Tiling Anhydrite screed with wet UFH

3mm grout joints, 1 expansion joint(giving sub 40m2 bays) and a decoupler(thin if sub 6mm tile) will be fine in my opinion. Not an expert on these screeds, as Ajax is, but I thought these gypsum screeds had less movement than traditional screeds?
I'd happily fix decoupler with anyhyfix, switch to cement based on top, incorporate a soft joint(Silicon joint as opposed to unsightly expansion joint) and sleep well at night.

Must admit, I wouldn’t have thought twice either @hmtiling, and with the backing of the adhesive company I use, use cement based adhesive. They have no problem specking their products on anhydride.
In fact just completed a 50m2 floor with 1800 x 900 that was initially a failure due to incorrect prep.
It has expansion joints at the two doorways.
It also has ufh which was initialised in the correct manor and is cranked up fairly high.
It’s a good install and I have complete confidence in it.
Personally I wouldn’t have dreamt of using 5mm joints, 3mm would have been my choice.
I’m certainly not trying to butt heads with you Alan, I’m just saying what I’d of been happy with.
If it is a thin tile, then as Harry says, definitely an anti fracture mat too.
And if water proofing was necessary, I have complete confidence in a waterproof one that I use also.
Obviously expansion joints in this situation are a given, agin tho, I’d of only gone for one.
That would conform to all the relevant information that I possess, so if things in the gypsum screed world have changed, then we as a forum, need updating.

No head butting going on here... I accept that from a technical perspective you could use 1 joint. In ust think 2, evenly spaced, making 5mx5m bays would make it more robust. My real concern though is the 3mm grout joint. I'e just done some calcs based on the coefficient of linear thermal expansion of porcelain, wich is about 4x that of the screed. Each tile based on a change of temperature from 5degrees up to 25degrees (I.e a cold floor up to an operational surface temp) would expand by 0.64mm so about 0.3mm in each direction. That is 10% of the grout joint width at 3mm which will place huge stress on the joint. I don' know the compressive strengths or coefficient of linear compaction for grout so csnnot say if it will accomodate that level of movement. It does worry me though...

I guess, having a slightly paranoid streak I just don' want the floor to fail and the screed to get the blame as it often does...

As regards adhesives I had to take a tile of a alpha hemihydrate the other day . Fixed with cement adhesive . Trust me the adhesive didn't want to leave the substrate .

Totally get the paranoia Alan, non of us want want our product blamed, including workmanship.
Reading your data certainly makes you ponder the possible outcome.
Some thin tile companies and Mapei are promoting 1mm joints, I heard it myself only the other night at a seminar. Granted it was on walls, and I appreciate they’re not subject to same direct heat sources generally, (heating walls is becoming more popular, and they’re being used as a primary heat source) but the tiles will still expand, will they not?

Ajax123 said:

No head butting going on here... I accept that from a technical perspective you could use 1 joint. In ust think 2, evenly spaced, making 5mx5m bays would make it more robust. My real concern though is the 3mm grout joint. I'e just done some calcs based on the coefficient of linear thermal expansion of porcelain, wich is about 4x that of the screed. Each tile based on a change of temperature from 5degrees up to 25degrees (I.e a cold floor up to an operational surface temp) would expand by 0.64mm so about 0.3mm in each direction. That is 10% of the grout joint width at 3mm which will place huge stress on the joint. I don' know the compressive strengths or coefficient of linear compaction for grout so csnnot say if it will accomodate that level of movement. It does worry me though...

I guess, having a slightly paranoid streak I just don' want the floor to fail and the screed to get the blame as it often does...

Click to expand...

Hi Alan,
I've found some info (picture attached) on the compressive strength for Mapei ultra colour grout as an example.
Unfortunately, couldn't find any data on BAL, Weber, Ultra, and Nicobond to compare but pretty sure they all in the same category.
In you opinion....
Would the 3mm joint filled with Mapei accommodate that level of movement.

Ajax123 said:

No head butting going on here... I accept that from a technical perspective you could use 1 joint. In ust think 2, evenly spaced, making 5mx5m bays would make it more robust. My real concern though is the 3mm grout joint. I'e just done some calcs based on the coefficient of linear thermal expansion of porcelain, wich is about 4x that of the screed. Each tile based on a change of temperature from 5degrees up to 25degrees (I.e a cold floor up to an operational surface temp) would expand by 0.64mm so about 0.3mm in each direction. That is 10% of the grout joint width at 3mm which will place huge stress on the joint. I don' know the compressive strengths or coefficient of linear compaction for grout so csnnot say if it will accomodate that level of movement. It does worry me though...

I guess, having a slightly paranoid streak I just don' want the floor to fail and the screed to get the blame as it often does...

Click to expand...

So I’ve had time to sit, read and digest your post now Alan, and these are my thoughts on it.
Most of the time we fit tiles at room temperature, typically around 18 degrees, so as long as the tiles have acclimatised, they too are going to be around that temperature.
So given that the temperature of the tiles is increased to the 25 degrees that you mentioned, as a comfortable underfoot temperature, the actual temperature range is from 18/19 to 25 degrees, in fact only an increase of around 6/7 degrees, how would that effect your calculations?
Because if they do start at around room temperature, and the floor gets colder the tiles will already be in an expansive state, so therefor it would follow that they would contract, making the joint wider effectively.
I appreciate I’m kinda splitting hairs a little but it seems that everything is balanced on a knifes edge these days, so it may well come down to being a little pedantic.
Sod all would get done otherwise.

3_fall said:

So I’ve had time to sit, read and digest your post now Alan, and these are my thoughts on it.
Most of the time we fit tiles at room temperature, typically around 18 degrees, so as long as the tiles have acclimatised, they too are going to be around that temperature.
So given that the temperature of the tiles is increased to the 25 degrees that you mentioned, as a comfortable underfoot temperature, the actual temperature range is from 18/19 to 25 degrees, in fact only an increase of around 6/7 degrees, how would that effect your calculations?
Because if they do start at around room temperature, and the floor gets colder the tiles will already be in an expansive state, so therefor it would follow that they would contract, making the joint wider effectively.
I appreciate I’m kinda splitting hairs a little but it seems that everything is balanced on a knifes edge these days, so it may well come down to being a little pedantic.
Sod all would get done otherwise.

Click to expand...

Really good points raised there but I am not trying to be smart or too pedantic . But most wet floor systems I have worked on have had air stats so to get the room into that comfort zone temperature spread does anybody know what temp the floor slab would be run at .

jcrtiling said:

Really good points raised there but I am not trying to be smart or too pedantic . But most wet floor systems I have worked on have had air stats so to get the room into that comfort zone temperature spread does anybody know what temp the floor slab would be run at .

Click to expand...

Alan stated above that an operational surface temperature of 25 degrees.

3_fall said:

So I’ve had time to sit, read and digest your post now Alan, and these are my thoughts on it.
Most of the time we fit tiles at room temperature, typically around 18 degrees, so as long as the tiles have acclimatised, they too are going to be around that temperature.
So given that the temperature of the tiles is increased to the 25 degrees that you mentioned, as a comfortable underfoot temperature, the actual temperature range is from 18/19 to 25 degrees, in fact only an increase of around 6/7 degrees, how would that effect your calculations?
Because if they do start at around room temperature, and the floor gets colder the tiles will already be in an expansive state, so therefor it would follow that they would contract, making the joint wider effectively.
I appreciate I’m kinda splitting hairs a little but it seems that everything is balanced on a knifes edge these days, so it may well come down to being a little pedantic.
Sod all would get done otherwise.

Click to expand...

I calculated based on an uplift of 20 degrees from 5 to 25 so you are quite correct that a smaller uplift would lead to less expansion. It would be approximately a third so 0.1mm in any direction. I'm not absolutely certain how this translates to a compressive force so need to think on it more.

Ajax123 said:

No head butting going on here... I accept that from a technical perspective you could use 1 joint. In ust think 2, evenly spaced, making 5mx5m bays would make it more robust. My real concern though is the 3mm grout joint. I'e just done some calcs based on the coefficient of linear thermal expansion of porcelain, wich is about 4x that of the screed. Each tile based on a change of temperature from 5degrees up to 25degrees (I.e a cold floor up to an operational surface temp) would expand by 0.64mm so about 0.3mm in each direction. That is 10% of the grout joint width at 3mm which will place huge stress on the joint. I don' know the compressive strengths or coefficient of linear compaction for grout so csnnot say if it will accomodate that level of movement. It does worry me though...

I guess, having a slightly paranoid streak I just don' want the floor to fail and the screed to get the blame as it often does...

Click to expand...

Interesting that i would of thought that porcelain was dimensionally more stable than the screed . Everyday is a skool day

Does a movement joint compensate for tighter grout joints? Is a smaller joint more acceptable with a smaller tile - more joints and a larger joint with large tiles ?

Localtiler said:

Does a movement joint compensate for tighter grout joints? Is a smaller joint more acceptable with a smaller tile - more joints and a larger joint with large tiles ?

Click to expand...

Yes a movement joint would mitigate some of the risk as it would help alleviate stress by allowing free movement. Yes smaller tiles allows thinner joints as the movement per tile is less.

jcrtiling said:

Interesting that i would of thought that porcelain was dimensionally more stable than the screed . Everyday is a skool day

Click to expand...

Porcelain has a coefficient of linear thermal expansion of 4 whereas anhydrite is 0.12 both x10 to -6 m/mK based on the engineering values I have. Sand cement is much higher than anhydrite hence the need for fewer joints in anhydrite screeds...

Thanks again to everybody for their input. I'm glad my post has prompted a debate which I've found really helpful in understanding the issues.

I am hugely relieved that a decoupling layer is not required - this was going to cause me the greatest headache of all and even more costs. I am perfectly relaxed about putting in extra hard or soft expansion joints and realise now that it doesn't matter that these aren't structural joints carried through the screed. The plan is to go for two, dividing the floor into 3 x 5x5m bays, as Ajax suggested.

I am inclined to stick to 3mm grout lines for aesthetic reasons and because the tiles are rectified (10mm thick). I had a long and very helpful chat with a technical guy at Tilemaster who was happy with this provided sufficient expansion joints were provided. I might incorporate additional soft joints if I start having sleepless nights worrying about this.

Ajax, I wonder in the failure case you mentioned whether expansion joints had been provided? and who picked up the tab for the remedial work. I'm guessing it would have been a big number.

In a similar vein, I wonder whether any supplier of materials would provide a watertight guarantee provided they were happy with the specification. Of course, if the job really did go belly up, the first port of call would be the tiler who did the work but I suspect that wouldn't get me very far. In my experience, the customer always end up paying when things go wrong since it's always deemed to be his fault!

Thanks again to one and all

Cheers, Jonathan