Does a House's LEVEL Change, Day to Day?

[Josh]

I'm getting ready to set some ceramic tiles on my kitchen wall for a backsplash. A few days ago I screwed a couple of very straight 5' steel bars end to end horizontally along the wall studs, to serve as a ledge to mount the first row of tiles on. I was very careful to get the two bars' ends completely flush with each other and level, using a good accurate 48" Johnson level. I checked their level later that night, and again the next day. Both bars were dead-on level. Yesterday I checked their level one more time prior to beginning the tiling and guess what: they're both off somewhere between an eight and a quarter of an inch.

I set the level at the same place on the bars each time, and used the same side of the level. The bars can't slip....they're still securely screwed to the studs behind the drywall. And their ends are still flush with each other, so I have a good 10' straight edge along the wall. But what's up with the changing level?

This house is a 50 year old single story brick veneer, built on a concrete slab in West Texas. Is it normal for houses or individual walls to shift that much from day to day? Maybe due to temperature variations?

[Doug Miller]

Explanation: you may have used the same *side* of the level each time, but turned it end-for-end for the second measurement.

Your level isn't as accurate as you think it is. Put the level on the bars, then turn it end-for-end, keeping it parallel to the floor, and check again. I'll bet it reads level one way, and not the other.

[norminn]

Why steel bars? Perhaps they expanded enough to force one end up or down a little bit?

[Jim Elbrecht]

-snip-

By accurate- do you mean you checked the level just before using? You swapped the ends, and the level was the same? [being careful to wipe the bottom both times so some errant little chip or whatever wasn't throwing you off]

-snip-

[first off- for my money, there is a world of difference between 1/8 and 1/4 in a tile job. You should be thinking in 16ths at least- and 32ds will make it look that much better]

If your house settled 1/4 inch in 4' overnight- you might want to put off tiling until you find out what your problem is. If you were in MN,[40below zero-- F. . .brrrr] then maybe it frost heave could do it-- and if it did- then that wall isn't a good place for tiles.

I don't think it was that cold in W TX- so I would suspect

1. level got whacked- or

2.operator error.

1. Serious trouble that you need to verify and correct before you tile anything.

-snip-

Jim

[ransley]

If your house shifted 1/4" overnight you would of heard walls creaking, maybe plaster cracking, doors and windows would not operate the same, some windows might not even open easily and some doors might catch on the frame, I jacked up a few places and thats what happened. Remeasure it.

[willshak]

Josh wrote the following:

Where you at the same 'level' of sobriety both days?

[hrhofmann]

Or maybe he needs new glasses.

[catguy]

West Texas uh. Check this out.....Paul

[RickH]

I had a house boat that used to do that all the time.

[aemeijers]

I'd be more inclined to suspect the different expansion rates of the steel bars and wall structure during temperature swings. Most people use an eyeball-straight 1x4 for what you are doing. Not like you are lining up laser optics or anything, here. Lay a marble on the bar, and see which way it wants to roll.

Having said that, this is Texas, and the laws of physics on your infamous clay soil are different at times. Had any rain or big temp swings outside?

I'd not lose a lot of sleep over it. Most people set the counter and upper cabinets before they set the backsplash, and fudge as needed so it looks right. Eyes looking at it will key off the back edge of counter and bottom edge of upper cabinets. Tile is seldom gauge-block uniform in size and square anyway.

-- aem sends...

[Frank Warner]

You must store your grog evenly on both sides of the galley to keep an even keel.

-Frank

[Josh]

As for why I used the steel bars: They were the straightest objects I had on hand, plus they already had a lot slots & holes in them so I could run a screw through them and into the studs easily.

Anyway, I did as many of you suggested and swapped the ends of the level. Sure enough, both bars are off 1/8" one way and dead level when the level's ends are rotated. Question is: which is correct?

I thought I could determine the "accurate" end by checking the bars' level with a couple of 24" Craftsman levels I have. One of shorter levels showed the same thing the 48" level showed (dead on one way and off 1/8" the other) and the other level showed a 1/8" error each, only exact opposite depending on which way the level is orienated.

Guess I'll go with the longer level, and be careful to keep it orientated the same way all the time. Like someone above said, tile rows don't have to be laser-precise level.

[Joe]

You need to check your levels and discard ones that are off calibration. Find a rigid flat surface longer than the level. Place the level on it, check the bubble and note the exact location (centering is not necessary). Reverse the level (180 degrees). The bubble should be in exactly the same location. If the level passes this test it is OK to use it for your projects. If it fails, put it on the curb and get a decent level such as Johnson or similar. Considering the frustration you are having with your work, it would be prudent to buy and use a laser level. Prices range from $20 to $400+ depending on what you want, but the cheapies from Harbor Freight are perfectly accurate for tile work and the like. Even a line level will do a good job. The level hangs from the center of the line and needs two people to get it right. When anchored at one end of the line, the other end is raised and kept fairly tight until the bubble is perfectly centered. Framers and carpenters often use these low cost tools for lengthy measurements. They are usually within

1/8" or so of a laser level line.

Joe

Joe

[mike]

You sure about that? Passing your test suggests that the level is ok. Failing your test suggests that the level is bad OR the test surface is not level.

If the level passes

[Me]

The surface doesn't need to be level.

If the bubble is not in the same place in both directions, it's definately defective (or has debris interfering).

He is not suggesting that the bubble should be centered at all.

[hrhofmann]

If you have now figured out that there are some errors with the levels, why not fix them all so that they give the same reading no matter which way they are placed. Most levels can be adjusted by shifting the vials slightly.

[mike]

Neither am I...unless you expect the level to be "true".

Try this. Go up on your sloped roof and stick the level in the direction of maximum slope. Mark the position of the bubble. Rotate the level 180 degrees. Mark the location of the bubble. If they're in the same position, report back.

The ONLY surface that puts the bubble in the same position regardless of level direction is a LEVEL SURFACE.

On a LEVEL SURFACE, there is NO rotation of the level that makes the bubble go to a different place.

Your level is "calibrated" IF you have a perfectly level surface AND the bubble sits in the CENTERED position.

[Doug Miller]

I think the disagreement here stems from different interpretations of the phrase "in the same position" -- the same position relative to the surface being tested, or the same position relative to a specific end of the level.

You're thinking the latter; everyone else here is thinking the former.

Designate the ends of the level as "A" and "B". Place the level on any surface you wish, with end "A" on your left and end "B" on your right. If the surface slopes upward from left to right, the bubble will be off-center to the right, toward end "B".

Now reverse the level left-right, so that end "B" is on the left and end "A" is on the right, and replace it on the surface in the same position as before.

If the bubble is off-center to the right by the same distance it was before, the level is true. That's what's meant by "in the same position": relative to the surface being tested. It doesn't matter that it's now off-center toward end "A" instead of "B" as it was before.

[DanG]

Do NOT use that level. The level is out of calibration. If you cannot reverse the level, the bubble is not within spec. It is almost impossible for an excellent craftsman with a very good hand level to close around any room. It won't happen with a level out of calibration. You are compounding an error.

This is one time to consider a water level (cheap) or to borrow a laser.

[hrhofmann]

e:

I always use paper shims under the "low" end of the surface that I am using as a "level" reference. Once I get the surface level, then it does not matter which position I use to place the measuring level, it shoud be off by the same amount no matter which side is to the "left. Once I am sure the surface is level using the shims, I then adjust the vial so that the bubble is in the middle. Once that is done, I go back to the original slightly unlevel surface and see that the vial indicates the same amount of unlevelness no matter which end of the level is to the left.