Questions Answered – Why Do All My Doors Come Out Twisted?

Question:

HI

I am an amatuer woodworker building some cabinet doors, but I am getting them twisted all the time.

Can you please advice on how to avoid this and any tools and techniques I can use to help it.

Many Thanks

Answer:

Why Do Doors Come Out Twisted? Great question.

For this there is no short answer so here we go. Twist in doors is not limited to a particular size of door so any door size can be built with or without twist.

I am often asked questions surrounding why twists occur in newly made doors and especially how to remove a twist from a door if that’s the case. It is a frustrating thing to happen because it does take quite a bit of work to make a door in the first place and whether you use machines or hand methods, anything twisted is always disappointing at best. The reasons for twists in doors are manyfold. 90% of twists usually revolve around human failure in some great or small measure so no point mincing words here. Mostly the flawed work can be avoided if you follow patterns of construction carefully. Here are some primary causes. You can mix and match whichever way you want because often there is more than one reason to a twisted door. Let me start by saying there is only one joint to make doors with and that’s the mortise and tenon joint. Often the joint or joints used for door making will be one or more of the common three, which are through-mortise and tenon, haunched mortise and tenon and blind or stopped mortise and tenon. Generally the same flawed work results in any and all three joints.

1. Poor stock preparation resulting in twisted rails, stiles or both.

2. Misaligned tenon cutting which is out of parallel to the long axis of the rail.

3. Misaligned mortise hole cutting which is out of alignment to long axis of the stile.

4. Incorrect shoulder alignment.

5. Out of square edges to stock resulting in pressure exerted during glue up which pulls the door into twist.

6. Over cramping from one side result in in compression of fibres that remains after the glue has set and cured.

7. Insufficiently dry material resulting in further movement in stock after completion of the work.

8. Clamps placed and or clamped out of wind.

9.Shoulder lines are unequally spaced on intermediate cross rails.

There are others but we must stop somewhere. Here are the possible and probable results of the flaws above:

The results of number 1 above may not be obvious at first thought but imagine if even one of the rails or stiles is even marginally out of wind. Though any and all other in-wind rails and stiles would indeed counter the one recalcitrant member to some degree, it’s unlikely that any such counteracting would or could fully correct it. Any twisted component is unacceptable in joinery in general and especially is this son in doors where only one side is fixed and the other is free. Inevitably this means the door will not align properly to the frame or main body or carcass of work.

At numbers 2 and 3 above I think that any misalignment of mortise hole or tenon will be the number one cause of twist in doors unless by some unlikely fluke an error occurs in both simultaneously and one counters the flaw in the other. Not a likely scenario but nice thought. Without using certain patterns of construction most woodworkers using hand or machine methods fail to recognise the essentiality of working dead parallel to the outer face of the rails and stiles. A half a degree discrepancy over so short an area of a tenon cheek or inside a mortise hole’s flat face will indeed telegraph through the whole and will for certain extend into irreconcilable twist.

If at number 4 you have misaligned the shoulders of the tenon this inevitably results in unequal pressure being exerted by subsequent clamping to each opposite shoulder of the same tenon. This causes tension in the very joint area itself both during and after glue up which then in turn pulls the door into twist. It may not be apparent, but shoulder accuracy in cut matters very much. This occurs all the more if the door has rebates (rabbets) but regardless of whether square edged all the way across or rabbeted, when a joint is pressed home both shoulders should seat precisely at the same time without the need of clamping pressure to close one because of the other.

When the edges of the door stiles are out of square either partially or wholly along the stiles as at number 5, subsequent clamping can pull the door into twist and the glue, when set, holds the whole door in twist. You should remember that when using glues such as PVA and others, once the glue has set, there is rarely any release of pressure inside the joint in the sense of bringing the joint to rest. True and well made joinery always results in harmony between the components, hence the name joinery, from the word harmos, from which we get the word harmony, means harmony.

Number 6 is not so uncommon and especially when using woods that readily compress such as oak, some of the softwoods and some hardwoods that have more compressible (spongy) grain. Even with the best clamps known for clamp heads clamping at dead 90-degrees there is some effort resulting in the tightest point being at the inside corner of the clamp head in relation to the bar. This then results in greater finer compression on the inside face or aspect of the clamps than on the outside face. This is one reason why, when clamping thicker material, say over 1” or so thick, that we place opposing clamps on the opposite faces to equalise the pressure too. This is especially so on thick table tops, the leg frames of benches, thicker doors and so on.

Number 7 is not so uncommon at all these days so choose a good supplier you know well if possible. On important work take along a moisture meter. and shoot for wood with a M/C somewhere between 7-12% or buy and be prepared to acclimate your purchase in an atmosphere conducive to the work or the environ it will finally be installed in if at all possible. Often woods that are dried rehydrate to some degree (even a small one) and distort after being into a project even after being sold. Often what happens is the wood is bought with the higher moisture content and is then machined straight away with moisture still in the wood. That wood then shrinks and distorts even marginally unequally when housed or installed in a drier climate whether a workshop or final destination as a finished product. Such climate changes inevitably lead to some measure of distortion in the wood.

I often see number 8 during glue ups and also see that combined with over tightening to close up the resultant gaps at shoulder lines. Make sure the clamps don’t slip off the bench top or one clamp isn’t sitting on top of something causing it to put wind in the project being glued. This a common occurrence in our anxiety to get the project glued up efficiently. Make sure also that after glue up you don’t move the work to somewhere that causes twist in the work too. Undue stress before the glue sets can ‘break’ the glued surfaces between joint aspects and result in permanent distortions such as twist.

Last but not least number 9. If one or two intermediate rails are longer or shorter for some reason this can cause one or both stiles to bow and bend unevenly and cause twist. All shoulder lines should dead on the same distance apart and without compromise. Finally, one thing I always do is check the door for wind before gluing up b standing the door on end placing only one corner on something solid and lowering the other side down to rest on the opposite stile. I then site one stile with the other to see if at all there is any twist in the door. If not then I am confident that the door is well made without twist. I then clamp the door so that all shoulders are tight but not overly pressured by the clamps and do the same. If no twist occurs then I feel confident to glue up. If twist has occurred at an unacceptable level, I consider each of the 9 points above before any glue hits the work and try to correct the problem by working on any flawed workmanship or replacing a part. Once a door is glued up in twist it is unlikely that the twist can be corrected by counter-twisting or indeed by any other action.

There is another post bottom left if you care to look.