Level dips in screed/UFH between joists

Discussion in 'Subfloor Preparation' started by pollen, Jun 28, 2021.

  1. pollen

    pollen Member

    9
    1
    3
    Hello all

    1850s victorian house. Dining room c35m2 has unknown UFH system screeded inbetween joists of suspended wood floor. Not sure about DPM (suspect not) but looks like 40mm screed on top of 60mm foil backed insulation board. Also no idea about relative humidity of base. We think it was installed around 8 years ago. We want to put down low tog underlay and 14mm oak engineered boards.

    There is an area of c12m2 which has a dip/slope to the wall, focused in the area where the pipes run. Dips are between 2mm and 18mm deep. Some other minor patching required to level up to joists in other areas. Photos below.

    Not sure on the best way to level or what products to use.

    All thoughts welcome!

    floor2.jpeg floor1.jpeg floor3.jpeg floor4.jpeg
     
  2. LKAB Gypsol Alan Jackson

    LKAB Gypsol Alan Jackson Well-Known Member

    64
    6
    8
    Thats one horrible looking floor screed. Not even sure where id begin but probably overlay the lot with ply mechanically fixed to the joists. Then smoothing compound to level the ply... not ideal though. Deflection and bounce would be my main concern.
     
  3. merit

    merit Well-Known Member

    7,969
    1,639
    113
    I would flood it with Ardex NA. Then put down a Triton dpm sheet then underlay and float the engineered wood. You need to make sure the wood flooring can’t be heated above 27 degrees or it can crack


    Sent from my iPhone using Tapatalk
     
  4. pollen

    pollen Member

    9
    1
    3
    Thanks both, much appreciated. Given the general state, I think it might be best to bite the bullet and cover the lot in compound. I'm hoping this will neutralise any potential bounce/deflection.

    I'm not going to attempt this myself, but so I am clear on how to probe professionals, can I ask whether ther there are clear benefits of latex vs. fibre reinforced (eg. ARDEX K 40 HB) for my use case?

    As I understand it, the NA needs aggregate added for areas deeper than c8mm, which given the size of the room, will further add to the costs. But a potential benefit of NA is the timber/screed base may not need priming. Is this right? Anything else to consider?

    On the 27 degrees point, there is a standard (non digital) thermostat in the room. I assume I need to get a digital one linked to an external floor temp sensor to ensure that it's never exceeded, where the sensor needs to be embedded into the smoothing compound?
     
  5. dazlight

    dazlight Super Moderator

    6,839
    1,615
    113
    Forget NA for this as the cost per bag will be very high. Ardex K40 HB will be better to use and flows better.
     
  6. merit

    merit Well-Known Member

    7,969
    1,639
    113
    Tbf I would use proflow instead of NA as Daz says it’s too expensive when it’s going that deep. The k40 water based moisture tolerant? I know ultra renovation is. That would work too


    Sent from my iPhone using Tapatalk
     
  7. LKAB Gypsol Alan Jackson

    LKAB Gypsol Alan Jackson Well-Known Member

    64
    6
    8
     
  8. LKAB Gypsol Alan Jackson

    LKAB Gypsol Alan Jackson Well-Known Member

    64
    6
    8
    Sorry...i missed my breply... If its bouncing now adding more weight wont remove it. Itll change slightly the bounce wave frequency but itll still bounce. To stop it bouncing yoj either need support from underneath or composite lateral support in the form of mechanically fixed cross members (noggins or boards)
     
  9. pollen

    pollen Member

    9
    1
    3
    Thanks Alan. It's not bouncing now - the screed seems solid between the stable joists (although I have no idea how solidly or otherwise the insulation board was fixed underneath).

    The bounce I was referencing would be from dips in different areas when we float the (wide and long) floorboards. Hence why now looking at a maximalist approach to level everything up, and bind together those odd bits where the pipes are cut in across the joists but not totally covered.
     
  10. pollen

    pollen Member

    9
    1
    3
    Hello all. Back again. We were recommended that liquid screed (with depths of 15mm to 55m) could level the lot, unbonded ie. on a pvc membrane. We can live with raising the floor level. However, at the last minute the screed supplier has changed their advice and had said that it is likely that with the subfloor being what it is, the screed will likely crack due to movement.

    As previously stated, the joists are stable and have no movement. The only possible movement is if we gently jump up and down in very specific areas where the screed has cracked. And even here this is minimal (1-2mm at most).

    Given that this is so localised, and if we assume that we will not go anywhere near the floor until it has completely dried/cured, am I right to think that the glued floating floor (with underlay) will act as a bridge across the joists and spread load evenly when walked on, protecting the screed.

    Or am I missing something critical which means we should abort and go for compound?
     
  11. LKAB Gypsol Alan Jackson

    LKAB Gypsol Alan Jackson Well-Known Member

    64
    6
    8
    Which liquid screed were you looking at? If its anhydrite its significantly more flexible than cement and 15mm unbonded is fine provided its the right mix design ie TS15 at that depth. Has the screed supplier said it WILL crack or it MIGHT crack. What depth joist are they at what centres and what spans... it's an unusual one certainly but if there is no movement thats the key. Its then about engineering. You need to be aware than any screed is at risk if cracking regardless of precautions. Cracking is not the same as failing and us pretty common. It's more a case of can you live with the cracks. Normally cracks are not an issue with bonded timber...
     
  12. pollen

    pollen Member

    9
    1
    3
    Thanks again Alan for the speedy response.

    Anhydrite. Yes it's a TS 15 mix. And 'might' not 'will'.

    On joists, they are all 10cm x 5.5ish (so roughly 4x2 inch), not evenly spaced but maximum centre 43cm, span gap 37.5cm. A couple have been strengthened by sistering, which given how new they look from above, assumed to be when the UFH was put in.

    Thanks for the key distinction between failing and cracking. Given that the cracks will be under the bonded timber floor, I'm just worried that small unseen cracks may lead to bigger ones and then crumbling/failing...
     
  13. LKAB Gypsol Alan Jackson

    LKAB Gypsol Alan Jackson Well-Known Member

    64
    6
    8
    At 430mm max centres 4x2 joists it would depend on how long they are. Assuming c24 grade, standard loading tables would allow a maximum unsupported length of 2.1m which means basically not very long. Sistering helps but not much with solid timber. I'd be more worried about shear strength than bounce especially given some of the joists are notched and probably penetrated. Based on a nominal depth of say 35mm you have a loading of 70kg per m2. This is pretty heavy for that grade of joist. For that weight I'd expect a minimum 6x3 solid ir 220mm I joist at max 450mm centres. For me it doesnt sound like a good idea to put a screed of any sort over this system I'm afraid. I make and specify Gypsol screeds and having been involved in developing ts15 I'm not worried about the screed itself but I am worried about the structure. I could send you the test data and you could ask an engineer to look it over but really I think you'd be better off overlaying with timber mechanically fixed to the joists.
     
  14. pollen

    pollen Member

    9
    1
    3
    Ah. I'd not considered overall weight loading. The sleeper honeycomb walls in other rooms are 1.2m centre spaced supporting the length of the joists, so we have assumed this to be the case here. Does this make a difference? Very happy to receive the test data too. But also recognise that this may not be the easier option after all!
     
  15. LKAB Gypsol Alan Jackson

    LKAB Gypsol Alan Jackson Well-Known Member

    64
    6
    8
    Yes that supporting distance makes a huge difference. I'll do some calls and come back to you. I'll have to make a few assumptions
     
  16. LKAB Gypsol Alan Jackson

    LKAB Gypsol Alan Jackson Well-Known Member

    64
    6
    8
    Right based on those joist spacings and dimensions the joists should easily accept 1.5kN/m2. The 40mm between the joists is about 0.4 so you have around 1.1 left. 55mm of TS15 will impart around 1.1 or just a bit less. You are therefore at the edge of the joists capability. This assumes the timber is c12 rated or better and is based on 450mm centres. It also assumes no sistering.

    I reckon you would get away with applying the TS15 personally. If you lay it on a polythene separating menu should be fine. Yes you may get some shrinkage cracking or a bit if creep but as itsca wood floor going on top should be fine. I wouldn't tile it though :)
     
  17. pollen

    pollen Member

    9
    1
    3
    That's great news. Most centres are c370mm so I think we have a bit to spare.

    Huge thanks again for your expert advice. Above and beyond. Looking forward to seeing what TS15 can do
     
  18. pollen

    pollen Member

    9
    1
    3
    And def no tiles :)
     
  19. pollen

    pollen Member

    9
    1
    3
    That's great news. Most centres are c370mm so I think we have a bit to spare.

    Huge thanks again for your expert advice. Above and beyond. Looking forward to seeing what TS15 can do
     
  20. mjfl

    mjfl Well-Known Member

    6,258
    1,593
    113
    Alan, what's the drying time of this TS-15 and how soon can this be ground off without more laitance occurring?
     

Share This Page