfaces, and with some careful bookkeeping and umming and aahing would this work for identifying uneven spots for some scraper love?

If this wouldn't work, could you please explain why?

Thank you!
Riaan
 
Two of your surfaces are never being compared to each other. As a result there are symmetrical curves and slopes that won't be detected. The three-surface process requires that every surface be tested against every other one.

It can be a useful test, though, and it's one I used in fixing up my Grizzly table.
 
Ah, I see, you need to reference them round robin style. Thanks for the explanation!
 
PS can you please share what you did on your grizzly table and what you discovered and how you fixed it? Thanks!

This bleedin' opera mini
browser isn't allowing me to edit. Aargh
 
if you are gonna scrape, i'd suggest getting a surface plate so that you may make your own reference
 
PS can you please share what you did on your grizzly table and what you discovered and how you fixed it? Thanks!

I'll leave out the blunders and dead-ends.

The table and the base bottom were, amazingly, flat as near as I could measure and parallel when the table was centered. However the angle between those surfaces was a complex function of position and I coud rock and wiggle the table. If I tightened the gibs enough to stop the wiggling at the worst spots they were too tight to permit full travel. All my millwork had a .1" pattern in the finish from the table rocking as I turned the handles.

First I scraped in the Y-slide casting using my glass-plate "surface plate", using a DTI to make sure that the top and bottom remained parallel. This was quite straightforward. Then I scraped in the horizontal surfaces on the table and base, using the DTI to make sure that they remained parallel to the table surface and the bottom of the base respectively. For this I used a straightedge, the DTI, and the scraped-in Y casting. I took off a lot more metal doing this than I would were I to do it again. Getting rid of all the rocking wasn't easy, and I assembled and disassembled the whole thing far too many times for testing. I then tried to scrape in the angled surfaces inside the dovetails and got rid of the biggest bumps but ended up lapping them which left me with a too-smooth surface: I should have flaked them. Note that it does not suffice to get these flat: they must also be parallel to their mates and their line of intersection must be parallel to the horizontal surfaces. I also made new steel-backed brass gibs.

I'm sure an experienced scraper would consider it an ugly job, but I got rid of all the wiggling, rocking, and binding and I can now tram the table and have it hold within .001" from end to end.

I now have a copy of Machine Tool Reconditioning so next time I can do it (sort of) right.
 
Sounds like a success, good job. Thanks for the feedback!
 
On many of the China machines and on badly worn machines, even a poor scraping job can make a large improvement.
When a quality indicator which registers one ten thousanths of an inch increments will no longer show a variation in a surface, the people who are good at scraping
know there's still a ways to go...
Still seems a bit like magic to me.
 
Yeah I'm still surprised that in today's age of high tech trickery, a human hand and eye still does the best job of making straight lines.

And they say there's no straight lines in nature. Rubbish! :)
 
Back
Top