Curling Screed

[MJNewton]

I am currently building an extension to increase the size of our lounge and have had tradesmen in to do some aspects of job, which included laying an unbonded 60-80mm fibre-reinforced sand and cement screed to finish flush with the existing lounge floor and a concrete threshold running along a set of doors. Here it is:

 

 

This was done 6 weeks ago and, as per the screeder's instructions, it was left covered with a polythene sheet for a week to help the initial cure and has been left uncovered since at a room temperature of ~16c - not for any particular reason; although I was mindful that the slower it dries the less issues I would likely have.

 

Unfortunately I noticed today that the slab has risen ~7mm along the join with the door threshold as seen here:

 

 

(The black sheet on the right is the separating barrier between the screed and the concrete slab underneath but it is very thin and so not affecting the measurement like it might appear to be doing).

From what I have read, unbonded screeds can be prone to curling as a result of the top surface drying out quicker than the bottom and so whilst it might not be that surprisng that this has happened I am keen to get it sorted. It's not just the height differential that concerns me (some self-levelling compound to bring the lower levels up could likely sort that) but more the fact that you can feel movement in screed if you jump up and down there. There's no cracking (yet!) but it is definitely flexing.

 

The screeder is coming out on Thursday to have a look but I would welcome anyone's thought on what options I have, noting that I'll be covering it with a 15-18mm engineering wood floor eventually.

 

Thinking aloud I thought I/we could:

 

1. Do nothing - not really a viable option (for me at least!) as it'll obviously compromise the final floor covering

2. Bring the threshold height up with some form of self-levelling compound - this still leaves a poorly supported screed in the affected areas which may cause issues a later date (although I imagine engineered wood flooring itself provides some strength?)

3. Drill holes and force some sort of resin in/under the screed and then bring the threshold height as required - don't know anything about this as a technique; it just sounds like a possibility to me.

4. Try and get the screed to settle back down by jumping on it and fill what I assume would be a resulting crack

5. Cut and remove/replace the screed as required

 

I did read that sometimes the curl can be reversed by putting a polythene sheet back over the affected areas for a week to move the residual water balance. I am surprised that this could work but I've seen a few mentions of it in what appear reputable places. Does anyone know anything more about this? I suppose it's got to be worth a try before Thursday just to see if anything changes.

 

Grateful for any thoughts!

 

Edited January 26, 2019 by MJNewton

[Mr Punter]

I have not heard of this before.  Did they put anything compressible around the edges?  Also, is it the same against the lounge floor?

[MJNewton]

It was laid straight on to the concrete slab beneath, albeit with a DPM between the two (as I beleive is fairly common for this type of screed).

 

The join with the lounge floor has some differences in height but this is likely as much to do with the less-than-perfectly-flat existing lounge floor than anything else. Also, at this end the screen depth was more towards 80mm compared to the 60mm or so at the door end and hence is likely a bit stronger.

Edited January 26, 2019 by MJNewton

[nod]

It looks to me like it has been laid high

Covering a screed is ant necessary 

There definitely should have been an exspansion joint between the two floors 

The flexible plastic type that you simply cut back when the screed has set 

 

if the screeders coming back 

Its a good start 

offer to pay for the self leveler 

If he sorts it out 

He should have it perfect in an hour 

[MJNewton]

It was bang on when first laid, and indeed for a weeks afterwards as I kept an eye on it as I had read a lot about different screed failures and was mindful how critical my joins were. It's only since I've taken my eye off it has it moved! 

 

I think that even with some sort of expansion joint the curling would still have happened with the same consequences re heights. If anything, reinforcement between the two floors would have better to prevent differential movement.

 

I've been reading more about it and it seems it may even curl back down (slightly, if not fully) as the lower levels dry out, and it could of course get worse still before it gets better so it may be a little early to be making good with levelling compound.

 

This sound hopefuly from BS8204-1 (Screeds, bases and floorings):

 

Quote

e) Drying. Slabs and levelling screeds should be allowed to dry out as slowly as practicable after curing
to reduce the risk of curling. Accelerated drying by forced ventilation or heating should not be used as
this can result in incomplete hydration of the cement and cracking and curling of the screed.
The possibility of damage to the floor finish over joints or cracks, such as rippling of thin sheet and tile
flooring, can be caused by a reversal of curling due to the redistribution of moisture within the thickness
of the screed once it is covered. This risk can be reduced by covering the dry screed with impermeable
sheeting for about 7 days before laying the flooring.

 

I obviously want this reversal to happen (we are a long way off covering it) and so will try covering it as I'd read elsewhere. The screeder is not coming until Thursday and so I've got five days to see if any reversal occurs.

 

 

Edited January 26, 2019 by MJNewton

[nod]

2 hours ago, MJNewton said:

It was bang on when first laid, and indeed for a weeks afterwards as I kept an eye on it as I had read a lot about different screed failures and was mindful how critical my joins were. It's only since I've taken my eye off it has it moved! 

 

I think that even with some sort of expansion joint the curling would still have happened with the same consequences re heights. If anything, reinforcement between the two floors would have better to prevent differential movement.

 

I've been reading more about it and it seems it may even curl back down (slightly, if not fully) as the lower levels dry out, and it could of course get worse still before it gets better so it may be a little early to be making good with levelling compound.

 

This sound hopefuly from BS8204-1 (Screeds, bases and floorings):

 

 

I obviously want this reversal to happen (we are a long way off covering it) and so will try covering it as I'd read elsewhere. The screeder is not coming until Thursday and so I've got five days to see if any reversal occurs.

 

 

You might be right 

I’ve been laying screed since I was sixteen 

never covered any screed 

Perhaps you can on a tiny bit like yours 

It would be very difficult to cover a 100 mtrs plus laid in a day 

[MJNewton]

With the screeder coming on Thursday and me having taken measurements today I'll see what it's like then and which direction, if any, it's gone in!

[MJNewton]

An update following the screeder's visit today:

 

I was all prepared for the inevitable 'it'll be fine' and 'must be something I've done' lines but I needn't have been so cynical as there was none of that! The screeder was really disappointed by what has happened and whilst he couldn't say exactly what the issue was (possibly a slightly-too-wet mixture causing it to shrink more than it would otherwise done) he is keen to get it sorted.

 

He came up with two options: 1) Cut out the bad corners and re-screed them, or 2) remove the whole lot and redo it. We agreed that option 2 was probably better even if it is more work, although I did offer to assist with the removal. He also said he would use Tarmac's 'Truscreed' screed this time which has an admixture requiring less water and therefore giving faster drying times. This will not only avoid delaying my project (it will actually be dry before the original would've done) but the shorter drying time also means less chance of drying-related issues. Truscreed is also suitable for bonding even at 80mm thickness and so that's what he'll do this time.

 

So, whilst I am very disappointed that we have this problem I can't fault the screeder's response and so this is a net positive in my book. If there was any negative it's that I discovered I paid more than I could've done given he was actually a subby to the firm I contracted and they added on a decent margin. Still, at least they quoted (and delivered) whereas most companies didn't even return my calls!

Edited January 31, 2019 by MJNewton

[Sue B]

That sounds like a good result - well done

[Onoff]

I'd have resin fixed a line of dowls into the edge of the existing slab then PTFE taped the projecting ends before screeding, thus creating a slip joint and helping to minimise the new edge curling.

 

Any mesh in the new screed?

[MJNewton]

From what I've read about Truscreed it seems pretty much impervious to curling with it being bonded and fast drying. It is available with fibres but I am not sure if that's the mix he'll be using (the original screed had them).

[MJNewton]

The screeder came on Saturday to dig the old screed up; I helped him with it partly as I did feel sorry for him having to rip out his own work but also didn't want to let him loose with a breaker (my breaker I hasten to add!) so near to the new doors without supervision!

 

He came again today and put the new screed down:

 

 

It's a modified sand and cement mixture so will be fast(er) drying and therefore less prone to issues. Time will tell!

Edited February 12, 2019 by MJNewton