Floating or Fully Bonded floor over UFH

[Stones]

In my local builders merchant today looking at flooring, and they had a rather nice engineered board at a good price.  The difficulty is that this particular board has a 'clic' fitting system, making it unsuitable for bonding to our concrete slab.  I have always been under the impression that the best solution with timber flooring over (wet) UFH was to bond it down with a flexible adhesive, however, floating a 'clic' floor in conjunction with the appropriate underlay does appear to be another (and seemingly quite popular) option.  

 

Instinct tells me heat transfer is going to be better if the wooden floor is fully bonded down to the concrete slab, but what in reality is the difference likely to be if we opted for a floating floor?  Poorer heat transfer and slower response time, but as we are planning to run the slab 24/7 at low temp would this really be an issue?  

 

Thoughts?

[ProDave]

I am wiring a new build that has exactly this. So they obviously think it's going to work.

 

Check what underlay they supply, in this case it's a "special" one for UFH that is perforated with lots of small holes. I am there again on Friday so I can see if I can find out what underlay they are using.
 

[Stones]

Thanks Dave,

 

They have proposed an underlay specifically for UFH, and I've no doubt it will work, its really a question of how well it will work, and given the low flow temperatures we usually discuss for new well insulated builds, is it really going to make any difference if the heat transmission isn't quite as good when compared to a bonded down floor.

 

Easier and cheaper to lay, and if it did ever need to come up a lot easier to remove.  

 

  

[Nickfromwales]

In a low energy home with high levels of insulation under the slab, I'd really not worry one bit about heat transfer / losses caused by the lack of contact caused by the underlay. 

The only point I'd raise is if you have this wood in some areas and not in others. The areas where the floating wood is laid may benefit from a very slightly higher flow temp but that may then overheat other areas ( particularly important when all zones are controlled by one stat, eg a single flow temp per zone ). 

Have you zoned the heating ? Is the wood throughout ?  

[Stones]

Good point, last house we rented had a mix of tile and carpet over UFH, and there was a slight variation in how warm the rooms felt, although I do think that wind direction (and consequently wind induced heat loss) had a far greater impact on how rooms felt.

  

The house will have a mix of tile (kitchen, utility, vestibule, bathrooms and a small downstairs space, the rest timber flooring (inc bedrooms).  Bedrooms have been zones but the rest is to be controlled by the master stat.  Bathrooms and utility being slight warmer not an issue for us, and the kitchen is open to the main room, so in reality probably not that great an issue.  

 

The plan is to operate the UFH such that the pump circulates regardless of whether heat is being called for to try and maintain an even a temperature as possible.

[Stones]

I've had a look for some data about how much of a difference there is in respect of heat transfer but can't find anything meaningful. Looking at various timber flooring suppliers, the bonded option seems to be favoured, whilst the UFH suppliers I've looked at seem to recommend a floated floor in preference to bonded.

 

[Stones]

Still not made up my mind on this.  What little I've been able to glean online hasn't really helped.

 

I'm guessing that bonding down a floor is going to take longer than floating a floor.  Anyone like to hazard a guess on exactly how much longer per m2?

[PeterW]

I did 25m2 of bonded bamboo in a little over 6 hours (over 2 nights) - and that was using a 80mm random length board so a pain in the %#*€ ... I reckon I could have done it in laminate in half of that time but the floor feels totally different and it's solid. 

[Stones]

Hmmm, would you say random lengths are best avoided?  I know a floated floor would have some 'spring' in it as opposed to a bonded floor.

[PeterW]

Random lengths are ok if you lay them random ...!! Don't be tempted to use the short ones at the ends, cut long ones and keep the "pattern" going. 

 

The plus side is randoms are easier to handle as they are smaller. 

[ProDave]

Our present (solid Maple) floors were sold as "random lengths". In practice that meant there were 4 different length pieces and a lot more of the shorter ones than any others.

 

The best tip I had was "close your eyes and pick up a piece" then use it, rather than looking at the pile. Even then you had to be careful not to end up with two adjacent joins in line etc.

 

I "added" to the randomness by using whatever off cut was left from one row, to start the next.
 

[Stones]

Is your Maple bonded down or floating?

[ProDave]

1 hour ago, Stones said:

Is your Maple bonded down or floating?

It's a structural t&g board fitted straight across the joists, secret nailed.
 

[Stones]

I'm really struggling to come to a decision on which way to go.  Cost wise, it doesn't look like there will be much in it (based on bonded installation taking longer than a floating installation), so it's really a case of what are the other advantages / disadvantages of both methods?

 

We've covered heat transfer, and accepting there may be some minor difference but nothing significant, the only other two things I can think of are a springy feel if we float the floor as opposed to a solid feel if bonded, and secondly, that cooling the slab (if we found we had to) might be a little more problematic if we went with a floating floor - condensation issue between board and slab?

[joe90]

Personally I dislike floating floors and the possibility of them sounding hollow or bouncing. In theory a fully bonded floor must be able to transfer heat, surely????

[Barney12]

I completely agree with @joe90 

Floating floors sound hollow, bonded sound solid.

Bonded are also less noisy under foot.

[pauldoc]

Hi Stones, I was just reading back through this thread and wondered what decision you came to? I currently have the same dilemma 

[Stones]

I went with fully bonded onto the concrete slab, and a floating floor in our one upstairs space. Have to say comparing the two, far prefer the solid underfoot feel of the bonded floor.  

 

Interestingly, there was no real difference in the length of time it took to lay, as all my flooring was T&G. It may well have been quicker (but would have carried a higher material cost) to lay a floating click system, but we wanted random lengths rather than long fixed lengths.

[pauldoc]

7 hours ago, Stones said:

I went with fully bonded onto the concrete slab, and a floating floor in our one upstairs space. Have to say comparing the two, far prefer the solid underfoot feel of the bonded floor.  

 

Interestingly, there was no real difference in the length of time it took to lay, as all my flooring was T&G. It may well have been quicker (but would have carried a higher material cost) to lay a floating click system, but we wanted random lengths rather than long fixed lengths.

Thanks  for your reply. Can I just check a couple more details pls. 

So if bonded it has to be T&G not click? Or can you use either? Most of what I have seen seem to be click.

Also did you use an underlay with the bonded? if so what kind

and lastly I'm assuming the screed has to be completely cured first. One day per mm?

 

[Crofter]

I'll have to decide on this myself at some point.

My floor is 16ft span suspended timber and has a bit of bounce in it already- probably inevitable unless I'd used much deeper joists.

At the moment I'm thinking bonded floor (bamboo) with a flexible adhesive, incorporating electric UFH mats. Sounds expensive, though...

[Tennentslager]

My 22mm reclaimed engineered oak (T&G) was nailed using the specialist brads and tool hired from HSS.

No glue or lining, fixed directly to the chipboard sub floor.

Fast and easy and it's rock solid.

Oh, and that's how it was done previously when it was a floor in a dance studio fitted professionally. I had to cut off the old brads before fitting and there was no glue on the back?

Edited February 2, 2017 by Tennentslager

[Barney12]

1 minute ago, Tennentslager said:

My 22mm reclaimed engineered oak (T&G) was nailed using the specialist brads and tool hired from HSS.

No glue or lining, fixed directly to the chipboard sub floor.

Fast and easy and it's rock solid.

 

Another alternative to brads and hiring a machine is to use tongue-tite screws which have a tiny head which is lost in the tongue. We used them on the engineered oak floor in the flat and they are absolutely brilliant as they push the tongue and groove together and pull the plank to the floor. Of course you can also easily unscrew a plank if something goes wrong. 

 

http://www.screwfix.com/p/tongue-tite-screws-3-5-x-45mm-200-pack/85991?kpid=85991&cm_mmc=Google-_-Product Listing Ads-_-Sales Tracking-_-sales tracking url&gclid=CMih87WJ8dECFe287Qod388P5g

[AliG]

I too am at this stage, just about to order up the flooring.

 

Most wooden floors I walk on in new apartments have a slight spring to them which I assume means they are floating over a small thickness of foam insulation. I quite like the way this feels, maybe it is just personal preference.

 

I spoke to a flooring company and they said they would stick the floor down and I spoke to the builder and he said that he would float it.

 

I think that people historically preferred to bond it due to concern about lack of heat transfer but as has been said I wouldn't;t worry about that in a well insulated house with low flow temperatures, so maybe it is just personal preference now.

 

 

[Crofter]

The advice thay came with my bamboo was to bond down over large areas (>6m or >35m2).

I wanted to go for T&G as I knew I'd be laying it on my own, and wrestling click planks together is not fun without extra pairs of hands. Plus T&G gives you ever so slight wiggle room.

I've not seen any official advice on the matter, but my feeling is that the simplistic "leave a 10mm gap for expansion" plays a part in this. Surely any flooring will expand/contract as a set proportion of its size, not as an absolute measurement. So if you have a 1m wide corridor, the flooring may be fine with much less than 10mm expansion gap, whereas if you have a 10m wide kitchen, you may need a much greater gap. Bond it down and it can't go anywhere, regardless of the size of the room. Although they still say to leave a 10mm gap, which is weird really.

The main downside may be that replacement of damaged boards could be tricky when it's bonded down- although @Stones managed to replace individual pieces where bonding had failed, IIRC.

[Stones]

I'm still not sure I like the feel of a fully bonded floor.  However, what I do not miss is the normal gap between skirting board and floor that you get on a floating floor where heavy furniture is placed against the wall.

 

Replacing boards (see my blog) was quite invasive if using the plunge saw method.  We have recently tried a different approach as we have two or three other boards that need some remedial work.  I should stress this is not an adhesive issue, but a level slab issue - the boards didn't bond in certain places as they were pulled up from the slab when fitting subsequent boards due to minor undulations.  You really do need a level slab to bond without issue. We recently tried the following product 

 

http://www.espares.co.uk/product/es1708231/fix-a-floor-repair-adhesive

 

with great success.  Clean and easy to do, and A LOT quicker than the lunge saw method.  In the test area we tried, after 24 hours the floor board in question, previously hollow sounded is now rock solid and sounds like all the other fully bonded boards.  For the test area a hole was drilled and a plug (cut from spare flooring) was fitted.  We have a couple more boards to do so will try simply drilling a 1/8" hole, injecting the adhesive then use wood filler to fill the drill hole.  I should be able to report back in due course as to which repair method looks best.