When I was in my teens (70 years ago), I ground and polished a 6 inch telescope mirror to make a Newtonian telescope. The final polishing step used celenium oxide and a pitch lap. The results were accurate to within fractions of the wavelength of light. I don't see why we can't do the same thing to flatten the back of a tool. Has anyone thought of something like this? Obviously, you would only do it for something very special because it would take hours to do. I thought of it after seeing an ad for a Blue Spruce chisel, where they boast of such flatness.
First, I’m guessing you were making a round parabolic mirror for your telescope not the central diagonal flat mirror. If this is the case, you weren’t trying to make a surface flat, you were trying to make a correct optical figure for the parabolic mirror. The pitch surface with the cerium oxide polishing powder wasn’t the exact opposite of the parabolic mirror, it was crowned in a different way, right? You moved the mirror blank in a prescribed pattern over the lap to do the polishing. Then you would test to see if the figure was correct, then modify the lap or your pattern to home in on the proper parabolic figure. Then test again and go back to polishing, over and over until the figure of the parabolic mirror was correct for your telescope design. It could take hours, days even, right?
The other thing that comes to my mind is why would anyone need to do this? Yes, Blue Spruce is apparently set up to do such flattening and honing and my guess is that it took them a good bit of time to achieve this with reproducible results. They probably have some pretty expensive machinery to do this too, I’d guess. But why does the average woodworker need to do this? Might be interesting to try, but I see no real advantage. The basic set of Japanese stones, diamond plates and or oil stones in good fettle should be able to get a woodworking chisel or plane blade flat enough to work well. I don’t know how one would easily determine if a surface was flat to within angstroms and in my mind, it isn’t needed for woodworking.
It is better to have our sharpening procedures tied to observable results, than abstract notions. A worker's ability to discern subtle differences in performance (in sharpness, geometry and edge life) are key to developing sharpening skills.
Your reply shows that you understand the process very well. The initial shape HAS to be a portion of a sphere. The polishing, in an extremely subtle way that I never really understood, moves the sphere to a parabola, which as you say is continuously monitored, using the Focault test, to the correct parabolic shape.It's based on diffraction physics. If one goes too far, it becomes a hyperbola, which is not what you want. But there is no reason, with appropriate adjustments,why one couldn't make a flat.
Of course, I understand that just about no one in his right mind would regularly do such a thing for normal woodworking. But wouldn't it be a blast to try it just once!
Incidentally, all the stuff to do it is incredibly cheap in cost. The Focoult test, for example, uses a juice can with a pinhole and a little bulb. Also, it occurs to me that one could rig up some sort of little motorand some levers,etc. to do much of the polishing.
I agree with you, of course. But I still find it interesting that Blue Spruce must do something like this, granted they would have a sophisticated set up. There's nothing abstract about it, about it though. This is one way to actually make optical quality lenses and mirrors. And, as a teenager, with essentially no funds at all, I did it and made a really fine mirror. It would work for a chisel, if you were nuts enough to take the time. I can assure you, if you did it, that you would never have to re -flatten the face of such a chisel-- ever!
I have Blue Spruce chisels. I also have Veritas chisels, and Veritas plane blades. All are flat, flat, flat ...
But flatness, per se, is not the target. It is being coplanar along the length and width. Like Japanese chisels. Or an undercut dovetail floor. However, if you are machining, and you can get the accuracy, flat is the same as coplanar.
A few more comments: thanks for catching, ‘cerium’. All these years I had selenium in my head. The pattern you mention is remarkably simple; one simply moves the top disk back and forth over the bottom disk while continually, moving around in a circular way. Or, what I did, which was to mount the disk on a fixture that could be rotated. You mentally think of just “sanding” the lower disk FLAT with the upper disk. What happens, though, is that the lower disk becomes the convex portion of a sphere, and the upper disk becomes its matching concave partner. So the upper disk becomes the mirror,, and the lower disk should be thought of as the grinding tool. The method results in curved surfaces — not flat — because the disks are the same size and the grinding process continually has the upper disk OVERLAPPIBG the lower disk so that the area of contact changes. And changing area of contact results in changing pressure. And a critical mathematical point is that the only two surfaces in perfect contact—IN ALL DIRECTIONS— are a sphere( and a flat surface( which is a sphere of infinite radius.) IF THE LOWER SURFACE WERE SUFFICIENTLY LARGE SO THAT THERE WERE NO OVERLAP, then the contact area and pressure would not vary, and the resulting surfaces would have to be flat.
. It would work for a chisel, if you were nuts enough to take the time. I can assure you, if you did it, that you would never have to re -flatten the face of such a chisel-- ever!
We abrade the back every time we sharpen. The first time the chisel is sharpened the fantasy is over.
What am I missing? My point is that you would not ever have to sharpen the back of the chisel. Sharpening the bevel shouldn't affect the back-- unless you mean that little bit to remove the burr. As for fantasy, Blue Spruce tools are not fantasy. I hope you don't think that I am actually advocating this approach for normal woodworking. I just thought it was an interesting idea.
My understanding is that you will always be fighting for a flat back since the dulling does not only happen on one side. Since the back edge gets dull too you need to further flatten the back to ensure precise cutting action. I primarily use Japanese Chisels and enjoy the hollow backs.
Every time you hone the bevel, you turn a burr onto the back side. If you don’t then rub the back on your sharpening stone, how do you get rid of the burr?
As I mentioned earlier, we abrade the back at every sharpening. That will destroy your optical flatness. Since we only abrade the inch or two near the edge, the back will slowly become slightly convex, even though the part near the edge remains flat. Take a look at a chisel that has been sharpened thousands of times.
I thought it was rosen that was used, and the glass swarf(?) was the abrasive.
I had a nerdy friend with a couple of glass discs in his pocket for a while. At school. To my non-physics brain, a four-inch diameter mirror was small potatoes next to one twice the diameter. I would have joined him but found sharpening knives and whittling was more entertaining.
How do you, or any typical amateur woodworker, know that these chisels are optically flat? Who among us has the tools and skills to measure that degree of flatness? I'm not disparaging Dave Jeske, but we should keep in mind that he sold the company to the marketing geniuses at Woodpeckers. For a reality check, take a look at a photomicrograph of a flattened surface of your favorite hardwood at the level of magnification you are discussing. A bale of hay would appear smoother and flatter, so what's the poiint?
To test for optical flatness microscopes and micrometers are not sufficient. You would need light diffraction methods. Some variant of the Foucault method should work. To test for the correct parabolic shape when polishing a telescope lens all that’s required is a juice can, a flashlight bulb, and a razor blade.
There are two separate aspects of an optically flat surface: the smoothness, and the planarity. While I'm not convinced either is helpful in a chisel, if I recall correctly from when I ground my telescope mirror (more recently than you-- it was only fifty-some years ago) to get a spherical surface you rub two things together, but to make it flat you need three. Each must be ground and checked against the other two, and only when all three match can you call anything flat. You need transparency to look for the color bands that say you are getting close to matching on a scale of wavelengths of light. You might be able to use a similar approach with two glass and one metal surface.
The optical flattening technique really is in testing, and in making inevitable biases cancel one another. The later stages are analogous to hand planing a board flat: check where it's out, fix that, and recheck. For chisels, grinding on a flat surface while trying not to rock the chisel sounds good enough. If you make a mirror polish and don't see distortion at the edges of the reflection, it's flat enough that I don't think it would be possible to distinguish from an optically flat chisel in use. But both wear and bending in use (not even sharpening the back) would likely take it quickly out of optical flatness.
The color bands that have been referred to here for shaping spherical or parabolic mirrors and lenses are called Newton's Rings, just in case anyone wants to study the techniques more deeply.
I'm doing it on an India stone, which is pretty close grit to a soft Arkansas, mostly because it's hard, resists dishing and generally produces less dubbing.
But on some of my old chisels previous owners polished only first 1/4" or maybe less, now they're quite out of flat and I can't think about anything else except maybe sacrificing the first 1/4". And it seems my newer chisels are going that way too, it's just I re-flatten them every once in a while. Was wondering maybe I should use a finer grit?
The flat only needs to be long enough for you to reliably find it by rocking the chisel while trying to extend the length of an already flat wood surface or while removing a burr.
