19

Even though I enjoy DIY, I also suffer from a certain amount of anxiety about it, especially when tolerances have to be tight. If the Internet's advice is that a certain thing needs to be square, or level, or plumb, I am never confident that I have achieved enough squareness, levelosity, or plumbitude. Should I fret about a 32nd of an inch? How about a 64th? Will the contractor I hire for phase 2 judge my handiwork in phase 1? These are the questions that keep me awake at night.

You'd think that laser levels would help my predicament. But they often don't, for multiple reasons:

  1. How wide is the beam, Part 1: if I'm trying to cast a level laser beam across the entire span of a 40 foot wall -- with the laser pointed perpendicular to the wall at the 20 foot mark -- the beam is going to be fatter on the ends than it is in the middle. If I'm trying to draw marks along that wall that just kiss the bottom of the beam, marks at the ends of the wall will be ever so lower than the marks closer to the center.
  2. How wide is the beam, Part 2: a beam in a bright room will seem narrower to my human eyes than the same beam in a dark room. Depending on the shape of the room, and the nature of the available lighting, some sections of walls can be less illuminated than others, making the beam on those less-lit walls seem fatter.

(To both of the above, you might say: "mark to the center of the beam, not the edge!" But what if I don't trust my hand-eye coordination to consistently mark the exact center of the beam at each point of reference that I draw? And what about when the pencil mark ends up just a skosh off of where your hand actually intended it to show up?)

  1. How perpendicular is the laser to the wall you are projecting to?: Suppose you are projecting a plumb beam onto a wall , to mark the center of some feature that you are going to hang. (Perhaps a vanity mirror, that you have also marked the horizontal center of.) If your laser device is not perfectly perpendicular to the wall you are trying to mount the thick framed mirror on, there will be a parallax error between where the beam hits the wall, and hits the mirror frame.

  2. How flat is the wall you are projecting to?: even if your laser level device is perpendicular to the wall -- when projecting a level beam, if your wall is out of plumb more on the left than it is in the right, the projected beam will appear to curve slightly upward (or downward) in the region that is out of plumb.

So I guess my overall question is: in residential finish remodelling, how do professionals decide what is an acceptable level of precision, and how do they achieve it in an imperfect house, with an imperfect environment, even with a laser level?

GogTheGuilder
  • 433
  • 2
  • 11
  • 15
    Often times, looks are more important than measurements, like averaging the tilt of a wall and a cabinet so that it's more plumb/level than the wall, but not so accurate that it looks crooked. – dandavis Jan 31 '23 at 09:47
  • [Highly related](https://woodworking.stackexchange.com/q/14406/93) – FreeMan Jan 31 '23 at 13:02
  • 8
    If you want to be this much precise, at 50m / 150ft you will have to take into account the curvature of the Earth – fraxinus Jan 31 '23 at 15:53
  • 17
    I hung a display shelf above our bedroom door. I had 3 choices: 1) level according to gravity (spirit level), 2) parallel to the top of the door trim, 3) parallel to the ceiling. (Oh, the joys of a 130 year old house!) I chose 2) because the other two options _looked_ wrong. – FreeMan Jan 31 '23 at 19:34
  • One of the important features of a laser is that the beam doesn't widen. It is collimated light. Unless you work in a thick smog with particles to scatter the beam, and projecting to replicate a mile of subway tunnel rail section - the widening is going to be less than your pencil mark diameter. – Stian Jan 31 '23 at 23:59
  • 3
    Just don't start wondering how smooth is smooth enough and you'll be fine... That worry is a time sink for real. – Stian Feb 01 '23 at 00:00
  • @GogTheGuilder, regarding beam width: you could draw a line to mark the width of the beam and then measure and mark the center of that line. – Celos Feb 01 '23 at 08:36
  • 7
    Move to a house like mine which is 300 years old. Nothing is straight or parallel to anything else. It's quite liberating. – Jack Aidley Feb 01 '23 at 08:49
  • I love all of these comments. Too often things are already unlevel, so you have to pick what looks best - not "true" level. That said, you could always do what I did and use a transit level because you had one laying around from that time you had to tell the contractors to tear up the pipes they just laid. – iDriveSidewayz Feb 01 '23 at 14:58
  • 2
    If you're worried about 1/64", go look up the thermal expansion coefficients for your materials and figure out how much things will shift between winter and summer. Or just humidity, the "landlord special" fitting of the door on my apartment won't stay closed without the deadbolt in winter but takes a bit of force to close in summer due (I assume) to the wood swelling. I suppose it could also be the whole structure shifting from thermal expansion but that's less pleasant to think about. – llama Feb 01 '23 at 15:41
  • 2
    and if you *really* wanted the maximal precision for your laser level, you could use ND filters (or crossed polarizers for continuous adjustability) to decrease the intensity until the only part of the beam you can see is the centre (assuming a gaussian beam which you may not get from a laser level, in which case you could obtain one by passing through an iris placed between two lenses arranged so that this is their mutual focal point... hopefully the level of effort and expense implied here is sufficient to make you happy with "level enough") – llama Feb 01 '23 at 15:54
  • 5
    It's been said that the difference between a pro and a DIYer is the pro knows how much tolerance matters for the job in hand, and optimises for speed within those tolerances. The DIYer doesn't know what is 'good enough' so builds to the best tolerances they can - which may be tighter but often takes much longer. – user1908704 Feb 01 '23 at 16:52
  • 1
    @FreeMan You had at least one more choice, which is often the best - average out the alignment between whatever is above, and whatever is below. – MikeB Feb 02 '23 at 12:23
  • 1
    Could have, @MikeB, but it's now _obviously_ parallel to the top of the door trim and, with decorations on it, the line with the ceiling is not visible, so the _unparallel_ to the ceiling isn't visible at all. (Except when on a ladder swapping out decorations). I'm now in the best of both worlds situation. ;) – FreeMan Feb 02 '23 at 13:49
  • I'll upvote just for _levelosity_ and _plumbitude_, but this is a good question. – gnicko Feb 02 '23 at 14:15
  • I wouldn't worry about it. Something that's spot on plumb now probably won't be after 5 years anyway. – Huesmann Feb 02 '23 at 18:51
  • 1
    Accuracy to 1/32" over 40' is more than 1 in 15,000. No-one will **ever** work to that tolerance in a building!! – Tim Feb 03 '23 at 09:12

5 Answers5

35

A good technical response by @martinbonnersupportsmonica but I’m going to broach the psychological side of the question.

In short, good enough is usually good enough. Do your best. Check your cabinet hanging at the last minute with a good quality spirit level (cough, cough stabila). Check that level against itself regularly. Things that want to be dead level/plumb: cabinets, doors, windows.

That said, making something look natural in its setting is important — more important than making it micrometer plumb. This means cheating with trim where you can, or splitting the difference when faced with existing conditions that aren’t perfect.

For example, a plumb upper cabinet butting into an out of plumb wall will have a terrible reveal. The solution might be a wide filler strip that tapers and cheats the eye.

Don’t fret about accuracy so much as visually pleasing results. You might know that something is off by a tiny bit, but almost nobody else will. And eventually, you’ll forget as well.

Aloysius Defenestrate
  • 18,266
  • 1
  • 30
  • 52
  • 3
    I have a long workbench that I needed level so things wouldn't roll around. It's a few inches below the opening of a half wall in a room that is *not* level. The left side is maybe 2" below the top of the half wall and the right maybe 4". It's remarkably jarring to look at. – Reid Jan 31 '23 at 16:29
  • 3
    *And eventually, you’ll forget as well.* Like my dad says, "after 6 months, no one will know or care". – computercarguy Jan 31 '23 at 21:20
  • 3
    @Aloysius Defenstrate: I hadn't heard of Stabila before your answer. But, I'd consider a $200 spirit level if it doesn't have a different opinion depending on which 180° orientation it was placed in. I'm tired of trying to average 2 bubbles in my head. – GogTheGuilder Feb 01 '23 at 02:22
  • 2
    A spirit level that can't agree with itself is a curse. Not to say that you have to spend a fortune on Stabilas, but they are nice. – Aloysius Defenestrate Feb 01 '23 at 03:42
  • I like this. If the work you do does not create a hazard and *you* are happy with it, then great! – Michael Mior Feb 02 '23 at 14:41
  • 1
    @GogTheGuilder - there's a good place for a spirit level that's not level - in the bin. It's not working right, so throw it away. Most are not adjustable - why should they be? They're either good or throw-away. – Tim Feb 03 '23 at 09:08
  • @Tim except /every spirit level I've ever bought/ gives a new opinion when I spin it around the other way. Surely not everyone here has bough those fabled Stabilas? – GogTheGuilder Mar 07 '23 at 02:02
  • 1
    @AloysiusDefenestrate thank you so much for this answer. It has helped me understand (A) to never keep going in hopes that angels will eventually sing, and (B) I am not Morty Sanchez, and thus I will never experience "true level". (Although, today I did get an eye twitch when I discovered that my 2 laser levels disagreed about what a "plumb" line looks like.) – GogTheGuilder Mar 07 '23 at 02:05
  • @GogTheGuilder - I must have 8 or 9 levels, of differing lengths and capabilities. I always check them 180 degrees before use, and some are maybe 30/40 yrs old. They wouldn't have got this far if they hadn't passed the simple test. – Tim Mar 07 '23 at 15:01
20

A quick search for "visible out of plumb" found this page, which is a detailed list of acceptable tolerances https://www.premierguarantee.com/insite/tolerances-in-internal-and-external-walls:

  • Flatness of internal wall: maximum 5mm offset from a 2m straightedge;
  • Plumb of internal wall: max 10mm out of plumb for a ceiling height of 2.5m or less;
  • Level of ceiling: max 10mm over 2m;
  • Squareness of corners: max 15mm deviation from a 500mm square.

If you are not comfortable with metric, 5mm is about 3/16th inch and 10mm is about 3/8ths. So not only should you not fret about a 1/32th, you shouldn't even fret about 1/8th.

I think the context of the page is "when will a building guarantee pay out", so these are quite generous tolerances — but it gives some sense of what is acceptable.

Oliphaunt
  • 105
  • 2
Martin Bonner supports Monica
  • 3,023
  • 9
  • 19
  • 2
    My concern isn't with the inexactly-constructed walls, per se'. Rather, it's with the more-perfectly manufactured accoutrements that need to be added in a precise way, to an imprecise environment. (Today, those items were: bathroom cabinetry and mirrors.) That page is a great reference, though -- thanks! – GogTheGuilder Jan 31 '23 at 07:32
  • 4
    @GogTheGuilder that's why shims were invented... – Solar Mike Jan 31 '23 at 08:21
  • 19
    You'd be shocked, @GogTheGuilder, how not-precise these "more-perfectly manufactured accoutrements" actually are. – FreeMan Jan 31 '23 at 13:03
13

It's only a problem when it's a problem, and every structure has flaws.

  • A relatively flat floor that's out of level 5/8" over 40' isn't a problem. No human can detect that. A wall that's out of plumb 5/8" in 8' is a problem. It can bee seen, and it probably affects many other things it relates to.
  • A cabinet that's hung 1/8" out of level isn't a problem... unless it connects to a 20' series of cabinets which will magnify that to an inch.
  • A roof deck with a 5/8" dip in one truss is a problem. It'll look terrible. A roof deck with a 5/8" swale over 6 trusses probably isn't noticeable and doesn't affect anything else.
  • A drywall corner with a rounded-out tape joint isn't a problem... unless you're fitting a granite countertop that you'd rather not round out to match.

Every situation is different. There's no simple answer, and carpentry is as much art as science. It's mostly wood, after all, so no matter how precise you strive to be it'll confound you. That isn't something you can control. You'll have to be ok with that or go build computer processors instead.

isherwood
  • 119,766
  • 7
  • 148
  • 349
  • 2
    FWIW, the room containing the workbench I described in the other comment is something like 3" off level over 10 or 12 feet (and stable over the decade or so we've lived here). Aside from not agreeing with the workbench, the room is just fine and feels plenty level. If we had a dinner table in there maybe that would be a problem. – Reid Jan 31 '23 at 16:34
  • 1
    3" over 12' doesn't seem crooked?? – isherwood Jan 31 '23 at 16:37
  • 7
    1/4" per foot, eh? Should drain well... – FreeMan Jan 31 '23 at 17:00
  • 1
    Computer processors take the art of dealing with things that are impossible to build perfectly to plan to the extreme. No two processors are the same (on the scale that matters, i.e. nanometers), they were carefully engineered so the differences don't matter. – Nobody Feb 01 '23 at 10:10
  • 1
    Yeah, but I can't see the slope in a microprocessor from my couch. – isherwood Feb 02 '23 at 13:49
6

The Six Month Rule

I always go for 120% accuracy, then settle for 80% when all is said and done.

The remaining gap will close back to 100% satisfaction all on its own in 6 months, when I am busy using the space for what it was intended and forget about all the flaws that nobody else sees either.

The Beholder Rule

For DIY work there is no hard standard. Level and aligned to the eye, smooth to the touch, and no glaring mistakes in sun light. If these criteria are not met it doesn't matter how accurate the work is. I use spirit levels, a laser level and numerous attempts at dry-fitting to ultimately arrive at a final install.

For Pro work there are precise standards because unhappy customers will find any flaw to complain about or to use as settlement leverage. No matter how great and smooth it all looks, if there is a technical flaw there is technically leverage.

P2000
  • 15,457
  • 26
  • 49
  • 1
    What exactly is accuracy above 100%? Are you allowing for settling or shrinkage or something? This strikes me as hyperbole. – isherwood Jan 31 '23 at 22:26
  • 8
    yes @isherwood hyperbole and attempt at humour. 100% being whatever you get when you follow normal procedures (e.g. measuring tape + spirit level) and "120%" after doing two more measurements, a laser level, two dry fits -once before and once after lunch-, the spouse's eye, father in law, and one more measurement from another angle. – P2000 Jan 31 '23 at 23:46
4

Regarding the worry that the laser line thickness will be different at one place than the other: It won't be, or not enough that you can tell. Laser light is emitted as a collimated (i.e. parallel) beam rather than the diverging or focused-at-specific-distance beam you would get from most light sources. There will be some loss of that characteristic as it goes through the cheap level's optics, but generally not enough to matter.

Mark Morgan Lloyd
  • 325
  • 2
  • 5
keshlam
  • 22,920
  • 5
  • 43
  • 79
  • In general, I'd agree with you. However, my **_el cheapo_** laser level would beg to differ... – FreeMan Jan 31 '23 at 17:01
  • 9
    That's a flashlight, FreeMan. – isherwood Jan 31 '23 at 17:38
  • 4
    Well, no, all lasers have a specification on their divergence. Picking a laser level at random (https://www.milwaukeetool.com/Products/Power-Tools/Lasers/3631-20#sp-specs) one will find the beam divergence in the specs (and with the wrong units there - should be milliradians not radians most likely, so about 0.15 degrees). – Jon Custer Jan 31 '23 at 19:14
  • 1
    Granted. But for this application 0.15 degrees is close enough to zero that it's likely to be the least significant source of error. – keshlam Jan 31 '23 at 19:21
  • Uh, yeah. One would expect there to be a difference in quality between Milwaukee and Harbor Freight... – FreeMan Jan 31 '23 at 19:31
  • Sometimes the difference between brand name and no-name is whether they bothered testing and calibrating before shipping... I needed one for a specific task, spent about $35 online, and it works more than well enough for the price. – keshlam Jan 31 '23 at 23:40
  • 2
    If in doubt, verify the marking by positioning the laser at different points in the room and checking that the line segments are a precise match. Frankly, I'd be far more worried about (a) the leveling accuracy of the laser head (b) the thickness of the pencil used to transfer marks to the wall and (c) my ability to position a (masonry?) drill with micrometer accuracy that I would about the properties of the beam. – Mark Morgan Lloyd Feb 02 '23 at 13:10