Whitworth-three plate method - Page 3

1. Stainless
Join Date
Apr 2013
Location
Damascus, MD
Posts
1,391
Post Thanks / Like
Likes (Given)
3980
790
I'm not very experienced with lapping, I mostly do scraping. Anyhow, I guess that the issue is that the plate on the top (the one that is being moved) is subjected to pressures and parts of it overhang from the edges of the bottom plate at the end of each motion: the bottom plate could grow low at the edges, while the top one could grow high at the edges.

This is all I could think about.

Paolo

2. Diamond
Join Date
Sep 2002
Location
Lawrenceville GA USA
Posts
6,175
Post Thanks / Like
Likes (Given)
697
1315
If you watch the first part of this video you might get a bit closer to understanding a different perspective.

Charles

3. Plastic
Join Date
Apr 2019
Country
AUSTRALIA
Posts
15
Post Thanks / Like
Likes (Given)
10
1
Originally Posted by Paolo_MD
I guess that the issue is that the plate on the top (the one that is being moved) is subjected to pressures ....

Paolo
Newton's third law tells us that to every action there is an equal and opposite reaction. In other words, any pressure (force) from the top plate onto the bottom plate exerts an identical pressure from the bottom plate onto the top one. That is why I can't see any need to alternate the top and bottom plates.

4. Plastic
Join Date
Apr 2019
Country
AUSTRALIA
Posts
15
Post Thanks / Like
Likes (Given)
10
1
Originally Posted by CBlair
If you watch the first part of this video you might get a bit closer to understanding a different perspective.

Charles
Charles, I watched the video and I think I can understand how he can end up with three identical but not-flat surfaces doing what he did, which was rub the plates without changing their orientation. My understanding is that part of the method involves rotating the plates around their 'Y' axes while doing the rubbing, to randomise the wear.

He rubbed them without rotating them but then rotated them to compare them. I'd like to see the experiment repeated with rotating them while rubbing them.

Ken

5. Diamond
Join Date
Sep 2002
Location
Lawrenceville GA USA
Posts
6,175
Post Thanks / Like
Likes (Given)
697
1315
He also mentions why that is a good method for round or square parts but it is not practical to rotate rectangular parts as you cant get even loading or spotting of rectangular objects. This is not exactly what you were asking about but it is relevant to the overall concept and it is something to keep in mind. It is also something that is not talked about anywhere else that I know of.

Charles

6. Stainless
Join Date
Apr 2013
Location
Damascus, MD
Posts
1,391
Post Thanks / Like
Likes (Given)
3980
790
Originally Posted by kenlip
Newton's third law tells us that to every action there is an equal and opposite reaction. In other words, any pressure (force) from the top plate onto the bottom plate exerts an identical pressure from the bottom plate onto the top one. That is why I can't see any need to alternate the top and bottom plates.
You're perfectly correct in citing Newton's third law. But, unfortunately, you're incorrect in its interpretation, since the lapping requires you to move the plates relative to each-other, creating some overhang, and you've to keep into consideration Newton's universal gravitation law as well.

As sketched here below, the distribution of forces on the surfaces is not even when the top plate overhangs the edge of the bottom one.

Paolo

7. Diamond
Join Date
Jun 2006
Location
Near Seattle
Posts
4,661
Post Thanks / Like
Likes (Given)
2898
1232
I take it that kenlip understands why the pieces have to be symmetric (sqauare, round) and one needs three of them.

The question about "what do you have to swap A from bottom to top" is addressed by Paolo, and by emperical observation. A related issue is that some mechanism, often one's hands, is moving the top plate, while the bottom is typically held much more rigidly.

As for why do A and B need to trade places but C not, that strikes me as wrong - all three plates should take turns on top and on bottom.

8. Plastic
Join Date
Apr 2019
Country
AUSTRALIA
Posts
15
Post Thanks / Like
Likes (Given)
10
1
Check out this video, from about 17 minutes onwards (the earlier part is just about how they measure the flatness).

There is no way these guys are flipping the blocks. Also, the blocks are not symmetrical. They aren't even the same size.

I'm confused.

Ken

9. Diamond
Join Date
Jun 2006
Location
Near Seattle
Posts
4,661
Post Thanks / Like
Likes (Given)
2898
1232
that is NOT the three-plate method - the three-plate method is a scheme to generate flat reference surfaces *without* recourse to instruments like the collimator he sites down, or even the use of the repeatometer, let alone things like interferometers or interference band measurements (no helium light sources in Whiteworth's time.) three-plate method is a way to make "very flat" from "nothing" (well, suitable material, some kind of abrasive or scraping process, and some thin consistent bluing material)

The three plate method dates to at least the 1830s and is normally attributed to Whitworth. Interferometers date from the late 19th century (Michelson of Michelson-Morely fame). Things like autocollimators date from 16th century or so, but I couldn't easily find a scheme for measuring flatness with one older than sometime in the 20th century.

10. Plastic
Join Date
Apr 2019
Country
AUSTRALIA
Posts
15
Post Thanks / Like
Likes (Given)
10
1
bryan_machine, Perhaps I misunderstood, but I didn't get the impression they were using any of the instruments to actually flatten the plates. They used them simply to measure and verify the flatness of the plates.

Once they measured the plates, and decided that their flatness could be improved, they then used lapping techniques to flatten them. How the lapping techniques they used relate to the lapping techniques of the 3-plate method is what I am questioning.

Granted, they had a (presumably) flat steel plate, which they used to lap the granite. Whitworth's system seems to require a specific sequence of alternating which plates are lapping against each other, symmetrical plates and alternating top and bottom plates yet they lapped only two plates of different size and shape and got flatness.

11. Stainless
Join Date
Apr 2013
Location
Damascus, MD
Posts
1,391
Post Thanks / Like
Likes (Given)
3980
790
Kenlip,
As Bryan_Machine clearly explained, the two things are absolutely different.
Let's try to better clarify the very obvious, first: in the case of lapping a granite plate (referring to the video), you have very flat cast iron laps charged with different grit of diamond powder. The target, i.e. the granite plate is not a lap by itself, since the diamond (or whatever abrasive you want to use) does not embed into its surface. Therefore, the lapping action of the granite plate toward the lap is negligible.

The other thing to pay attention is that, when things are done by professionals with a lot of experience, they appear overly trivial. One thing that I want you to note is their comment that they prefer to lap larger plates, than smaller one. they mention that, with small plate (or, in any case, small in reference to the lap used), there is more risk of lapping the plate convex, for the same reasons discussed here above, when trying to create three perfectly flat surfaces from nothing.
However, the main reason why he has placed the surface plate on the top of the roughing lap is that the lap is larger (and heavier) than the plate and, by moving the surface plate always within the boundaries of the lap, without nay risk of introducing any convexity/concavity.

Lastly, once more, lapping a surface with already flat laps has nothing to do with the Whitworth's method. Tom Lipton has actually a series of videos about creating three circular laps using the Whitworth's method using non-embedding abrasive.

Paolo

12. Diamond
Join Date
Jun 2006
Location
Near Seattle
Posts
4,661
Post Thanks / Like
Likes (Given)
2898
1232
@kenlip - i get your question now. Paolo_MD made a good explanation, but I'm going to beat on this horse some more since it's the kind of distinction that can be very hard to grok.

In the surface plate refurb video by Lipton, various tools ALREADY KNOWN TO BE VERY FLAT are at hand, and instruments based on properties of physics RATHER THAN AN ARTIFACT are at hand. (Collimators, etc.)

You don't have to use the Whitworth method to make a flat object if, for example, you have a device (like a collimator, interferometer, etc) that will tell you its shape. The process might be a bit non-determinstic, but you could eventually work your way to flat.

The Whitworth method is interesting because it allows derivation of a flat surface when external references are not at hand.

In a different video Lipton demonstrates 3-plate work, and if memory serves eventually uses an optical flat with helium line to check the plates (though he took one of those videos down for some reason.) That technology was unknown in Whitworth's time.

And the need to check flatness without a reference to start from (key point), and the geometry, is why it's a 3-plate method rather than 2-plate....

(1830s - no interferometers yet, no lasers, so far as I can find no autocollimators at least for this application. How do you make the first surface plate? The three plate method....)

13. Diamond
Join Date
Feb 2004
Location
peekskill, NY
Posts
24,183
Post Thanks / Like
Likes (Given)
0
4391
Originally Posted by bryan_machine
As for why do A and B need to trade places but C not, that strikes me as wrong - all three plates should take turns on top and on bottom.
Also, each plate needs to turned 90 degrees every other spotting, or a saddle-shaped error develops. The absolute
best discussion of this that I've ever seen is in Moore's book.

14. Plastic
Join Date
Apr 2019
Country
AUSTRALIA
Posts
15
Post Thanks / Like
Likes (Given)
10
1
bryan_machine and Paolo, your last couple of postings have helped a lot to clarify some of the technicalities. As for the concept of 3-plate method, I think I understand the idea of the iterative lapping of the plates with the various sequences and rotations.

However, there are still some points that I can't work out. For example, in The Whitworth Three Plates Method — Eric Weinhoffer , in step #6, he says, "Finally, the single convex plate is lapped against the Blue control plate"

I can understand how, if one was utilising scraping, one could selectively work on the convex plate, with the flatter plate as the reference/control, to get the convex plate to match the reference plate. However, I can't understand why, if lapping the plates, the convexity of the upper plate doesn't introduce a concavity (or multiple waves) into the lower plate.

The plates are all made from the same material, so the concept of the abrasive imbedding into the softer cast iron plate and the imbedded particles abrading the harder granite wouldn't apply here. Hence my thinking that both plates will be abraded, leading to a loss of flatness in the lower plate in step #6.

After step #5, there are two matching and relatively flat plates. One could surely stop here and use either of these plates, if the flatness was within one's desired spec. However, to get improved flatness, one needs to repeat the whole process, perhaps multiple times, so one needs to get the third plate lapped. Is there any other reason to go to step #6? For the final round of the sequence, would one not want to stop at step #5?

15. Plastic
Join Date
Apr 2019
Country
AUSTRALIA
Posts
15
Post Thanks / Like
Likes (Given)
10
1
Originally Posted by jim rozen
The absolute
best discussion of this that I've ever seen is in Moore's book.
I have ordered Foundations of Mechanical Accuracy by Wayne R Moore, 1970. Is that the book to which you are referring?

Ken

16. Diamond
Join Date
Feb 2004
Location
peekskill, NY
Posts
24,183
Post Thanks / Like
Likes (Given)
0
4391
Originally Posted by kenlip
I have ordered Foundations of Mechanical Accuracy by Wayne R Moore, 1970. Is that the book to which you are referring?
Ken
That's the one I was mentioning, yes.

17. rhb
Plastic
Join Date
Apr 2019
Country
UNITED STATES
State/Province
Arkansas
Posts
49
Post Thanks / Like
Likes (Given)
5
20
Wayne Moore's book is a masterpiece. Prepare to be amazed.

A precision level vial (e.g. 4"/2 mm) will measure flatness to 20 millionths per inch without the need for 3 plates.

A plate ground plane and parallel to match the dimensions of your precision machinist's level with 3 carbide ball supports, slightly flattened on a grinder will allow measuring the flatness of a plate without the need for generating 3 plates.

One hundred and eighty years ago Whitworth introduced the 3 plate method. While still valid and useful, gravity is an excellent reference of very high accuracy and reliability.

A level supported at 3 points defines a plane and its relationship to vertical. So a plate which supports a precision level via 3 points provides a reliable means of defining a plane surface over short distances.

The level measurements give you the derivative of the surface. If you integrate that you have the surface.

If you sense the level vial by means of a capacitance bridge, resolution of a few millionths per inch is easily achieved. Suitable vials from AliExpress are under \$12 delivered from China. In lots of 100, 4 mm carbide balls are \$0.50 each on eBay.

The primary limitation on measuring flatness is temperature stability.

18. Here you go... They were upright when I opened them...flipped now...

19. A few more.. Plus 3 points on these plates are laid out withe the points at 25% but not all 3 points are created equal. 3 points under a Sip Jig Bore or say an old Pratt & Whitney Jig Bore, the 2 points are placed under the heaviest sides or under the columns and the single side is under the lighter end. Blanchard Grinders and Heald ID grinders 3 points are located near the extreme outside ends. This can start a whole new conversation, but an easy way to explain it is they built the machines 3 point system to make them self aligning and easy to level. A rebuilder is working for the most part on some sort of rectangle shape and we place a triangle under the part be it a saddle, table and we place the 2 point side under the heavy end and 1 point under the light side. I say using 3 points is also a "lost Art" and the more you practice using it the more perfect you can be.

20. Plastic
Join Date
Apr 2019
Country
AUSTRALIA
Posts
15
Post Thanks / Like
Likes (Given)
10
1
Bryan_machine and Paolo, you last couple of postings have helped a lot to clarify some of the technicalities. As for the concept of 3-plate method, I think I understand the idea of the iterative lapping of the plates with the various sequences and rotations.

However, there are still some points that I can't work out. For example, in The Whitworth Three Plates Method — Eric Weinhoffer , in step #6, he says, "Finally, the single convex plate is lapped against the Blue control plate"

I can understand how, if one was utilising scraping, one could selectively work on the convex plate, with the flatter plate as the reference, to get the convex plate to match the reference plate. However, I can't understand why, if lapping the plates, the convexity of the upper plate doesn't introduce a concavity or a series of waves into the lower plate.

Both plates are made of the same material, so one shouldn't get the same wear pattern one does when lapping a granite block with am iron plate, where the harder material is the one that gets worn down by the abrasive. Both plates should undergo change.