GBA Logo horizontal Facebook LinkedIn Email Pinterest Twitter Instagram YouTube Icon Navigation Search Icon Main Search Icon Video Play Icon Plus Icon Minus Icon Picture icon Hamburger Icon Close Icon Sorted

Community and Q&A

Can I skip self-leveling compound and smooth concrete slab with tile mortar?

smoke_teff | Posted in General Questions on

Hello all. I’m finishing my basement, which has an old concrete slab floor I plan to tile. Slab isn’t in terrible shape, but has a few cracks, dips, high spots–at most there’s 1/2″-1″ variance in floor height over a 10-foot radius. 

Most online sources say I should smooth or level the floor with self-leveling compound in advance of the tile. I don’t really care about it being ground level–just want it flat enough the tiles don’t crack.

I’m wondering, can I achieve a smooth enough effect by just applying mortar slightly thicker in the dips, trying to make a flat plane with the tops of the tiles as I go?  That is, can I get away with not making the floor perfectly flat before applying mortar and tile?

The tiles I’m going to use are 6″ X 6″, about 3/8″ thick.

GBA Prime

Join the leading community of building science experts

Become a GBA Prime member and get instant access to the latest developments in green building, research, and reports from the field.


  1. GBA Editor
    Brian Pontolilo | | #1

    Hi Smoke.

    I've never done what you are describing but I have seen people do it, basically using different sized notched trowels to spread the mortar, with good looking results. I can't say how those floors held up over time. I have used leveling compound and it is a pretty easy process and creates a really nice surface to work on. I'm sure some others will weigh in with their experience.

  2. onslow | | #2


    This sounds an awful lot like a question I replied to about a week back. I would say that 1/2-1" variation is well beyond the fudging with trowel notches I described. The small tile size will echo the floor variation pretty closely as far as the big dips go. The bumps are more of a problem. What do you do when all the tile around the bump are set? Trying to target what is level across the area you described would be very dicey.

    It would be best practice and make a better looking floor to level it out as much as you can first. I prefer to use the vinyl modified cement over the self leveling since your slab may also have a general pitch in one direction. Also, self leveling can get tricky the deeper you go. A 10' radius means a 20 foot circle which is north of 300 sf.

    If the tile has a tumbled edge, you will have a bit more grace than a hard crisp quarry tile. I just set a couple hundred square feet of it and it is very unforgiving. Droopy corners will shout at you every time you enter the room. I may be overly anl, but I like flat and true floors.

    Get yourself a big straight edge and really analyze the dips and bumps. Check the overall level and try to see if 1" is really present. As for the cracks, well, you are on your own there. If the slab shifts again, the tile will separate at the grout lines if you are lucky and crack all over if you're not. If you spring for the separation membrane, do your leveling work first.

    If you have not done a lot of tile work, the notch fudging is more art than science. Good luck.

  3. oldbungalow | | #3

    our contractor did this in our 1920s basement but much smaller area (bathroom and hallway). He didn't use straight up mortar though- he used mortar with an arcylic admix, which provides strength.Large format tile mortar may be similar. I don't recall the amount of variance but the final result has been rock solid and the tiles are 1'x2'.

  4. Expert Member
    BILL WICHERS | | #4

    I’d think it would be a lot easier to apply one of the self-leveling floor leveling compounds and then tile over that. I don’t see how you’d ever get things even with troweling only with more than a few very small dips or cracks. With floor leveler, you have a nice, flat, even surface to work on.


  5. onslow | | #5


    Just want to explain a bit more clearly why I don't favor the self leveling this time. Imagine a shallow saucer that you want to fill level to the rim. If the saucer is anything but dead level at the rim, the self leveling material will flow to where gravity tells to go. A tipped saucer will leave you with an un-fillable divot to deal with. Since Smoke's floor has cracks and a discernable dip of up to an inch, I would be suspicious of just how true the floor plane is.

    Checking the overall field for level first might find the lucky circumstance of a perfectly level perimeter. In that case, filler up. More likely there is an overall tilt to one direction, possibly two. Even partially filling the divot will still leave a certain amount of hand work. I never said hand leveling was easy. For an area of the size described it will be very tricky. Once the real character of the floor area is established, placing leveling spots at about 3' or less intervals will help keep one on target. It is all a royal PITA and part of the reason people think tiling costs too much. Tiling through the dip with a small tile like Smoke's can result in the grout lines reading like a gravity well diagram.

    I once missed (got lazy) a relatively small 1/2" deep saucer on a patio that I was setting 16x16 tile on. As I set the first one crossing the saucer, it was immediately clear that I was hanging air under a quarter of the tile and in trouble, what with active thinset waiting on me. A more attentive check of the next tile position out showed me the dip vanished . I managed to pry the tile back up and re-grade the thinset with a 1/2x1/2 notch in the deep part of the dip. I feathered back the notch height as I exited and then carried on with a laid over 3/8. It all worked out, but I learned to check the floors much more closely.

    Ironically, this little divot would have been the exact condition I described with the saucer analogy. The client knew the whole patio had settled to one corner and the new tile would have to echo the tilt. Of course, no one would see or feel the unless the had a level or closely watched the scribed trim on the house side. I had checked the perimeter for flatness, noted the tilted plane and then made the error of only cursorily checking the field.

    KT - Super flex thinset is wonderful stuff and handy for small patches, but it is more costly than the generic vinyl patching cement. Smoke sounds like he will be needing a fair bit. What I have on hand is often the determining factor.

    1. Expert Member
      BILL WICHERS | | #6

      Thanks for that, I see what you mean. With a tilted floor, you’d end up with a lot of leveler on one side that could build the floor up so much as to be a problem on that side.

      I have something of a similar issue with a room that was originally a patio but got built into the house as a regular room. It has a hump in the middle. I’ll probably have to grind that part of the slab down (really not looking forward to that), but leveler would build up the edges where the doors are too much.

      I’m more used to slopes toward central cracks where leveler works pretty well. Too bad concrete can’t be more like a nice, flat piece of steel :-)


    2. smoke_teff | | #7

      Thanks for the detailed response. I can offer a bit more information on my situation.

      I was originally thinking of using self-leveling compound myself, but was apprehensive as it would be my first time, I would have to rent the equipment, and I found it difficult to estimate how many bags I would need (in any event it seemed it’d be about $7-800 worth of material). I brought in one cement contractor who recommended “gauging” the existing floor with a roughly 1” new layer of concrete (not self-leveling) which he would finish flat, if not ground level. He quoted me $2500 for roughly 400 sq ft. This was much more than we wanted to pay. He also suggested we could leave the floor as is and just go with a floating floor of vinyl click-and-lock planks, and not sweat whatever minor unevenness might telegraph through. The ease of application of the vinyl planks is very tempting, though I already have all the ceramic tile I would need--so buying the planks would mean an extra ~$800+, an extra usage of extracted/synthetic materials, and lots of wasted ceramic. Plus I'm not sure how well the planks would perform on the floor in its current state--would the occasional bump/dip mean the seams between tiles would stretch or come undone? Would you notice and be bothered by every bit of unevenness you walk across?

      Another cement contractor came over last night after I posted this message, this time one who specializes in leveling surfaces. He said he would use self-leveling concrete, but would mix it slightly thicker than usual—this would make it spread out more slowly, so he could get it to achieve a flat plane overall without it flowing until it achieved ground level. This would seem to solve the “saucer” problem you illustrate, though it would leave some minor tilt in the floor plane. Though the floor is close to level as is. The basement is roughly 12’ wide by 35’ long. There’s a height difference of about an inch in each of those directions. This contractor says his application will probably cut that difference in half, which I am happy to live with rather than doubling my cost of materials to achieve perfect levelness.

      The second estimate was also much more affordable: $1200 (nothing like a $2500 estimate to make $1200 seem cheap). The guy gets great reviews and specializes in this work, so I trust he can do what he says.

      I now think I’m likely go that route, having the slab professionally smoothed/flattened and applying the tile myself. It would also include a lot of the work I'd have to do myself before applying any floor material--scraping and prepping the concrete, patching the larger cracks, even grinding down the larger bumps. I already have a bunch of the roll-on elastomeric membrane to prevent cracking and help with waterproofing--as recommended by someone else on this message board. Plan is to paint that over the flattened surface, then the thinset/tile on top of that. Happy to be dissuaded if there’s some obvious problem I’m failing to consider.

  6. onslow | | #8


    It is of course your money, but I would tend to avoid the plank flooring. It may be thicker than sheet goods or tile squares, but I suspect that over time they would conform to the floor deviations in a way that would look bad. Vinyl has a way of conforming to the tiniest of lumps. If the second fellow really is well reviewed it sounds like a better way to go. I am too far out of the paid work loop to know what charges people ask these days. Of course the virus slow down is a strong motivator. Since you will be doing the labor on the tile setting it might be palatable to accept the cost under the time savings you will have by not fighting a wavy floor. Be sure to ask how long whatever material he used needs to cure before coating with the membrane. Best of luck and don't set out more than 10-12 tiles worth of thinset at a time. You have a task ahead.

  7. NickNation | | #9

    This is way past the date of the above posts but thought it might be helpful for others since i was here looking for answers myself.

    I have a similar situation and had similar questions but went ahead with self leveling myself and would like to offer the following advice to other beginners. Hint: it was a lot of work. And for those above that think that because your floor is a tilted birdbath and it’s all gonna flow to the low end, it’s not. It’s not at all like water.

    I had 450 sf, basically a 14x30 rectangle (varies past the stairs to basement) with a 9x6 nook. Floor was fairly uneven and one end was about an inch higher than the other although it looked okay just looking at it. Until you put a straight edge on it. Hydronic heated floor and was probably hand finished with high spots at all the corners and around stairs going to basement.

    I would advise getting a 10 foot straight edge because a 6 foot or less doesn’t give you the right picture. Invest in an aluminum one that you see in the YouTube videos or make one with the factory edge of thick plywood and at least 4 inches wide so doesn’t flex which is always very straight. Or wide mdf nailed to a 2x4. It became apparent that the perimeter walls were about right even if not level and the rest of floor was lower but wavy or bumpy. I mapped out the highs and lows with marker on floor.

    I thought I could just cover the place with self leveling and I got family involved, 6 people and me with multiple 6 gallon buckets, measured water ready, 2 drills and paddles. So it was a production line with mixing and cleaning and got it all poured in under an hour using 16 bags and I went with the max amount of water allowed (levelquik rs) for flow. Itwas 10 degrees celsius outside which helped in slowly down the setting. It went fairly smoothly and it looked like a beautiful lake of cement. But 4 hours later I was disappointed to find out that it wasn’t flat. Remember the goal is flat not level.

    First mistake was pouring too much near walls and/or slopping it up against the walls. This forms high spots. And then spending too much time trying to smooth after. I would try to pour away from wall and gently bring it near but don’t worry about it. If low you can always go back later and fill it by screeding from the flat floor. It’s much harder to fix high spots. Also, quickly smooth it and that’s it. If you play with it too much, even just a few back and forth strokes, you’ll get ridges. And you don’t need to dump the whole bucket at once as it flows once dumped, even 10 minutes later but once on the ground it starts setting soon.

    My high spots remained high spots. In fact I think they act as dams for the compound. Remember that self leveling compound is only self leveling to a point. It needs help. I thought you could just fill the place up and it would fill low areas with my help. Nope. I mean kind of. I should have removed the high spots first, even if it means covering them after because it would be smoother. And you don’t need to rent a grinder. Just use your cheap air hammer and chisel and cut it out. I would quickly sharpen chisel blade on bench grinder occasionally and goes better. The leveler will fill the rough spot. One good thing about leveling first and then using air chisel is that it removes more easily than the concrete with much less dust. Then just smooth it with patching cement.

    So I spent weeks trying to make this newly poured self leveling floor flat and I eventually did. If I could do it again I would have filled all the “ birdbaths” first with leveliquik and/or patching cement and removed high spots. And then poured the leveller over the whole thing to make it nice and flat and uniform.

    Another thing for getting floor ready or after trying to fix your pour is leveling one direction first (the long direction that you will lay tile), then the other. It seems very confusing where to start with a bumpy wavy floor so just do this. So I filled and flattened long direction even though there were dips perpendicular. Then after the whole floor was done I did the other direction and it got flat in ALL directions. I was amazed. Then just a bit of patching.

    One final tip is if you’re filling some small birdbaths, draw circle with marker and just mix enough for that by using ratios, don’t guess. Roughly estimate the volume from the chart on the bag. So if going 3/8 thick it does 14.7 sf. So 50 lbs/14.7=3.4 lbs per sf , then do ratio of water in litres (not quarts) because easier to measure litres in fractions using measuring cup. I estimated volume by measuring sf of area not including a couple inches on edge for feathering and it was pretty accurate. So if area was 2x3=6, that’s 6x3.4lbs=20.4lbs. I weigh myself on scale then weigh myself holding pail of powder because it’s too light to just weigh pail. And if 50lbs is 5.44 litres water then 20.4lbs is 2.22L water. Then pour just enough to cover bottom of straight edge and saw back and forth to screed and add more as needed, scraping back anything that goes over line on ends. Then after an hour, scrape up edges to improve your previous feather edge and scrape up drips. Much harder later

    Good luck beginners, you can do this !

Log in or create an account to post an answer.


Recent Questions and Replies

  • |
  • |
  • |
  • |