Thread: Where to get flat reference surface

Originally Posted by rob streeper

What I've found here (Plane sole flattening, again...) is that even seemingly insignificant deviations from flatness make a difference if you're objective is a flat finished surface.

Compare the data presented in posts 49 and 96 of the above linked thread. It's readily apparent that the plane lapped to 3 microns produces a smoother finish over the tested range of cutting depths and that it will take heavier cuts with consistency. It now leaves a flatter surface than my 6" jointer with a Shelix head does. The plane is now lapped to 1 micron but I haven't had time to do more testing.

Sure, it makes for very interesting testing and discussion topic, but just because super flatness can be produced doesn't necessarily mean it's useful to for practical wood workers and especially given the main audience on these forums of DIY wood workers.
I can understand these small differences are needed for accurate fitting in metal work, but the way wood moves, even on a daily basis, and that other larger scale dimensions play key roles in WW, does it really matter, especially for DIY WW.
The question them becomes what overall difference does super flat wood really make to any finished wood product and what level of flatness will produce acceptable (not invisible) results especially for DIY.
An interesting question to ask would be, professional cabinet makers in the last few hundred years seemed to a have made perfectly acceptable if not outstanding furniture without any tools to this level of flatness - what level of flatness did they work with?
It all reminds me of screen / camera resolution - where the eye can only ultimately see differences of ~0.1mm, then one might ask apart from making camera companies more $, what is the point of going any further?

Originally Posted by BobL

Sure, it makes for very interesting testing and discussion topic, but just because super flatness can be produced doesn't necessarily mean it's useful to for practical wood workers and especially given the main audience on these forums of DIY wood workers.
I can understand these small differences are needed for accurate fitting in metal work, but the way wood moves, even on a daily basis, and that other larger scale dimensions play key roles in WW, does it really matter, especially for DIY WW.
The question them becomes what overall difference does super flat wood really make to any finished wood product and what level of flatness will produce acceptable (not invisible) results especially for DIY.
An interesting question to ask would be, professional cabinet makers in the last few hundred years seemed to a have made perfectly acceptable if not outstanding furniture without any tools to this level of flatness - what level of flatness did they work with?
It all reminds me of screen / camera resolution - where the eye can only ultimately see differences of ~0.1mm, then one might ask apart from making camera companies more $, what is the point of going any further?

Why? Because it's there.

P.S. Yesterday I had occasion to use the plane studied in the other thread in the field and it's a much sweeter tool in the hand now.

Originally Posted by rob streeper

I went around the house just now with my Lapmaster flatness gauge. I took several points on three different granite countertops, a glass table and a mirror, All showed variations in flatness of 0.0002" or more, some were as high as 0.0005". The flattest was one bathroom counter with only a couple of spots that were > 0.0002" out of flat. None were flat enough to produce an optimally flat plane bottom.

Just clarifying - over what distances were the above flats determined?

Originally Posted by rob streeper

You can make your own flats by using three pieces of countertop or thick mild steel plates and some diamond abrasive but given the amount of work involved a small granite machinist plate is a bargain. Start with something like 60 micron grit. Label your plates A, B and C. First lap A and B together until they make full contact. Then lap B and C together. Finally lap A and C together. Then proceed to the next finer grit. When you get down to 1 micron grit (several days later) you'll have three flat surfaces. This procedure is used for truing lapping plates and it works very well.

Is there a detailed guide on how to do this?

Originally Posted by yoboseyo

Is there a detailed guide on how to do this?

The Whitworth Three Plates Method — Eric Weinhoffer

It depends on how flat the surfaces are to start with but by way of a guide, 3 , 12 x 12" CI plates machined flat to start with take about 4 days to get flat.
I've not done it by apparently it's murder on you hands, arms and shoulders.
The only really practical was to do this is with an oscillating arm that can move the top piece automatically over the piece underneath and use running water to wash away the debris.

If you want to get a piece of CI flat, and assuming it has already been machined flat and your a have a suitable reference surface against which to compare, it's faster to perform what is known as a "hand scrape".
This won't get it as flat as the 3 plates method but it is considered flat enough for metal work machinery so it should be more than flat enough for woodwork.
Below shows a picture hand scraping - its done with a special TC blade in a holder - it's murder on your hands/arms and shoulders.
At a scraping weekend class I attend I attended in Melbourne in 2014 we stated with a dozen or so blokes.
One guy lasted an hour and 2 others didn't turn up on day two.
The aim of the class was to scrape this block to within 0.02 of mm AND square to 0.02 of a mm across all 6 surfaces.
Getting a surface flat is relatively easy - getting it square is harder.
It took me all of saturday and I finished my block on Sunday at about 3pm.
SF3.jpg


The reference surface in this case was this massive granite block is smeared with a very thin layer of blue grease and the side that is being scraped is rubbed against it.
Marco2.jpg

Here's an example of a surface approaching the 0.02mm flatness
The evenness of the patter is what identifies the degree of flatness.
FOTD.jpg

Here's what my block looks like now after it has been blued - its the small one in between the other two.

Originally Posted by BobL

Just clarifying - over what distances were the above flats determined?

The gauge has four points of contact. The distance between the end points is ~238 mm. The indicator is in the center.

Lapmaster flatness gauge contact side.jpg

Originally Posted by rob streeper

The gauge has four points of contact. The distance between the end points is ~238 mm. The indicator is in the center.

Thanks.
Sorry should have asked this in my previous post.
Given that flatness measurements usually are given in distance per unit distance. do your flatness measurements quoted for stone benches and float glass take the distance between the outer contact points into account between distance into account?

No, just the reading on the indicator gauge. I did look at some of the surfaces with my smaller flatness gauge and there was very little variation noted. This is the smaller gauge, the base is about 84 mm long and the indicator contact is in the middle.

PW flatness gauge.JPG

Here are a series of measurements of our flattest countertop.


Countertop 10.jpgCountertop 9.jpgCountertop 8.jpgCountertop 7.jpgCountertop 6.jpgCountertop 4.jpgCountertop 3.JPGCountertop 2.JPG


Here are some measurements of a JET tools 9" X 12" surface plate that I bought a few years back for $59US shipped.

Jet surface plate 2.JPGJet surface plate 5.JPGJet surface plate 7.JPGJet surface plate 1.jpgJet surface plate 4.JPGJet surface plate 8.jpgJet surface plate 3.JPGJet surface plate 6.JPG


And here is the gauge, the reference block and the JET plate on the countertop measured above.

flatness gauge, reference block and Jet surface plate.JPG

The countertop is concave in the middle and is ~10X less flat (?) than the JET plate.

I looked up the flatness of surface plates, there are two criterial - local and overall flatness
Based on the size of your gauge, for a 238 x 238 mm plate;

A Grade AA plate of that size should be better than 43 µ" (micro inches), a grade A better than 86 µ", and a grade B 172 µ"

Your measurements for some bench tops and glass of 200-500 µ" are therefore only about 1.2 - 3 worse than a grade B surface plate?

Either I have messed up the arithmetic, or that would suggest that your granite tops and float glass would be more than flat enough, especially for the average DIY wood worker.

My measurement of "un-flatness" of 4000 to 7000 µ" for a piece of ~12 x 10" float glass may perhaps be more indicative of its flatness ?

If I recall correctly the JET plate is a grade B and my measurements appear to be consistent with your calculations, perhaps a bit better and getting close to grade A for this particular plate.

P.S., how are you putting special characters in your posts?

Originally Posted by rob streeper

If I recall correctly the JET plate is a grade B and my measurements appear to be consistent with your calculations, perhaps a bit better and getting close to grade A for this particular plate.

Thanks.
So, are you still suggesting that your float glass and granite tops are not flat enough for use by an average DIY woodworker?

P.S., how are you putting special characters in your posts?

I'm using a Mac so it's "option m" for µ, for a PC I believe it's "alt 230"
https://theasciicode.com.ar/extended...-code-230.html

Originally Posted by BobL

Thanks.
So, are you still suggesting that your float glass and granite tops are not flat enough for use by an average DIY woodworker?



I'm using a Mac so it's "option m" for µ, for a PC I believe it's "alt 230"
https://theasciicode.com.ar/extended...-code-230.html

I haven't tested it of course and am reluctant to say yea or nay without some testing, all I stated above was the fact that countertops, glass and tile aren't flat.

I have found (detailed in the other thread) that my test plane needs to be really, and surprisingly, flat, lapped to 6 µ's or better, to take a consistent 0.004" thick shaving. Further, the test plane was flat to within 0.0002" or so after the first stage of lapping and its performance was pretty bad (see the first post of that thread).

Good enough for "DIY woodworking"? That's a personal decision, my home remodeling work has shown me that some people are well satisfied with things that don't immediately collapse, leak or start a fire but most aren't.

alt 230 = µ. Thanks, I didn't know that worked online.

Originally Posted by rob streeper

Good enough for "DIY woodworking"? That's a personal decision, my home remodeling work has shown me that some people are well satisfied with things that don't immediately collapse, leak or start a fire but most aren't.

Note I did say "Average woodworker" and not just any WWer

There will always be a wide spectrum, but my guess is the WWer in between "satisfied with things that don't immediately collapse, leak or start a fire" , and persons such as yourself, would be more than satisfied with something that approaches the flatness of a B grade surface plate. Whether they would be satisfied with my piece of float glass is something else. Perhaps it can be said that in between what you and I have observed, a woodworker won't know the flatness of a piece of float glass unless it is checked out?

We also should bear in mind that super flatness only gets a woodworker so far. And although it's a very important first step it's pointless getting a piece of wood super flat if you can't then get it square or parallel.
Sometimes we forget that WWers of the pre 21st century didn't have any of this fancy gear and still managed to turn out very impressive pieces of furniture.

Despite this I will continue to read and enjoy your stuff on flatness and thanks for posting it - I guess that also shows perhaps where I sit on the spectrum