Straightening a twisted frame

[MitreKnown]

I've made the internal frame for one of a set of garage workshop doors. The doors are to be well insulated (90mm worth) to keep the garage workshop warm. The frame is going to get covered in plywood (12mm external, 5mm internal) and there is a window at the top. The problem I am having is the frame twisting. I can't tighten the last corner without the frame taking on a twist and I can already tell it wants to twist as it is easier to pick up one bottom corner (the fully closed one) than the other. The 4x2 used for the stiles have an amount of twist in them and I think that's transferring to the whole frame. None of the 4x2s will be perfect, so I can't fix this by just selecting the best. They have at least been dry stored for about 12 months though. Also, any slight out of square on the rail cuts will create twist.

What can I do to combat this twist? I don't imagine the plywood will do much to resist it, but would love to be proved wrong on that thought as it would be a lovely/easy solution.

Note: as viewed the frame is upside down.

 

[deema]

Absolutely nothing. As the seasons change it will keep moving.

 

[MitreKnown]

Absolutely nothing. As the seasons change it will keep moving.

That's bad news. Am I better off making the door more flimsy so that it can be pulled straight by the door frame when closed and it's not an issue when open? Would be a shame to lose insulation thickness, but I guess I could make the stiles and rails from 2x2 timber instead, or perhaps even 25x50 batten

 

[Sachakins]

Could it take 18mm ply each side to pull almost flat first?

 

[MitreKnown]

Could it take 18mm ply each side to pull almost flat first?

That's a good though. Adds quite a bi of weight, but would be stiffer.

 

[Spectric]

That's a good though. Adds quite a bi of weight, but would be stiffer.

Also more secure and better for attaching locks and hasp.

 

[Sachakins]

Did you use the timber as it was, or did you mill it on planer thicknesser first?

 

[Terry - Somerset]

If the two skins are glued and screwed to the frame whilst held flat, would the door act like a torsion box and resist twisting.

I understand this is why most cheaper doors tend to be flat - not the quality of materials specified by B&Q or Wickes etc -but because of the way they are constructed.

Just an idle thought!

 

[Jameshow]

If the two skins are glued and screwed to the frame whilst held flat, would the door act like a torsion box and resist twisting.

I understand this is why most cheaper doors tend to be flat - not the quality of materials specified by B&Q or Wickes etc -but because of the way they are constructed.

Just an idle thought!

Is it really a torsion box?!

 

[Dee J]

If the timbers are basically straight, but with a twist, then you can cut the joins to allow for that, so the twist doesn't transfer to the door.

 

[Terry - Somerset]

Is it really a torsion box?!

According to Wikipedia - I accept not always the best reference:

A torsion box consists of two thin layers of material (skins) on either side of a lightweight core, usually a grid of beams. It is designed to resist torsion under an applied load. A hollow core door is probably the most common example of a torsion box (stressed skin) structure. The principle is to use less material more efficiently. The torsion box uses the properties of its thin surfaces to carry the imposed loads primarily through tension while the close proximity of the enclosed core material compensates for the tendency of the opposite side to buckle under compression.

A door frame + skins on either side glued and screwed are effectively a torsion box. Your design may be improved with a little more internal frame to fix the skins to.

 

[MitreKnown]

A door frame + skins on either side glued and screwed are effectively a torsion box.

I think this could work. Having made the frames, they are pretty heavy. I have hinges rated to 160kg a pair but still.

I am wondering about changing the plan to use 25 x 50 batten without the diagonal and doubled up on the stiles. Then I can insert a rectangle of 50mm thick PIR insulation (Celotex) which could act as the internal support of the torsion box. It would reduce the weight and potentially help vs the twist, but the insulation is expensive.

It looked so easy on paper

 

[MitreKnown]

Did you use the timber as it was, or did you mill it on planer thicknesser first?

No though I guess I should have

 

[alex robinson]

Do you need to tighten it? I am not going to get involved with the torsion box definition, but most of the strength is going to come from the skins. That corner joint will contribute very little. I would argue you don't even need the diagonal in there.

 

[MitreKnown]

Do you need to tighten it?

I guess not, but I'd like to fill it. I'll then have to plane the frame back to square

 

[simonridout]

I have recently made a pair of doors, each using two 11 mm OSB skins. I put 150 X 50 pieces ‘inside the sandwich’ where the hinges were going, a 150 X 50 across the middle, as strength for the lock, a Arbus van lock and then 100 X 50 backing pieces top and bottom for security bolts, before filling in the rest of the perimeter with 50 X 50 before infilling with 50 mm rockwool. They were very nearly flat, I chose the least flat door to fix shut with top and bottom sliding bolts, into the floor and lintel. Then added a ‘storm flap’ on the inside of the fixed door and the outside of the other door. This was sufficient to ensure that both doors were true when shut and secure. Each door was 2100 X 1100 mm. I could only just move them singlehanded. Hinges were galvanised hook and strap gate hinges, straps bolted to the doors and hooks bolted to the door frame. Overall an easy task, which would had been a doodle with a second pair of hands.

 

[MitreKnown]

I have recently made a pair of doors

I think I have been too ambitious going for 4x2 rather than 2x2. I did it because the walls have a filled 100mm cavity so I wanted to match that, but I think 50mm thick door cavity is still very good. I am awaiting a quote for PIR, but it's available for ~£20 per door, so I am tempted to change to that.

• 5mm plywood inner skin
• 50x25 batten doubled up on the jambs. going with that to match the 50mm PIR thickness (will double check batten is correct size, not little under)
• 50mm PIR with L=0.022 equivalent to ~75mm of Rockwool / Ecobeads which was previously my chosen insulation as I have some free
• 12mm plywood outer skin

U-value should be around 0.022/0.050 = 0.4 W/m2.K

Each door was 2100 X 1100 mm

That's pretty much identical to my door size, but I have to make three pairs.

I am including a 400 x 1100 space at the top to put a glazed unit in. That will have a U-value of around 1.2 W/m2K so despite being a small proportion of the door almost half the heat loss will be through that.

 

[Sam Kilby]

It’s all a bit coulda woulda shoulda now but -
Cls and any other builders framing softwood is fairly shocking in terms of its suitability for joinery projects. You can sort through the pile and try to identify non bananas, but I don’t trust it to not turn into bananas.

One way to mitigate this getting worse is to utilise proper joinery at the corners.
Could be as simple as several big (15mm dia) through dowels drilled with auger bits through whilst all clamped where you want it.

Preferably par your own timber effectively to start with. Even if it claims it is already, it isn’t really. If not possible, make oversize then somehow trim back to into square after, but the joints at a least ought to be closed up.

For this kind of thing I would probably start with joinery grade redwood, slightly more expensive but worlds apart in terms of reliability for using in joinery..

If you have a look online you can see all the tricks framers have to do to negate the interesting shapes of this timber. Mainly using wedges or shims, kerf cutting the backs, and a month after install coming in and hand electric planing to remove bellies.

 

[TomGW]

I guess not, but I'd like to fill it. I'll then have to plane the frame back to square

I think that, at 4" + skins thick, the latch side will definitely need to not be square, in order to close snugly in its frame. You should give some serious thought to angling the latch side of both the door and the corresponding side of the door frame to allow it to fit snugly.

 

[MitreKnown]

Cls and any other builders framing softwood is fairly shocking in terms of its suitability for joinery projects. You can sort through the pile and try to identify non bananas, but I don’t trust it to not turn into bananas.

I agree for lots of uses but had hoped I would be OK here as I hand picked and dry stored the timber for >6 months before building. I wanted to use treated timber as its an exterior door. I think I could be facing the same issue with PSE softwood.

I think that, at 4" + skins thick, the latch side will definitely need to not be square

That's correct and I worked out I needed 6mm of gap in the middle just for that, so extra for wiggle. I was planning to add timber to create a form of rebate.