Prepping stock for joinery - How do you know when the stock is straight/flat enough?

[David Fegen]

When prepping stock for joinery, how flat & straight do you make your pieces - when do you know you can stop?



I know the answer is 'as flat/straight as possible', however, this comes at a cost - every iteration of planing and then stopping to check takes time.


With experience, this is probably based on intuition knowing what works and what doesn't work. However, being a beginner (1-2 years) I don't have enough experience with stock-prep & joinery to have built up any intuition.


My general approach for faces & edges:
1) jointer (or jack) plane until consistent full-length/width shavings
2) straightedge on top, turn to feel how straightedge moves on the workpiece (i.e. checking for humps/valleys)
3) tilt Veritas aluminum straightedge on edge, look for light-gaps


The problem is, there are always light-gaps. But from what I understand, it's nearly impossible to stop all light from passing through, especially when tilting a straightedge on its edge (I hear machinists have 'light rooms' where they measure photons).


It seems like you can achieve 1) & 2) within a range of being straight/flat - as evidenced by there always being some variety/size of light-gaps.


I just make sure the light gaps don't look 'too big.' I don't have a quantitative measurement for it, I just subjectively compare to previous workpieces I've prepped which may be correct or incorrect (this is where I'm missing the intuition from experience).

------------------


This seems to be a more general discussion on the topic, but not specific to prepping stock for joinery: https://sawmillcreek.org/showthread.php?269772-Standards-for-board-flatness

-----------------



Maybe I'm being too analytical about this...


When does everyone stop - do people have certain telltale signs on when enough is enough?

[Maurice Mcmurry]

I only use "light proof" gap ticks on work three feet long or less.

[Scott Winners]

You will know how good your stock prep was by how well your joinery comes out. Layout and execution also matter of course. Once you have a thing assembled, look to see what errors there are, then figure out what caused them and do that step better next time.

For dovetailing, I need to do a better job on endgrain prep next time. My most recent is OK, but it looks like the errors I see in the finished cabinet are from having sub optimal end grain prep. I think it is easier to layout dovetails when the board is s6s, but end grain prep is harder. I have moved away from trying to layout dovetails on stock that is s4s. The temptation is to only deal with the end grain once after the cabinet is glued up, but layout is tricky.

[glenn bradley]

As mentioned, a perfect fit at one end and a problem at the other means "not true enough". This will vary with your requirement but for a stretcher that will join at both ends, dead flat and true is what I do. For a leg that will fasten at one end a little, little bit of twist can be tolerated.

[Jim Koepke]

Hi David and a belated welcome to the Creek.

For prepping stock for joinery matching a straight edge isn't necessary. It needs to match the piece to which it will be joined.

Maybe an old post on making straight edges/winding sticks will help > https://sawmillcreek.org/showthread.php?290331

jtk

[George Yetka]

Ive been working with letting material sit in garage as long as possible to acclimate. Once I have waited a couple days I will mill to close to final size. Then Ill stack it and clamp to bench or weight it and finish flattening the next day. That hidden stress is a killer

[Rafael Herrera]

It depends on what you are making. In my case, this is a shelf for growing lights. The raw material was 2x12 lumber. Some of it twisted as I ripped it, the final dimensions were approx. 1 1/4" x 3". the exact size didn't really matter I aimed for not too bulky, but not too thin, half lap joinery, glue, 1/4" dowels and a few screws. Not all the pieces were perfectly straight, but they were long enough that on assembly they flexed enough to fit. Leave some material for final smoothing so you can get rid of marking lines, dings and smudges. For smaller pieces, more precision may be needed, but you're in control, you need to figure out where precision matters and where it doesn't.
473551

[Richard Coers]

Your entire piece of furniture is only as good as the stock preparation. If you don't want quality in your work, then stop early. I build furniture that my family will be proud to own when I'm gone. I do not want them to think I cut corners because making a piece of stock right might have taken another hour, 50 years from now.

[Luke Dupont]

We're woodworkers, not machinists.

I think in the modern age we have an obsession with precision, far and beyond what is necessary. Many people preach precision at all costs -- an approach that no true craftsman working with time and money constraints would endorse. Precision beyond what is necessary or pleasing is wasted time and effort and adds nothing to the finished product.

For most purposes, a little light showing through here in there is perfectly fine. Just make sure there are no bulges in the middle on which the straight edge would rock, and there are no large gaps in which you could, say, slide a piece of paper in or something. That's already plenty accurate, maybe even more than necessary, for the vast majority of work, excluding maybe tool making.

For fun, you might occasionally try making a box or something with quite imperfect stock, unevenly thicknessed and everything. Then square up only, say, one face and one edge, and work with as few square faces as possible. This will make you think about where you actually need precision, and to what level you actually need it, and how you can achieve beautiful and accurate results in the absence of perfectly dimensioned stock. Another thing you can do is see how quickly and efficiently you can produce passable work, as would have been the goal of many craftsmen in the old days. I suggest these as exercises only, so you can learn from experience where the line is between what's a waste of time, what's quite good enough, and what's lacking or problematic.

I began thinking about this after talking to some professional joiners who still make their living this way here in Japan, and also after seeing people like Frank Klausz cut quick and dirty dovetails in just a few minutes.

Again, more precision for the sole purpose of more precision is utterly pointless. If it's not helping you to create something more beautiful or functional or adding satisfaction, it's excessive. The key is finding that line, which will no doubt change depending on the project and application.

[Maurice Mcmurry]

After the Jointer, Plane, or table saw with a lapped partner board. I sand with a known to be flat, square, and straight sanding block and or shooting board. I lap my sanding blocks and shooting boards as if I was truing the ways on a machine and use them to sand high spots until no light comes through. Mostly on things like stair treads and smaller items. I have a big, thick glass table with lights underneath. It is not as flat as a granite surface plate but works well.

[James Pallas]

Please don’t start trying to get a machinists tolerance into woodworking. If you do you will do a lot of tail chasing. If you want to find out how good you are doing measure it. Flatten an edge put a straight edge on it and try feeler gauges to check the fit. It’s very easy to see .0001 with light. If you really want to find out put your board on a surface plate and measure with an indicator. Then go take your straight edge and your feeler gauges to the best furniture store in town and measure some of that. If you can get wood to .005 to .010 your golden. Remember it won’t be the same the next morning. As is said YMMV.
Jim

[Jim Koepke]

Please don’t start trying to get a machinists tolerance into woodworking.

My woodworking improved quite a lot when the tape measure was put away and pieces were cut by comparison and story stick marking.

If two pieces need to be the same length, use a shooting board.

The same width is easy with a plane.

Some of this is covered in the straight edge & winding stick post linked to in my previous post.

jtk

[Prashun Patel]

When flattening a face, I don't check for light. I rock a straight edge and at various points. Rock-free is good enough for me - on the face. It's time consuming to check for light in several positions on a large piece, and it's unnecessary; rock free will mean even the glossiest finished piece will appear flat enough to the eye.

When jointing an edge, I check for light with a trusted straight edge. But only to get me within shooting range. The real proof is light clamping the 2 mating edges together with thin pieces of paper at various points. If none of the pieces move under light clamping pressure, the edge will glue up perfectly. I glue up edges one board at a time this way. It means many tedious trips from the glue-up table to the bench to make corrections, but it's the only way to know for sure. In fact, on most longer panels, I aim for a slight spring anyway.

[steven c newman]

Face: My hands can tell how flat a surface is, and where any high/low spots are. edges....jointer plane, check with steel yard stick...not so much for light, but for it rocking on any high spots...if it is low in the middle, yard stick will wiggle. Once the first edge is close....then I build up from there...goal being that the 5 board panel glue-up can stand up, on its own, will out clamps or glue....and no gaps showing. Then I can glue the panel up...


BTW: I do have a set of "Feeler" Gauges....and they stay in the tool chest with the rest of the old Ignition Tools. When was the last time they set the points in a distributor, anyway....

[Jim Koepke]

When was the last time they set the points in a distributor, anyway....

??????

Didn't everyone use a matchbook cover for that?

jtk

[Tom Trees]

It depends on the work, whether going for the authentic old style era ala Follansbee,
or whether one has reason for doing accurate stuff.

If the latter, then I have yet to see an easier speedier method for stock prep than the flat bench approach,
with timber not pinched between anything and flipped often.
This approach includes a long reach angle poise lamp which has a 7.5 or 8" shade to shine under the work to check for gaps.

Trying to fight flat takes extra time, and the planes won't work as well as they need to in most cases, i.e reversing grain,knots, figure etc,
and all that is just making extra steps for oneself.

Tom

[Jason Buresh]

It all depends on the desired end result.

Are you looking to have something featured in fine woodworking? Are they exposed or hidden surfaces? Are you recreating and antique piece of country furniture?

Each one will have a different answer.

Your satisfaction with the project should be the main factor of how critical you are though. But don't let perfection get in the way of enjoyment or completion of the project.

[Jack Dover]

3) tilt Veritas aluminum straightedge on edge, look for light-gaps


The problem is, there are always light-gaps.

Just FYI, tilting is less accurate. If I understand it correctly, you're tilting a straightedge so that it is resting on a "rib" where two faces meet. These straightedges do not guarantee straightness of a "rib" (technically an "edge"), only the straightness of the reference "edge" (technically a "face", just narrow). If a side is not precision lapped, it could be slightly wavy and then the "edge" will also be slightly wavy, while the reference edge (a "face") could still be perfectly flat and straight. That's why there's also "knife edge" or "beveled edge" straightedges - these are guaranteed to have a straight narrow reference face, ground almost to a line. Not that it's important for flattening wood though.

You will always be seeing light when using a straight edge on wood, except maybe on dense stable woods with tight grain - tropicals, etc. Fibers are all different, some compress more than surrounding, some spring back, some not, some swell from humidity more and so on. That doesn't matter either, you're looking for "hills and valleys" and whether a board is cupped or bowed. If you see some undulations on a face (because an iron is cambered, etc) - they most probably don't matter too, unless too severe. Like, a board from a planer is all wavy and rippled, yet it's perfectly workable and can be truly flattened right before finishing. Ikea doesn't even bother with removing planer marks, and people are pretty happy with it.

If anything, getting a board into cuboid shape is way more important for joinery than flatness.

[Charles Guest]

A component has to be out of twist. Some other conditions, if mild, are tolerable depending on where the piece will go in the project.

[Jim Koepke]

Ikea doesn't even bother with removing planer marks, and people are pretty happy with it.

Happy people may even pay more for that "rustic look."

"Less is more." - Gov. Jerry Brown

He just didn't realize he was talking about the labor vs cost of Ikea products.

jtk

[Scott Clausen]

As some have stated you will need a solid (flat) reference edge and eliminate twist. For the rest you can feel and see the major deviations. Use strait edges if you don't trust your eyes or to double check. I find if I get to focused on perfection I start creating other problems, kind of like oversteering on a slick road.

[Matthew Hills]

Chasing flat/square also trades off against workpiece thickness, which can be a concern.

Your stock will make a big difference in how hard it is to get flat/square/smooth.

Learning to judge the effect of any imperfections is part of the learning process. I got much more interested in milling after trying to build a few things out of hardwoods from home depot.

Matt

[Maurice Mcmurry]

I have a big, thick glass table with lights underneath.

I am getting some things out of moth balls today. Here is an image of the light table and a 2 foot shooting board.
My surface plate is an old saw top. It is not truly flat but close enough for most woodworking.

474444 474495

[Luke Dupont]

Chasing flat/square also trades off against workpiece thickness, which can be a concern.

Your stock will make a big difference in how hard it is to get flat/square/smooth.

Learning to judge the effect of any imperfections is part of the learning process. I got much more interested in milling after trying to build a few things out of hardwoods from home depot.

Matt


Great point. This is always a consideration when I'm deciding how flat to get something. Ideally I look to preserve as much thickness as possible, and sometimes I'm working to a target thickness that I don't want to go under.

[John C Cox]

All I needed to do is look at old, hand made furniture... what I saw... Often as not, the jointed show faces are done very well. Jointed not-for-show casework is made strong/stable, but not pretty, and non-show stuff living in "nowhere land" gets almost no prep.

Yes, every single scrap of wood precisely dimensioned and trued is nice. It's a necessity in high volume industrial production where things must be completely interchangeable and rely on simple unskilled assembly via jigs and fixtures. Recreating factory stuff can be maddening, because all the parts are so ridiculously accurate... Ironically, they excel at making giant piles of chairs with identically loose joints that wobble loose after an hour or two of actual use, and cabinets with identically cranky drawers that wrack every time you pull them.

I would say, figure out what needs to be accurate and put your time into that, and then don't sweat the rest so much. It will help keep your energy and creativity intact

[steven c newman]

2 biggest tools I use....Palm of my hand ( because everyone else will be running THEIR hand over it) and the MK II Eyeballs.....IF it looks square to me, then it will look square to anyone else..

Actual squares come up when I start the cutting process....and during assembly.

Remember....we are dealing with a material than can change as the humidity changes....and will never stop changing... What was perfectly sized when you left the shop..MIGHT be either a hair bigger, or smaller the next time you go to the shop....Just wood being..wood.

[William Fretwell]

My workbench is on a tilt over it’s length of 0.02 degrees, trying to not let it bother me. It used to be flat, the concrete floor is still flat but wood moves. I will check again in the summer, my humidity goes from 22% to 95%.

Planing wood one day releases tension that moves it by the next day. Your precision drawer made in the winter won’t budge in the summer.

You can flatten it until there is nothing left to flatten. The mortice and tenon joints and overall structure do keep things straight but don’t wait three days to glue it up!

If you are happy with the result, it’s straight. For drawers you need to be very happy. The finish after the final shaving is more important.

[John C Cox]

That's a good point too - older, proven designs had a lot of commonality in certain features because wood moves a lot. Consider that before ~1960, air conditioning and seasonal indoor humidity control was nonexistent. Furniture moved a lot seasonally... The designs incorporated sufficient allowance for this movement, and often part of the design concealed this extra clearance.

So if your designs rely on ultra-precision to work; say inletted drawer faces with invisible gaps, they probably won't even open or close in spring and fall.

[Anuj Prateek]

Been down that rabbit hole. Eats too much time and distracts from actual project. I have still not stopped doing it completely but it's much controlled.

If you like numbers. Stick with something like 0.015" - 0.020" over 3'. That's 0.005" - 0.007" per feet. It may be a lot for some and too strict for others. Point is pick a reasonable number based on results you are currently getting without spending excessive time on flattening.

Edit:

Don't tilt straight edge to see light. Only bottom part is guaranteed to be flat. Imagine if straight edge has a bow length wise then you will always see light.

[Richard Hutchings]

"3) tilt Veritas aluminum straightedge on edge, look for light-gaps"

I don't know where you got that idea but don't. Straight edges are not designed for this and will most likely bow a little as you do this. If it's bowed, there will always be light. This was mentioned above but I felt like adding my 2 cents :-)

After looking at those beautiful Veritas Straightedges, I may have to retract this. They indeed look pretty stiff but I personally wouldn't trust it when leaning it.