Standards for board flatness

[Tyler Bancroft]

I have a relatively poor eye for assessing the flatness of a board, so I've started using a straightedge and feeler gauges when I'm uncertain. What's a reasonable standard for the flatness of a board? In other words, if I place a straightedge in a variety of positions on the face of a board and can't slip a .025" feeler gauge underneath it anywhere, would you consider that sufficient? .015"? I realize this likely differs depending on the intended use of the piece.

[Warren Mickley]

Welcome to the forum, Tyler.

Instead of using the straightedge after you flatten the board, you want to use it as you flatten the board to guide your work. Use the straightedge and winding sticks to determine what are the high spots and then plane only those high spots. Don't take full shavings when there are already low spots on the board.

Although I have been planing boards by hand for almost fifty years I have never used a feeler gauge with a straight edge. If the straightedge rocks, take out the hump in the middle, if there is light under the straightedge plane the two ends and avoid the middle. I tried a feeler gauge this evening; I had trouble finding a place to put .001" underneath. I think it is faster and easier to achieve precision if you are constantly checking.

[Tyler Bancroft]

Welcome to the forum, Tyler.

Instead of using the straightedge after you flatten the board, you want to use it as you flatten the board to guide your work. Use the straightedge and winding sticks to determine what are the high spots and then plane only those high spots. Don't take full shavings when there are already low spots on the board.

Although I have been planing boards by hand for almost fifty years I have never used a feeler gauge with a straight edge. If the straightedge rocks, take out the hump in the middle, if there is light under the straightedge plane the two ends and avoid the middle. I tried a feeler gauge this evening; I had trouble finding a place to put .001" underneath. I think it is faster and easier to achieve precision if you are constantly checking.

Thanks, Warren. I frequently check with a straightedge (or edge of a jack plane) while working the board, but I find I can see light under the edge even when a .015" feeler wouldn't slip underneath. I can get a board to that state, but I'm not certain just how flat "flat" is.

[Tyler Bancroft]

To give a bit of context: I'm presently flattening some 12" wide maple boards. I noticed last night that I could see light under a straightedge easily even though a .015" feeler wouldn't slide under it - a bit of experimentation tonight found that it's possible for even my mediocre eyesight to see light under a .005" hollow. I can work down to .001", if necessary, but is that a good rule of thumb? Essentially, I'm trying to find out what peoples' standards are for flatness so I can train my eye to detect that.

[Warren Mickley]

Well, there are certainly places where we use different standards, and experience can help here. For instance the face side might need greater precision than the back side. For preparing a face side I usually have my trying plane taking shavings in the .002 range and would generally take a partial shaving if I thought it would improve flatness. For preparing edges to be glued, I use a jointer with .001 shavings. I don't want to see any light under the straight edge or when I place one board on the other. And if one board is in the vise and you set the other on top, you can feel even better than you can see if there is wind or hollow or whatever in the joint.

Last year I made some crown moulding and fascia for an 18th century house. For one component the face side just needed to be planed smooth for painting. the back side could be left as it came from the sawmill as in the original. The edge that was moulded needed to be quite straight for looks and for making the joint, but the other edge was hidden and did not have to be worked at all. In fact the width just needed to be within an inch.

On the other hand, for a carcass side in furniture, careful preparation pays dividends all the way down the line in joinery, assembly and finishing

[Tom M King]

It depends entirely on what you are going to use the board for. 15 thousandths is the thickness of a matchbook cover-what used to be used to set auto points in ignition systems, before electronic ignitions. So that much would be noticeable on about anything. You wouldn't want that much of a hump/hollow on the top of a piece of furniture, but it wouldn't matter on a picnic table. I've never seen any kind of "standard" for flatness by measurement.

[Charles Guest]

I have a relatively poor eye for assessing the flatness of a board, so I've started using a straightedge and feeler gauges when I'm uncertain. What's a reasonable standard for the flatness of a board? In other words, if I place a straightedge in a variety of positions on the face of a board and can't slip a .025" feeler gauge underneath it anywhere, would you consider that sufficient? .015"? I realize this likely differs depending on the intended use of the piece.

A hump in the middle of a board, edge, or panel is almost always bad, a slight but more or less regular concavity usually not.

[Tyler Bancroft]

Thanks to all who replied. Past a certain point, is flatness purely an aesthetic issue? In other words, as long as joints are fitting properly and spilled coffee isn't draining off the edge of the table, should I be worried about getting flat to .001" instead of .010", if it's not visibly different?

[Warren Mickley]

Thanks to all who replied. Past a certain point, is flatness purely an aesthetic issue? In other words, as long as joints are fitting properly and spilled coffee isn't draining off the edge of the table, should I be worried about getting flat to .001" instead of .010", if it's not visibly different?

It is always nicer to have well prepared stock. Here are some problems that arise when things are sloppy.

Thicknessing. We flatten the face side and then use a marking gauge to scribe the thickness. We plane to the line. If the line is not all in the same plane, it is difficult to plane to the line.

Wind. If you are making a frame and panel door and there is wind in a stile, that wind will be magnified when you go to put the frame together. Adjusting tenons or whatever, just to make the frame lie flat is work.

Carcass. When you are gluing a carcass (I use hot hide glue) it is nice if everything fits together nicely If a carcass side is slightly hollow in width, it is difficult to even mark dovetails on it, let alone try and straighten it when gluing together. And when you are gluing that last corner together, you want it to fall into place like it belonged there, not to have to pull on one piece and push on another just to get things together.

Jointing. If there is a slight hollow or slight twist, take care of it. If this is a lot of work, work at it until it becomes trivial. This will pay dividends.

When you get to that point when things are within the .015 of flat you mention, most of the work is done. So it is not a lot of time to make it better.

[Pat Barry]

Thanks to all who replied. Past a certain point, is flatness purely an aesthetic issue? In other words, as long as joints are fitting properly and spilled coffee isn't draining off the edge of the table, should I be worried about getting flat to .001" instead of .010", if it's not visibly different?

My opinion is that flatness depends on the surface and planned finish. If its a tabletop, for example, and you are going to go eith a high gloss finish, then you better have a very flat surface. If its going to be a matte finish, then absolute flatness isnt as important. I don't have a number for how flat is flat. If you are working to a number then, to be clear, you need to talk in terms of flatness per distance, such as 10 mils per foot, because the basis (denominator) is equally as important as the numerator.

[Kory Cassel]

My opinion is that flatness depends on the surface and planned finish. If its a tabletop, for example, and you are going to go eith a high gloss finish, then you better have a very flat surface. If its going to be a matte finish, then absolute flatness isnt as important. I don't have a number for how flat is flat. If you are working to a number then, to be clear, you need to talk in terms of flatness per distance, such as 10 mils per foot, because the basis (denominator) is equally as important as the numerator.

+1 Very subtle dips or waves will show up clearly in a high gloss finish and not matter at all for other applications. Of course then it just means you have to put so many coats on that it vastly increases the amount of time necessary to achieve the unbroken gloss compared to getting a truly flat base to work from. Each new layer of finish increases the required drying time compared to the previous layer you see. I've never measured so precisely, maybe I should as I usually end up using much more finish than I would like to have done. If you get to the bottom of it, please let us know.

[Prashun Patel]

IMHO your light test is a better thing to focus on than mils. For edge joints if light pressure eliminates the light or visible gap, then I think it’s fine.

For board faces, flat is often only important for aesthetics. In this case, relative flatness over a foot or so can be seen, but wider than this not so much. So for this, I trust my smoothing plane to tell me when the surface is flat enough.

For testing square I rock a square along the corner. If it does not feel like it’s rocking I think it’s good.

My method is to use good square and straight edge references and then trust my eye and fingers to judge fidelity to those trusted references.

[Robert Hazelwood]

0.015" is a lot, really. Especially on the smaller components that comprise most of fine furniture making. It should be relatively easy to get in the 0.005" range without too much fuss, just taking off high spots and then taking full-length shavings. Then with a bit more effort you can get it to 0.001", or to where there is no light visible under the straightedge.

Flatness is pretty important aesthetically, though aesthetics don't require the same level of precision as when joining boards. The gold standard here is the ability to take full length shavings, unbroken from one end to the other, and with overlapping strokes across the width, all with unbroken shavings. The board may not be perfectly flat at that point (it will often have a slight convexity), but it will be very close and can be guaranteed to not have the kind of troublesome dips that show up badly when a finish is applied.

Take Warren's advice about checking frequently as you go. It helps you understand how the plane and board surface are interacting, so that you can learn how to adjust your technique to suit the board at hand. The goal is to make a flat surface while taking the minimum number of shavings, so you need to figure out the topography of the board and develop a strategy for how to correct the defects.

[Tyler Bancroft]

Thanks to all who have replied so far - for someone learning solo, this site is fantastic.

Some checking found that the cutoff for no longer seeing light under my straightedge is somewhere around .003-.005" - does that seem normal?

[Jim Koepke]

I have a relatively poor eye for assessing the flatness of a board, so I've started using a straightedge and feeler gauges when I'm uncertain. What's a reasonable standard for the flatness of a board? In other words, if I place a straightedge in a variety of positions on the face of a board and can't slip a .025" feeler gauge underneath it anywhere, would you consider that sufficient? .015"? I realize this likely differs depending on the intended use of the piece.

Kind of like what others have said or implied, the flatness depends on the requirements of where the board fits into a piece and what kind of piece it is.

Today a piece of mine only required it to be flat to the bench. In other words if it didn't rock on the bench it was flat enough. It just needed the high corners knocked down.

A table top would be different and require a different approach.

jtk