can we talk, calcium sulphate screeds

[Ajax123]

Calcium sulphate screedforms a perfect compliment for an underfloor heating system being suitable for use with both wet systems and electric systems. The nominal septh of traditional screed over underfloor heating would be 75mm (yes I know sometimes it is laid thinner) with anhydrite you are able to reduce the depth of the sacreed significantly to just 25mm to meet BS8204 part 7 and 30mm to meet the requirements outlined by most manufacturers. This offers a number of advantages with heated screeds but before we look at them let us look at the standard design requirements and some of the possible pitfalls.

1. It is very important that the screed is divorced from the substrate by means of a Damp Proof Membrane, usually a 1200 gauge polythene over the substrate. It is common to see this folded up the walls behind the edge strip but in my opinion this is un-necessary as there should be a damp proof membrane under the concrete which serves to protect the srtucture. provided the screed does not bridge this then the DPM over the concrete can just be cut almost flush to the walls as it is only there to prevent the residual substrate moisture from rising through the floor and extending the screed drying time. I also see a lot of installations where the DPM has been moved upward on the zone ot on top of the insulation. This is a bad idea if you are using underfloor heating pipe clips to secure the pipes as these will clearly then breach the DPM so it becomes a seive instead.

2. The screed depth should be designed to meet as close to the minimum as possible. I see a lot of designs where the insulation is designed to meet the u-value required and then the screed placed to make up the zone. What would be better would be to design the insulation depth based on the minimum depth of screed (normally 50mm nominal). placing the screed too deeply results in significantly extended drying times.

3. The screed should not be placed at belos the nominal inimum design depth as this can lead to overheating at the surface of the screed and the posibility of pipe mapping and plastic settlement cracking over the pipes.

The screed should be designed with consideration for expansion and contraction. Whilst it is much more dimensionally stable than traditional materials it should still be joined across all door thresholds and where there are abrupt or significant changes to aspect ratio or direction. Joints should also be placed at the interface between independently controlled heating zones, remembering that generally there will be up to 8 curcuits within the same zone. All of these joints should be formed using a full depth isolating former ideall of a compressible material and should be between 5mm an 12mm wide. All joints should be fully reflected through the tiled surface. Remember uncoupling membranes are not designed to replace joints.

Now to look at some of the benefits. The screed has a very high thermal conductivity at around 2.2W/mK. The fact that the screed is fully self compacting means that there are no air pockets trapped within the screed which would lead to a less efficient heat transfer. Additionally the screed flows freely and completely around the heating pipes to offer a full contact surface for the heat transfer. The material is laid thinner so that there is less material to physically heat up. All of leads to greater efficiencies in the heating system meaning effectively that to get the same heat output compared to traditonal materials you need put less energy in. This makes anhydrite perfect for use with geo thermal heating systems as well as other more traditonal systems. It allows trhe screed to act more like a radiator than a storeage heater.

The fact that it can be laid thinner than traditional materials means that the design of the insulation can be considered in a more flexible way. For example if there is no desire or no possibility to alter the depth of the overall floor zone we can look to change the depth of the insulation. This allows the use of potentially more environmentally freindly insulations to meet or exceed the required U Value. When the U-Value is calculated using for example a poly urethane insulation having a r value of 0.023 at say 75mm I can look at changing the insulation type to an EPS100 polystyrene at 100mm with an R aValue of 0.030 and generally exceed the u-value we achieved using the PU. this would offer a cost saving in most instances. Alternatively and possibly more desirably in many smaller installations we can simply increase the depth of the insulation to enhance the u-value significantly and make the floor more environmentally freindly and in the long term cheaper to run.

The alternative of course it to physically reduce the depth of the floor zone although this can generally only be acheived at design stage because once the build gets past ground floor level it is too late. In this instance though reducing the screed depth by 25mm meansd that less material is used in the overall construction. This would often not suit traditional masonry construction where floors are usually designed to fit in with brick coursing. Where it can be done though the reduction in material use leads to lower levels of emboddied energy in the whole construction. I have seen instances particularly where the timber frame systems have been used where an entire course of bricks per floor has been removed with very significant material cost savings to the builder as well. The possible reduction in floor zone is up to 35mm.

Where electrical underfloor heating systems are used it important that the cables are controlled by means of floor stats instead or as well as room stats as there can be a risk of the floor overheating.

Turning the general floor temperature down not only leads to energy savings but also reduces the risks associated with thermal expansion and contraction of the screed and the inpact this has on the finished floor coverings.

The temperature of the underfloor heating system should not exceed 55oC for anhydrite during the comissioning phase and it is generally recomended that this is limited to 45oC during the operation of the floor. This would normally lead to a surface temeprature which is comfortable.

The surface of the screed in operation should feel at best slightly warm to the touch based ont he temerature differential of this (maximum 27oC) and the temperature of the average human hand. If the floor feels hot it is generally too hot.

 

[Ajax123]

Another major benefit of calcium sulphate is its ability to be force dried. The underfloor heating can be comissioned as soon as the screed is more than 7 days old. This means that the underfloor heating can be very effectively used as a means of force drying it. If you were to force dry traditional screed it will cause cracking, crazing and curling. None of these are likley to happen with calcium sulphate provided it has been correctly installed. Of course it is not impossible for cracks to appear and the fact that the underfloor heating can be comissioned very early means that any cracks which do appear can be attended to very quickly. It is imperative that an underfloor heating system is comissioned properly before tiles are laid. The recomendation in the relevent standards is that the underfloor heating in traditional screeds should not be comissioned for at least 28 days following installation of the screed. This means that if you want to use the heating to dry the screed if you use calcium sulphate it will probably be dry before you can even switch on the heating if you use a traditional screed.

Unfortunately it is not recomended that surface damp proof membranes be used on heated screeds but the fact that anhydrite can be force dried means that it is largely unecessary anyway.

The comissioning procedure takes approximately 10 days from start to finish although it can be extended in order to use it to force dry the screed effectively. Bear in mind that just heating the screed will not coause it to dry. It will of course force the moisture into the atmosphere above it. If this moisture is not removed either by extraction, ventilation or dehumidification (Not my personal favorite method) the relative humidity in the atmosphere will quickly equilibrate with the screed and drying will stop. This is true of all substrates.

 

[Ajax123]

Some stuff about moisture


The accepted drying time for sand cement is 1mm per day based on 20oC and 60% RH. so technically a 75mm screed takes 75 days to dry. Modified screeds can dry more quickly. The thing to remember is that cement based adhesive, based on the simple fact it is cement, will tolerate moisture abuse cos cement loves water. The cement simply conitnues to hydrate untill the water is all used up. It is less likely then that a cement based adhesive will delaminate if the screed is "reasonably dry" i.e. typially in my experience 6 weeks but I accept your 30 days if it has worked for you in the past. This is the case even if the adhesive (as most do) says the floor should be below 75% RH. If you do a moisture test on the material it will yeild up to 4% water to be considered dry. This is because some of the water is retained permanently within the cement pore and keeps the cement in a permanently alkaline condition
Calcium sulphate binder is added to sand and water and some additives to make a calcium sulphate screed. There are 2 distinct types but the ultimate result is the production within the screed matrix of Gypsum. Once all the source material i.e the binder is used up any remaining water - called mixing water or workability water - is surplus to chemical requirements. As the water will contain water borne sulphates and it moves around within the screed matrix predominantly in an upward direction if there is cement available for it to react with it will do so forming mainly ettringite and also some thaumesite although the latter is a very small proportion. The nominal dry calcium sulphate screed will have around 0.5% water by weight in it. to get to this level based on the same drying conditions described above will take approximately 1day per mm for the first 40mm of depth and then another 2 days per mm for each mm over the 40mm depth. i.e. a 50mm screed takes around 60days to dry in a good drying environment. Unfortunately the biggest issue we have in the UK is that people do not take this into account in the design stage and it often goes in much too deep. If you use a gypsum based adhesive there is no risk of chemical reaction and so like the cement on cement scenario it will be more tolerant to the effects of moisture. The gypsum based adhesives I know of can generally be laid when the screed is up to 85% RH which is actually closer to 1.5% moisture. i.e. almost 3 times wetter.
In terms of sanding the screed the advent of the now low laitance systems have reduced the need for this in terms of laitance removal because the laitance is bound within the screed matrix just as it is in a sand cement sceed. The down side is that the surface of the screed is likely to be much harder and smoother than the old fashioned stuff. This means that primer do not penetrate the surface as easily and as such it is often necessary to use an orbital sander to prepare it to be primed. ONe thiig I see time after time is overkill on this sanding. As for who does it .... it is generally a contractual thing. The screeders can do it and some do as a routine. Some offer it as an extra over and it is up to the main contractor to arrange it as required. Some say it needs to be sanded bu they do not offer sanding as part of the service. These latter are not my favorites but there is some merit commercially in some projects so there is place in the market. In terms of preparation for floor coverings the screed should be clean and free from surface contaminants such as plaster and mortar snots etc. The best way to remove these is by sanding. The preparation for floor coverings should in my opinion be part of the floor covering contract and as such the responsibility should be dealt with contractually between the main contractor and the flooring contractor. It is difficult to assign this part to the screeder because with the best will in the world he is unlikley to know the level of prep the flooring contractor is likely to require.
The same method of preparation should be carried out on traditional sand cement screeds, concretes and the like so in this respect anhydrite does not differ. The industry has unfortunately become fixated on the requirement for sanding. I think that because there are sometimes issues sticking the floor coverings down they get into their heads that they need to create a very rough surface in the mistaken beleif that this will help. It also provides a very convenient scapegoat for the primer and adhesive manufacturers to blame when there are issue. I have been on many problem sites where the lack of sanding and of course that devils screed anhydrite has been blamed by everyone concerned but on proper analysis and investigation in many cases the floor would have failed regardless of the level of sanding due to other problems with the build. Often the lack of a DPM or some other moisture ingress, in more than one case a leaking pipe under the screed has been a route cause.