I scraped a boudler into a master square, here's my report

[mtraven]

Hopefully you didn't toss out the outer offcuts. You're thinking of a three surface setup for flattening the inner square? If you use one or two of the offcuts and flatten the cut face to the same degree as your inner cubic square then those offcuts when used on all the faces of the square will give you want you are after.

nah, the inner square is flat enough at this point. I've just generally wanted to do the 3 pate method, I'd want to make a surface plate that way. I do plan on using some of those cuts to play around with 3 plate stuff, work out some of the basics. Hoping to get a few small prisms out of what left there. But what I really need is a 36" straight edge, so I think my focus is gonna turn to tooling to make that flat...difficult without a big surface plate

If this is an April fools joke, then har har har.
But if you actually made a flat granite square
from a rock in the yard, then Very cool Indeed !
Not time well spent, but maybe you have more
time than money or live in a remote part of the
world. Interesting anyhow.
-D

yes I did actually do it & yes my tooling budget is very limited. Not even sure what a block like that might cost, probably a couple hundred at least? I spent $12 on diamond dust and yes a lot of time. But that was largely time spent learning, I consider that time well spent. I did need the master square, but it was largely an excuse to get into the world of lapping. Being able to precisely shape granite, opens up a whole realm of relatively inexpensive, precision tooling

I just watched your power scraper video.
Very legit.
Subscribed!
-Doozer

thanks, trying to post more videos...I have lots of projects, but I hate editing!

[Paul Alciatore]

I am by no means any kind of expert on making mirrors or lenses. The one telescope mirror, which is still unfinished, is the total of my experience. But I can answer two things here.

First, as far as I know, there is no good way to apply the three plate method to a narrow straight edge. The reason why and the reason why I qualified my statement about it's use with the word "proper" is because the method requires that each of the three must be worked or checked against the other two IN ALL ORIENTATIONS. It is not sufficient to check them with only one axis aligned. It is possible for the straight edges to have a twist which will match both of the other two and when they are turned end for end (180 degrees). Take a strip of paper and grab the two ends and twist it a bit. Then turn it end for end and observe the same twist: they match. At the very least, the plates must be worked or checked at 90 degree increments. A straight edge simply does not have enough width to allow for a real check at 90 degrees. So yes, the two dimensions, length and width, should roughly match. As much as a 1::2 ratio might work, but I would not try much more than that.

As to making lenses for something like an autocollimator, that is very similar to making a telescope objective lens. One makes a point image of a far distant light source (a star) and the other makes a parallel bundle of light rays (which would converge at infinity) from a point light source. The optics are the same, the light just goes in opposite directions. And both uses require the use of a well corrected lens. The minimum correction is that of an achromatic (corrected for two colors of light) lens, which is generally made with two different lenses which have two, different indices of refraction. That is a minimum of four surfaces that must be ground and polished. And they must be properly spaced, which includes the thickness of the finished lenses. So you would think that a telescope objective lens would work for a autocollimator. But have you seen photos of telescopes? They are many feet (meters) in length. This is to allow the use of objective lenses with relatively long focal lengths which are easier to correct for the many possible distortions. In the terminology commonly used in relation to camera lenses, these telescope objectives are "slow" or have a large F number. The F number is the relation of the lens' focal length to the diameter. F = fl / d. In a telescope a large F number is not a big problem. But a slow lens would product an autocollimator that is many feet in length so a "fast" lens or one with a low F number is needed. Lenses with large F numbers are easier to design and produce than those with smaller F numbers. To bend the light in a shorter distance, larger curves are needed in the glass. And those larger curves also produce other optical distortions. This is why camera lenses, where a short overall length is highly desirable, often have four or many more individual elements/lenses in their designs. Fifteen individual lenses are not unknown in fast camera lenses.

So if you want an autocollimator that will fit on your surface plate, you are going to have many individual lenses to make. And each of them must be calculated to have the exact correct curvatures and refraction index. Such designs are challenging, even to the professional lens designers. Computers are heavily used to trace the rays through them. And even if you can find a design that is already complete, then you need to get the necessary glasses and grind every element.

Personally if I were to make an autocollimator, I would look for either a camera lens or a surplus lens from commercial or military optics. I would make the mechanical parts and be satisfied with the surplus optics. The focal length and diameter of the lens will tell you how big your autocollimator will be. Of course, optical paths can be bent with mirrors or prisms. One source of surplus optics:

https://www.surplusshed.com/

A quick check there showed a number of 42mm diameter, achromatic lenses that might make good choices. They were all under $10.

I had been scraping a bunch of steel & CI...had a granite tile on my bench, thought, what the hell, give it a try. To my surprise it felt a lot like cast iron & if it didn't dull so quickly, granite would actually scrap faster than CI (imo). I needed master square, durabar is pricey, rocks are free, the rest is history.

and yes I will do it again, though likely with a heavier use of laps. I expect scraping with take a smaller, but crucial role in future granite projects. I'm gearing up to scrap my lathe, still need a good long straight edge, a prism or two & maybe some measure tools...granite seems like a good candidate for all that.

@Paul Alciatore thats interesting stuff, seems like you know a lot about making lenses...I might need to bounce some ideas off you as I contemplate the viability of an autocollimator build. I had assumed the lens(es) & mirrors would need to be bought (costly, I imagine), but reading your description of the process makes me wonder if I couldn't just rub one out. I don't mean to underestimate the task, I expect it would be a significant challenge.

you touched on "proper" 3 plate technique, is there any such thing for a set of straight edges? or as I suspect, does the technique rely on the plates having a roughly 1:1 length to width ratio?
----------------


no one wants to talk about the pyramids, eh?

[BCRider]

Guy Lautard in one of his volumes of "The Machinist's Bedside Reader" books detailed a procedure for making and lapping a pair of squares to a very high degree of squareness. It involved doubling the errors by making TWO pinned together squares and using a lapping button to lap the edges. A little lapping then offer the squares to each other along a known straight edge against a light table (or handy window with some paper as a diffuser) to see any gaps. The same could be done with two long straight edges. Precision pins with a light press fit?

I also wonder about the idea that if a set of three surfaces can make a set of three flat plates if given enough time then why can't three thin edges do the same? I'm not sure how that would work with granite since even a granite straight edge in practical terms is more a long and skinny surface plate than it is a true straight edge. Or would a "V" edge lapped to a narrow flat qualify?

[mtraven]

paul-- thanks for the reality check on making lenses, I needed that. I was thinking that lens would be hundreds (i have no frame of reference). and its not the only precision optical component in an autocollimator.
thanks for opening my eyes to surplus optics. there's a great surplus store near me that I've been going to for 20+ years, they have a whole optics isle that I haven't ever looked real closely. you got me thinking about where else I could get parts, pretty sure rifle sites use collimation. and dont DLP tv's have some precision mirrors in them? I have a few of those laying around, they might be handy on the sled.


BC-- i've never actually worked through the 3 plate, but I have read many accounts of it--with small variances. but all of them involve exact (usually 90deg) rotation between rubbings, parallel is fine, but perpendicular is gonna destroy both surfaces if the L:W is way out of whack. That being said, I -think- it should be possible to prove 3 edges against each other, I just don't think that can be the technique for actually removing material.

[BCRider]

I recall Tom Lipton doing a set of plates and showing the movements. So I did a YT search with the keywords "flat three plate method" and got Tom's video along with some others. I didn't watch the others but I suspect they'll deal with the proper methods for how the plate are worked and how often they move around.

For flat edges such as lapping and proving the edges the method there was to pin the parts together so they are locked and then to set them slightly raised and use a brass or aluminium cylinder as the lapping tool. Rotating it so it wears evenly while lightly touching up the edges of the steel straight edges or squares. This is the last part of course. The initial work could be done using the mill... And with squares given a bit of a kick if your mill is like mine and is out of square by the amount I found it to be....

[boslab]

I am envious of your new pet rock, it is the straightest rock I’ve seen, congratulations on an amazing bit of stonemasonary, I’m sorry to say I straighten bits of rock with a bolster and 3lb club hammer, so tenths are literally 1/10”! Well done ( you could make a really accurate pyramid)
mark

[JoeLee]

I applaud your efforts but wouldn't a piece of granite counter top be a more suitable material ? I had thought of making a 45 deg. and a 30,60,90 deg. square from one. The thickness was about right. Cutting it would be no problem but making it true would be. I'll be waiting on your results.

Anyway............ Fred would be proud of you.



JL.................

[mtraven]

I am envious of your new pet rock, it is the straightest rock I’ve seen, congratulations on an amazing bit of stonemasonary, I’m sorry to say I straighten bits of rock with a bolster and 3lb club hammer, so tenths are literally 1/10”! Well done ( you could make a really accurate pyramid)
mark

thanks! I actually started with a 10# sledge--there was a really obvious fault running through the rock & it need to come off---plus I REALLY didn't want it to break late in the process, figured slapping it around with a sledge might help show any defects. Plus its great stress relief (me & the rock)

I applaud your efforts but wouldn't a piece of granite counter top be a more suitable material ? I had thought of making a 45 deg. and a 30,60,90 deg. square from one. The thickness was about right. Cutting it would be no problem but making it true would be. I'll be waiting on your results.
JL.................

were you thinking it'd be better because of the type of granite? or because is already somewhat flat and more or less parallel?

if your referring to the master square, its ~3 x2.5 x5.5", so counter top granite wouldn't have been thick enough. Plus I had that blank roughed out to within about 80 thou after just a few hours of cutting, it was 30 hours to finish it. so starting with nice blank, would be nice, but getting to that point is the easy part.

That being said, part of the reason I took this on was, I have a nice granite drop from my sink that I intend to make a few edges like you described. This little block was a good learning experience(and proof of concept) gearing up for others. Though for the long edge I want to make, I'll be taking ~ 6" rip & using it on edge, more like a camel back edge. I dont really have need for a 45(i dont think?) but a 30-60-90 triangle or prism is certainly in my future. Does present some challenges in measurement that I not yet sure out to handle, like testing to make sure its exactly 60. Probably need a sine bar & gauge blocks to check that properly, I have neither. I think the 30-60 prism would be much easier, as you have 2 parallel edges.

[JoeLee]

thanks! I actually started with a 10# sledge--there was a really obvious fault running through the rock & it need to come off---plus I REALLY didn't want it to break late in the process, figured slapping it around with a sledge might help show any defects. Plus its great stress relief (me & the rock)




were you thinking it'd be better because of the type of granite? or because is already somewhat flat and more or less parallel?

if your referring to the master square, its ~3 x2.5 x5.5", so counter top granite wouldn't have been thick enough. Plus I had that blank roughed out to within about 80 thou after just a few hours of cutting, it was 30 hours to finish it. so starting with nice blank, would be nice, but getting to that point is the easy part.

That being said, part of the reason I took this on was, I have a nice granite drop from my sink that I intend to make a few edges like you described. This little block was a good learning experience(and proof of concept) gearing up for others. Though for the long edge I want to make, I'll be taking ~ 6" rip & using it on edge, more like a camel back edge. I dont really have need for a 45(i dont think?) but a 30-60-90 triangle or prism is certainly in my future. Does present some challenges in measurement that I not yet sure out to handle, like testing to make sure its exactly 60. Probably need a sine bar & gauge blocks to check that properly, I have neither. I think the 30-60 prism would be much easier, as you have 2 parallel edges.

I was thinking that granite would be a better choice because of the quartz content, how ever much that would be in a black granite counter top. I think that a rock is just too soft and more fragile than granite.
I found a piece of counter top, it's about 1" thick or so, polished on one side. I would think that would be thick enough for a small square or triangle.

JL............

[BCRider]

But...... granite IS a rock ! ! ! To be fair it does depend on what sort of rock mtraven stated with. So MT, do you know the type of rock that you have there?

A friend of mine is into stone counter tops but not that low class granite stuff. She's keen on full on quartz for a bunch of reasons. I wonder how quartz would be as a plate, square or straight edge?

[mixdenny]

Darn. I had a beautiful pink granite rock that split in half during construction. I kept it on display for years. I finally tossed it in the woods next to my house, I wonder if I could find it.

[mtraven]

as BC said, granite is granite. As far as I can tell, the rock I picked was a pretty hard granite with a very fine structure to it. I don't have the ability to look at the makeup of the rock, quartz content for example, so thats all I can tell you. I read a little about stone selection in doing this project, turns out not even all granite tables are actually granite, often they are gabbro.

I found a piece of counter top, it's about 1" thick or so, polished on one side. I would think that would be thick enough for a small square or triangle.
JL............

perhaps, really depends on how long you want to make it.


re: quartz

quartz countertops are actually quartzite--an engineered (man made) stone. 90% quartz, 10% epoxy. the epoxy would worry me for stability, unless a special low/no thermal E epoxy is used..which I doubt. The reason higher quartz content is desirable is, its hard & good for durability. But as with most hard things, strength is sacrificed. **generally speaking** pink granite has more quartz than black. Because of that, pink/white surface plates are thicker to make up for that difference. So with quartzite, you might run into problems like that, circumvented by a thicker table if ya like ...but there is still the epoxy that might change over time, not to mention may create problems when lapping being so much softer than the quartz.

and I'll just mention, the durability concern is much more a problem for job shops, using their plates on a daily basis. as a home shop guy, I really doubt that block will wear in my lifetime. and if it does, I'll just re lap it.



that being said, I have worked with a bit of basalt in the past. It seems like a good candidate, very hard, very fine grain structure...yet I've never seen a surface plate made from it (its got a characteristic green hue to it). I wonder why that is? Maybe its grain structure is too small do to its rapid cooling (basalt is a volcanic rock, granite is igneous). Everything I read says basalt is softer than granite, though my diamond saw says differently.


edit: from the wiki on gabbro: "Slow-cooling, coarse-grained gabbro is chemically equivalent to rapid-cooling, fine-grained basalt"


so I think it is, as I suspected, a matter of too small a grain structure.

[rmcphearson]

My hat is off to you sir for completing this project!

[Jammer Six]

After watching and reading about this, a stupid question: when you start, you have to have a reference surface that is both larger than your largest surface and flatter than your intended standard, are these correct?

[Paul Alciatore]

There are no stupid questions. And sometimes when a very basic question is asked again, a new understanding can be the result.

For the vast majority of the methods discussed here and elsewhere, YES that is true.

BUT those larger and flatter surfaces must be produced somehow. And that "somehow" is the three plate method; well it is at least one of the "somehows". The three plate method does not require any additional or preexisting reference surfaces. It generates surfaces to whatever level of accuracy the maker wishes. And it also provides that craftsman with a pretty good means of testing just how good those three surfaces are.

Now there are companies who make things like surface plates and straight edges and granite squares. And those companies will have their own methods for producing and measuring those surfaces and edges. But historically, in the end, it all goes back to the three plate method which was for centuries the basic method for generating a flat surface.

Flatness can also be measured by optical methods. One optical instrument used is the autocollimator. But that is a whole 'nother subject.

After watching and reading about this, a stupid question: when you start, you have to have a reference surface that is both larger than your largest surface and flatter than your intended standard, are these correct?