I'm putting in a radiant floor whereas the aluminum plates (sorry - the cheap ones!) sit on top of sleepers above the subfloor.  What is the best practice for putting floors over this? 

1) I'm interested in putting 3/4" wood flooring over the plates.  Do I use rosin paper (or felt paper) between plates and wood floor?  Or, do I just staple floor down and skip any paper?

2) What about tile?  I'm guessing a coat of mortar (or thinset) right over exposed plates and pex, 1/4 Hardibacker, thinset again, and then tile.  Sound right?  What about these Schluter Ditra stuff also?

I'm asking it here because  of the unique install do to the aluminum plates.

I'm DIY, so not a professional with a bunch of experience in different floors, but I have exactly the same heating system. I also used the cheap plates, my calculations indicated that I didn't need more, and good engineering includes not wasting money. For me minimizing the height buildup to improve efficiency and no fasteners to minimize risk to the heating system was of importance. So I went with an engineered floating floor (I used bamboo, got it on sale) with a thin, high density pad. I will make the observation that is doesn't feel like a stapled down 3/4" wood floor, it has some give when you walk on it, but has been working well, and after 3 years still looks new. If you are going to go staple down I'd probably go without the paper, my research suggested that there is not a lot of reason to use it anymore anyway, at least that way you can see where you are putting the staple, and maybe glue down those last couple of rows that you can't get the gun into rather than nail them down.

As far as tile I used the Ditra, again the no fasteners was important to me. Cement board when properly installed is screwed down as well as thinset. With the Ditra I just followed the instructions, with the modified thinset below, and unmodified above, and again, in 3 years there has been no issues. The only thing I did out of the ordinary was to use a little extra thinset below, and force it into the plates and tubes to improve conductivity.

As far as heat transfer I did experience one cold snap of 18 degrees F below design temperature (living on the high plains can be wonderful), which did hold the heating system on full time, and the floor was able to deliver all of it, which means I was transferring about 14 Btu/(ft^2-hr) using 95 degree F water, so pretty good.

Thanks dave111 !

I'll be asking a couple more questions in the near future and looking forward to your responses. Looks like we are on same page when it came to plates and efficiency vs cost!

Well done Dave. The only thing missing is a proper heat load. Nothing wrong with thin plates in a sandwich radiant floor.

Howdy Badger;
Did get a heat load calculation, Manual J. Had that done before I sized anything. Mind you I figured I'd need a 3 ton, the Manual J came in at 2.5, so I was happy.

Want a job? :-).

Hi all, we are facing the same thing right now with tiling over hydronic heat/plywood sleepers/pex in aluminum transfer plates. We are using marble tiles (a tough kind, does not curl and not very fragile) in one area and slate in another, all relatively small rooms (bathrooms, mudroom). If you look at industry tiling guidelines, Schluter info, etc., they all call for a pretty thick stack-up, in most cases a 1.5-2 inch mudbed with the pex at the bottom, wire mesh over pex, then at least 3/4 mud over that. Then Ditra, then tile. So that's a pretty thick layer. So is Bekotec (I think a little over 2 1/4 inches by the time the tile is in).  Not to mention expense of these methods...

We live in northern Cal, so unlike MN and similar climates, we want a responsive system that we can turn down at night and on warm afternoons, and not have to then worry about taking hours to warm up a huge mudbed to heat the bathroom. So your method above sounds like a good approach, and would also help us with height issues (trying to avoid a big height change going from entries to rest of house). But I'm curious how well stone tile will hold up on that? We've reinforced our subfloor, but it is still a wood subfloor over a crawlspace, so definitely more deflection than on a slab. Were you using the sleeper/thinset/ditra/tile method over a slab subfloor?
Also, how was it to get the Ditra smooth and level? In watching Ditra install videos, it looks like it would be a challenge to get is smooth over the plates and the pex. So we had considered adding 1/4 CBU, but that of course brings its own challenges and downsides....
Thoughts?

JKH, if you only need a part-time heated floor because of a moderate climate or because of low living use factors, you may indeed be better off with a low mass and highly responsive emitter, and perhaps even an electric emitter too (e.g., Schluter-Ditra, SunTouch, Thermosoft, etc). Full slab high mass emitters are more appropriate for a full-time heated floor as they are the most efficient emitters (i.e., they require the lowest hydronic supply temperature to provide a given required heat gain).

Heating system operational cost decreases with increased emitter efficiency. Emitter installation cost can vary significantly and largely depends on whether the project is new construction or a remodel. For example, if you have new construction requiring a full slab floor, there is little additional cost to place PEX prior to pouring the full slab floor. If you have a remodel where you need plates or need to pour a thin slab just to create the emitter, the additional cost can be very significant.

If you elect to do a thin slab/tile emitter, you should carefully adhere to the subfloor stiffness requirements and carefully follow the specific manufacturer pour recommendations. If you need to deviate for some reason, you should consider consulting with an engineer before experimenting. In terms of efficiency, thin slab emitters fall between full slabs and plate systems.

You should always accomplish a building heat loss analysis and a heated floor design analysis to fully understand what will be required and how much it will cost. Lots of bad numbers and bad suggestions get tossed around on forums like this and only proper analysis will determine credibility and what is best for any given project. John Siegenthaler’s Modern Hydronic Heating is the best source for credible information on this subject. You should also consider doing a ROI analysis to fully understand if spending your hard earned dollars in this way makes good sense for you.

Just wanted to bump this, curious to hear about the experience of others who have tiled over hydronic heating/aluminum plates and how that went and how it has held up. Anyone? We have talked to a tiler or two willing to try the above-mentioned methods, but most want to pour a thick mud floor and don't want to use "new-fangled" stuff.

Sailawayrb, thanks for your reply. We would like to heat the house, not just the bathroom floors, and right now it is 39 degrees outside and damp....may not be Minnesota, but definitely could use some heat! We were redoing all the floors down to the subfloor anyway, so it seemed the perfect time for a hydronic system, and we installed it much like the OP.
We do tend to turn the heat way down at night and for 6-8 hours during the day. With properly done thermostats, our system will hopefully figure it out, but at the same time, we'd like to minimize the large amount of slab to heat up, so the system is more responsive. Thus our search into tiling over hydronic heat without pouring a thick slab. Looks like plenty of people have done it, but it's not strictly by the TCNA guidelines, simply because they haven't tested it (as far as we can tell).
We did run a heat loss analysis, though I would say it was a bit rough, since we are on solar (in rural area so only other option was propane). We plan to use an electric boiler (or heat exchanger, researching that now and appreciate any suggestions). So we did not need to be too exact as far as that aspect.
We have referenced John's book a lot, really helpful to have that resource.
As far as ROI...welll....nope, may not be there in a strictly economic sense, but how to calculate the comfort and possible reduction of arthritis and asthma etc. compared to other heating methods? We're looking at it a bit like people look at steam showers or swim pools or sports cars - strictly necessary? Nope. Hydronic heat is a luxury item for us. Not that we want to throw money away, but from a strictly economic standpoint, we have enough firewood around to heat with out woodstove all winter, so that heating method would make the most economic sense. We'll keep that as backup (and because it is actually illegal to install a woodstove or fireplace now in our county..) But still looking forward to warm floors!

We would highly recommend using Schluter Ditra anytime that you finish a HR heated floor with tiles.

Schluter Ditra Engineering Handbook

Yes, one often has to trade economics for other considerations. We would certainly agree that heated floors are the definition of comfort and health. Nevertheless, we always have our clients consider ROI before making these trades just for due diligence.

We don't advocate or get involved with HR plate systems (only HR concrete slabs), so we can't offer any first-hand guidance on how well HR plate systems hold up over time. New HR concrete slab systems should last for centuries. We do sometimes hear from folks who say HR plate systems systems can get somewhat noisy because of plate expansion/contraction during the heating cycles, but we suspect that the installation approach/quality has a significant effect on this issue. There are several companies who advocate and sell HR plate systems who should be able to provide you good guidance/recommendations on how best to finish a HR plate system with tile. These days, you can obtain quality tile that will emulate any finished look that you desire (e.g., stone, wood, etc).

I am not sure what you see as the downside of cement board. I've tiled a couple second floor baths over it with nary a crack a decade later. Deflection isn't a worry. Once the tile bonds to the underlayment, it isn't moving up and down.The purpose of the underlayment is to isolate the tile from movement in the structure. (As an aside, I use polymer fortified Versabond thinset, available at the orange box store. Do NOT allow it dry on your tools.) These floors were not radiant, granted, but it comes down to bonding surface available. With wide pex spacing, narrow spacing and large tile, I see no reason not to use CB in strips for the build up, which would put the plates in direct contact with thinset. For small tile pieces, put quarter inch over the plates. I've never used Ditra, principally because it costs more per foot than half inch hardie backer. Clearly Ditra is lighter, quicker and cleaner. Cut CB outside with eye and lung protection, and position a floor fan to blow the dust away from you.

Make that narrow plates

No problems yet, and it has been 3 years so far. I've got a wood structure under it, although I did increase the stiffness when I did the design, I'm up at 640. As far as leveling it I used a four foot level as a straight edge, working the Ditra over the thinset until it was level. I also used a rubber mallet for those stubborn spots. Overall leveling it was not that hard. I also was concerned as to stack height, I used a porcelain tile (1/4), and with the Ditra (1/8) and thinset (1/8) for a 1/2 inch stackup. Almost a perfect match for the engineered hardwood I used on the balance of the floor. Like you I wanted a quick response on the floor, although my reason was to simplify the design. I wanted the air/house response to dominate, not the floor, which greatly simplified the control system.