Level as straight edge

[jussi]

I seem to recall someone saying that a level as a straight edge, I think it was Sarge. I'm thinking about this one. Which is on sale btw.
http://www.amazon.com/Empire-e70-48-...8986808&sr=1-8

[LCHIEN]

I have suggested that they make good straightedges...
but the .001" per foot or so they spec is not straightness but error from level... so I don't really know for sure.
I can suggest how to check...

I think I got my 72" Stanley level from Lowes for around $30 on clearance, maybe that was a real good deal.

[jussi]

I can suggest how to check...

How?????????

[LCHIEN]

How?????????

I posted these a couple of weeks ago:

We would assume that the two sides of the level are parallel. This is easy to check with calipers... like the $16 harbor freight digital ones.

Once parallel, then its pretty easy to check the level for straightness even with a bowed board. If the board is slightly bowed, place it so the low point in in the middle.

The level has two edges A&B. place A against the board, measure the gap with your feelers at a couple of places along the length.

Now flip your level so B is against the same places that A was. Again, measure the same places with the feeler gages. Subtract the gap

If the gap is the same then the level is flat. If the gap changes in one direction, so that lets say its 5 mils higher in the middle and 2 mils higher between the ends and the middle, then the level has a bow of about 2-1/2 mils over its length (half the difference in gap in the middle).

If everything is the same, then the level is pretty straight.

Here's another way to test:
take a relatively flat piece of plywood at least as long as the level.
Lay the level so its on its side and you are using the bottom as a straightedge, draw a line on the plywood the length of the level.
Now flip the level over so its got the opposite edge against the line just drawn. Align it so the level is at the line at the ends. Draw another line using this edge. If the two lines drawn are conicident the level is straight. If the lines diverge in the middle, then the level are not straight by half the gap between the lines (which you should be able to measure if big enough.) It's very important to keep the pencil near vertical with the actual lead against the edge and maintain this angle the whole length.

I don't think I'd worry about 5 mils or less over 6', probably 10 mils is OK. once you get over 16 mils you're at 1/64th inch and now its a judgement call. Bear in mind that a 0.5mm mechanical pencil lead is .019" (19 mils) thick.

[jussi]

Cool. thanks for the tip.

[lcm1947]

Loring's post is certainly an excellent way but how about this and I think Loring actually mentioned it some time ago in some post or another and that's where I got the idea. Since you have to go to a store to buy one anyway just compare them to each other. Example; hold two levels together and see if any light comes through. If not I would imagine they are pretty level. Anyway that's what I did when I purchased mine and I bought the FatMax 24" & 48". I did test others as I was hoping to spend less the what the pricey FatMax's run but couldn't find two that didn't allow light to come thur.

[cgallery]

I posted these a couple of weeks ago:

Using a bowed board to check a straight edge, brilliant!

[jarhead]

...Since you have to go to a store to buy one anyway just compare them to each other. Example; hold two levels together and see if any light comes through. If not I would imagine they are pretty level...

Isn't it only true if both levels compared are level. If you picked up first level that is level and the second is not, then you'd conclude that both are not level. What are the chances that you'll pick up 2 levels that are perfectly level? I guess if you have a friend with a LV level, borrow it and use that as a reference.

[p8ntblr]

What are the chances that you'll pick up 2 levels that are perfectly level?

Then I'd just pick up another level until I find one that is. And if I couldn't find another one I'd pass on the company entirely. After all if they can't make more than one level that's flat then they're probably not a company I want to do business with.

[lcm1947]

I think p8ntblr got it. If both do match up then I would assume both are level. If the first one you picked was level and the second wasn't then you'd see light between them anyway. You need to keep checking until you find two that don't allow light through. What's the chance of both being un-level exactly opposite of each other to where they wouldn't allow any light through. Both are level. Anyway it makes sense to me. Let's ask Loring, he'll know.

[LarryG]

Every level is not going to be level unless you hold it level.





(Sorry, couldn't resist ...)

[lcm1947]

Now that's helpful LarryG.

[LCHIEN]

I think p8ntblr got it. If both do match up then I would assume both are level. If the first one you picked was level and the second wasn't then you'd see light between them anyway. You need to keep checking until you find two that don't allow light through. What's the chance of both being un-level exactly opposite of each other to where they wouldn't allow any light through. Both are level. Anyway it makes sense to me. Let's ask Loring, he'll know.

if you're asking...
You treat one level (lets call it A) as the unknown bowed board I described, then test the other level (B) as I described in the procedure. It really does not matter whether A is bowed or not.
You check the difference with feelers. Flip over B and check with feelers again. If the difference with the feelers is the same with B flipped, then the whole of B is flat. If the difference is bigger on one side of B then on the other then B has a bow. You may want to check more than one point.
This assumes that the Level B is uniformly thick and its sides therefore are parallel to each other. This can easily be checked with calipers - more than say .005" in thickness would be pretty suspect, I think.

Click on the following for a full size view. Item A in this case is some reference board, it doesn't have to be straight. Could be another level or a board or a table top. That's the beauty of this method.

In the top example, the level "B" is bowed, and the space between them d1 and d2 are different. This method magnifies the difference.
In the bottom example, the level "B" is straight and d3 and d4 measure the same.

If the level is badly bowed, you don't even have to use feelers, you can just tell by looking at the change in gap when you flip "B".

[SARGE..g-47]

Morning Jussi...

It was me in the case you mentioned and I did get the Empire "True Blue" you were looking at on sale.. I took a Veritas straight edge to HD and checiked for straight. With .003 from stem to stern and that's good enough without having to go to a machinisst straight edge.

So.... do what ya have to do as the True Blue will suffice.

Regards...

[LCHIEN]

As a follow up I measured 4 levels I own.
Sears 24" machined aluminum
Stanley 48" extruded (lightweight) aluminum #42-076
Stanley 72" painted extruded (lightweight) aluminum #42-720 (which is their low-end level)
Ace 48" extruded heavy aluminum - I adopted this lost puppy and it has seen a really hard life as testified by the scratches and scrapes, and dents on the end caps and the mud and dirt. I think it was used by a contruction crew and fell off the truck, really.

Anyway, the thickness varied 2.191-2.193 for the 24" sears
2.128-2.129 for the 48" Stanley
2.308-2.312 for the 72" stanley except for the ends where the paint was thicker and it was 2.315 and 2.318
and the 48" Ace was 2.207 to 2.212 for all its misuse and abuse.

I lined up various levels next to each other (avoiding the painted ends of the 72" STanley which was easy because it was longest) and there was not much if any visible gap after flipping any of the levels.

I used someone's suggestion and held a 100W trouble light behind them and wow I could see a gap here and there! But I tried to "poor man's feeler gauge" it with a sheet of paper (.0035" -that's 3.5 mils) and I could not even start to slide it through the gap where I saw the light.

My conclusion is that my levels are pretty much flat to much better than .005" bearing in mind that the painted one has some additional .004" thick (half the paint on each side) bumps on the ends.