How straight is straight?

[davidtu]

For those following my ongoing saga of breaking in myself to my new tools (jointer, planer, even the saw!) I've been struggling to end up with nice flat straight edges. It occurs to me, tho, that I don't have a good metric to judge it by... how straight is straight... or straight enough?

In this case, I am working on a small box (gotta start somewhere!) and I suppose the answer might vary depending on the project (does it?!) In any case, the way I am "measuring" my straight edges so far is by pressing them against the side of a 24 or 48" level... it looks like its been machined flat to me... has those little swirling circles... and then I hold it up to the bright sky (note: this technique will only work for another month in Seattle ).

So, what is the correct way or better way to do it?

What is an acceptable margin of error for a given project?

Thanks!

ps: I think things are going pretty well at this point... finally got some good results with the jointer (tho still need practice) and the saw seems "fixed"... I readjusted the fence parallel to saw and tightened it up). Thanks to many of you who have been helping.

[drumpriest]

That is a good enough technique if the level is straight. Which it may or may not be. Also, if it's an aluminium level, you can flex the level somewhat. You need a reference plane. I made one (assembly table), but for edges you can do it with a factory cut edge on some MDF. MDF is both stable and flat. The factory edges should be straight, use it to test you level. You can also hold 2 jointed boards to each other, if they match up, your edges are straight. Test the boards both ways.

I personally will glue something up if I can bend it together by hand. (tiny amounts of "not straight" over a small area arn't really a problem).

[Tom Hintz]

Over the years my definition of "straight enough" has grown more and more precise as I saw how much easier it was to produce clean, straight projects when the stock was machined as close to perfect as I could get it.
Though the "no light" method is actually a very good indicator, I use very precise straight edges to check most pieces, just to be sure.
As I gained experience on my machines, and in setting them up correctly, the edges I produced grew more accurate. I still check and continue to strive for no error.

[davidtu]

Surprised to learn an aluminum level can "flex"--they are aluminum--I thought it was a very inflexible material (e.g. bike frames of aluminum don't flex vs. steel).

I should say that I am looking for ANY light showing thru between the level and the edge... but I am accepting as straight if I have to angle the pieces back and forth in order to see the slightest bit of light.

And a little off-topic, but should I cut pieces 1/32 too large in order to run it back over the jointer again after the table saw? Or will I presumably gain the table sawing skills to get a gluable edge? (right now I am getting curved saw marks on the sawn edges.)

[gmack5]

(right now I am getting curved saw marks on the sawn edges.)

The "Curved" Saw Marks imply that either your Fence is not properly alligned or your technique is flawed. You need to take steps to correct the problem.

[LCHIEN]

I got a 6' level for checking straight, yes they can flex (anything can flex given enough force) but I try not to stress it to that point.

Swirl marks means your saw alignment is off. Usually the rear of the fence too close to the blade, the rear of the blade is probably making those swirl marks. Check the direction of the swirl marks to see which end of the blade is making them. If the front of the blade is making them, then you need a featherboard... you are using one, no?

[davidtu]

If the front of the blade is making them, then you need a featherboard... you are using one, no?

Yes, I am using a featherboard.

Hmmm... sounds like I have to re-check the fence... isn't this where we came in? Actually, I will have to see if I am still getting the marks... I am asking b/c of some pieces that I'd cut a few days ago, and that might have been pre-readjustment. Also, I was using the jointer, trying to get a perfect edge, and that's whereI was using the level to see the light--those I could get to almost no light... just the slightest hint of a line if I angled the piece back and forth.

Are you guys able to get absolultely no light, no matter how you angle your view? Not even a hair? And, for using a straight edge are you just seeing if it lies on it nice, or looking at the light w/ that. Seems like straight edge is less forgiving than the level (b/c of its thickness, or lack thereof).

Regarding the saw marks, could it be related to the size of the pieces? These pieces were pretty small; 8-14" long, narrow slats (cut to 1.5" wide from maybe 2.5")

[Tom Miller]

Concerning the saw marks -- another thing to check is the alignment of your riving knife. If not properly aligned, it can pull the work piece away from the fence.

Regards,
Tom
p.s. Once you get your edges for jointing perfectly flat and parallel, you'll learn about "spring joints".

[davidtu]

Concerning the saw marks -- another thing to check is the alignment of your riving knife. If not properly aligned, it can pull the work piece away from the fence.

Regarding riving knife position, I've seen it suggested two ways: 1) centered behind the blade, 2) aligned to the fence-side of the blade's teeth. Which is "right"?

[Tom Miller]

Regarding riving knife position, I've seen it suggested two ways: 1) centered behind the blade, 2) aligned to the fence-side of the blade's teeth. Which is "right"?

Aligned to the fence-side of the blade's teeth makes the most sense to me.

Regards,
Tom

[cgallery]

Your eye can perceive light-passing gaps of less than .001" (one-thousandth) very easily. Note that I say less than. It depends on the angle and intensity of the light, but .001" is easy to see with a dim bulb and a casual glance. A lot of people use the sun and, of course, you would be hard-pressed to find a more intense light.

You could be easily flexing the level .001". Or the piece you are checking. Or both together! And heck, I doubt the level is even straight to .001" to begin with!

What you need to do is find out if your reference (level) is straight. Or get a reference that IS straight.

You can try to use installed window glass and a feeler gauge to check your level. This can be time consuming because you have to use several windows and record your results across all of them to make sure they windows aren't distorted in the frames. And if you have a long level you may have to find a commercial establishment to find large-enough pieces of glass. You can get strange looks from passers-by.

But it is cheaper and more fun than spending $100+ on a piece of reference steel.

Good luck!

[cgallery]

Are you guys able to get absolultely no light, no matter how you angle your view? Not even a hair? And, for using a straight edge are you just seeing if it lies on it nice, or looking at the light w/ that. Seems like straight edge is less forgiving than the level (b/c of its thickness, or lack thereof).

The straight-edge may be less forgiving because it flexes more and creates a gap. The level may work better, but you may have to scrape it to make it straighter.

I know that if I attempt the test you are conducting (moving piece from jointer to reference straight edge and looking for light in the gap), I will always see a little. Probably (guessing) .0005" to (if I'm being sloppy or the tool isn't cooperating) .001".

I have two brand-new Starret 12" combination squares with hardened steel rules that I think they machine to .0002" or so per foot. When I lay them on a piece of glass, touch them edge to edge, and backlight with a bright light, I have no problem seeing a light gap.

[LCHIEN]

The question should not be how straight is straight, but how much error is straight enuf?

Placing two jointed boards edge to edge doubles the error and makes it easier to see gaps.

Boards are always flexible, the wider and thicker thay are the less flexible they will be to allow clamping to bring them together. More boards in a glueup will also build up problems if you have to close up gaps on each pair. So the allowable gap depends on several things.

I think if you make glue-ups and you can close the gap with little pressure then it's OK. It doesn't have to be perferctly straight. That's life.

If you want to quantify it, use two boards and a feeler gauge.

.001" is probably ridiculously small over a 4-6 ft. board.

P.S. just offhand, gluing up 3/4" boards with biscuits, I would say a combined gap of 1/64th inch or less over 4-6 feet would probably be pretty good.

[sbs]

Surprised to learn an aluminum level can "flex"--they are aluminum--I thought it was a very inflexible material (e.g. bike frames of aluminum don't flex vs. steel).

Steel is 3x as stiff as Al.

[drumpriest]

All bike frames flex, just a matter of how much before it fails. Aluminium bike frames are considered to be "stiffer" than steel frames for bikes, thus the introduction of carbon fiber front forks and rear seat stays, but I think that has more to do with geometry than the inherant material properties.

As Loring and I both said, one great test is with 2 jointed boards, flipped to double the error. If that's a good fit, you won't have any troubles with a glue up. I've never made a cut so good that NO light passes through, but you don't want any gaps.