Thread: Wood Movement

Originally Posted by John Samuel

The grain of the runners runs across the grain of the panel and I glue them to the panel.

What panel are you referring to John? The drawer bottom? A pic would be handy.

I assume you're gluing the drawer support to the sides of the cabinet. I've seen them break hide glue & come away when done like this.
I route a dado in the side and glue the front of the drawer support into it, leaving the rest of it free to move.

John, I hope you do not feel that we are nagging here. I think you are taking your chance sometimes. Whenever you make a panel, you must allow it to move. In your latest project you glue 2 very large Red Cedar panels to a frame. It gives the panels no room to move. This is just too risky.

The way I would do it is glue the front end of the frame to the panels, and the rest held together by sliding dovetails .
cedarpanel.JPG

Originally Posted by Wongo

John, I hope you do not feel that we are nagging here. I think you are taking your chance sometimes. Whenever you make a panel, you must allow it to move. In your latest project you glue 2 very large Red Cedar panels to a frame. It gives the panels no room to move. This is just too risky.

The way I would do it is glue the front end of the frame to the panels, and the rest held together by sliding dovetails .

Awww ... Scott ...

It never occurred to me that you were nagging. You were being helpful. People like you have advanced my understanding considerably. My problem is not that I am taking my chances. It is ignorance.

As I said earlier, you and others like Ian straightened me out with movement in doors, but at the time I did not make the connection with the frames of cabinets. In retrospect, it is bleedingly obvious, but when I waltz into the shop I tend to turn off the brain and do rather than think. It's a big reason why I have a shop ... relax ... feel the timber ... do something with my hands rather than with my head ... but this "switching off" has led to me missing the nature and importance of timber movement.

When I first got the tools I turned a couple of old cabinets over to see how they were made and pretty much replicated the construction methods I saw. Either I missed something fundamental or I copied a poor cabinetmaker, or both. To make matters worse, I lifted a construction method from a book. Over the weekend I took a closer look at the book. In the example in the book the cabinet was constructed from ply with edge banding ... not with solid timber. Sharp intake of breath!

At any rate, I seemed to have been very lucky ... and by now it is obvious I need all the dumb luck I can get. Because most of my cabinets have been made with low movement timbers such as NG Rosewood and cedar, I have not had a catastrophic failure ... so far ... no guarantees for the future. The only issues I have had has been with doors and a tabletop in Vic Ash.

So, what does the future hold? It seems to me that:

1. If I make cabinet side panels the way I make some doors, -2 mm veneers over MDF, that fastening frames to these panels ought not be an issue.
2. I could make the sides of cabinets with the frame and panel method. Then fastening frames (top and bottom) ought not be an issue and drawer runners could be made sturdy enough so they only needed to be fastened to the stiles, front and back (like the rails).
3. If I use solid timber panels for the sides of a cabinet (now less likely to happen ... too much potential for problems), I could use sliding dovetails for the frames and fit the drawer runners into dado slots, glued at the front and screwed elsewhere into oversize holes/slots.

Does it sound as if I am finally ... at last ... on the right track? Any other comments/suggestions?

Finally, below is a day bed made recently. I now see that the side and back panels are unable to expand where they are glued to the legs. I could, in the future, make the side panels (say) three separate pieces of timber with space between them, or I could make it with the frame and panel method. Any other ideas/suggestions?
IMAG0320[1].jpgIMAG0324.jpgIMAG0311[1].jpg

Thanks

Hi John

Originally Posted by John Samuel

So, what does the future hold? It seems to me that:

1. If I make cabinet side panels the way I make some doors, -2 mm veneers over MDF, that fastening frames to these panels ought not be an issue.
   this is right. The MDF is stable -- the veneers are just for show
   
2. I could make the sides of cabinets with the frame and panel method. Then fastening frames (top and bottom) ought not be an issue and drawer runners could be made sturdy enough so they only needed to be fastened to the stiles, front and back (like the rails).
   Not a good idea. Frame and panel construction for the sides gets a big tick.
   the drawer runners can sit on a web frame where the front to back members are glued into the front rail, but the rear tenons are not glued to teh rear rail so they can slide in and out in their mortices
   
3. If I use solid timber panels for the sides of a cabinet (now less likely to happen ... too much potential for problems), -- no reason you shouldn't use solid timber -- I could use sliding dovetails for the frames and fit the drawer runners into dado slots, glued at the front and screwed elsewhere into oversize holes/slots.
   again the drawer runners can be fixed at the front and attached using an unglued mortice into the rear support

Does it sound as if I am finally ... at last ... on the right track? Any other comments/suggestions?

you're having fun, you're building stuff you like and the family appreciate

smile and enjoy life

Thanks Ian,

Both you and Scott have been very helpful, and I appreciate your guidance.

If I make veneered MDF panels with (say) 45 mm stiles and rails rather than edge banding (which I dislike) is it still OK to dado/biscuit and glue the web frames to this panel?

I have only one problem with the recommendations so far. It pains me not to have the web frames (top, bottom and for drawers) screwed, glued and dadoed or biscuited into place. A little voice in my head (which loves to over-engineer) tells me that, especially for the top and bottom frames, a sliding tenon or dovetail can't possibly be strong enough. I know this voice tells lies, but I am still struggling with it. In addition my skills will need to improve to make the snug fitting dovetails or mortises and tenons this method calls for. Further, my inclination would be to put screws into oversize holes/slots to hold the back of the web frames in place, or am I listening to that voice again?

What I would normally do is to slide the drawers directly on the web frame ... it sounds as if you might be doing something additional ... am I missing something?

Thanks,

Hi John

have a look at this sketch of a web frame

Ian,

Your sketch was enlightening. I had imagined that the web frame would have all joints glued, and that the M&T /SD joints connecting the front of the web frame to the side panels would be glued, but that the M&Ts/SD connecting the rear of the web frame to the side panels would not be glued. It never occurred to be that the rear of the web frame assembly would not be glued.

If it is not my intention to use web frames for some drawer runners to slide on, but to make up drawer runners do you think it OK to cut a dado into a solid timber side panel, to slide a runner into that dado, to glue the front into place but to hold the centre and rear in place with screws in oversize holes/slots in the drawer runners?

One more question. Consider the sketch below which shows a web frame joined to a solid timber panel and legs.
Web frame and panels.jpg

There is a gap between the web frame and the solid timber panel. If I wanted to fill that gap to fasten (say) drawer runners, could I glue onto the panel a strip of timber where the grain was running in the same direction as the panel? That way, I could cut a dado into it to fit the drawer runners (or even the web frames). I am wondering if the possibly different rates of expansion might still cause grief.

Thanks.

Hi John

your sketch is much neater than mine

traditional timber runners, would be attached to the yellow members in your sketch -- where the web frame buts against the panels, the yellow members would be widened to nearly fill the gap.

you wouldn't normally use metal drawer runners with F&P construction, but if you did, the yellow members in your sketch would be replaced with diaphragms (I can't think of a more appropriate term) made of ply or veneer board to which the runners would be attached

Hi John,
In addition to what has been said there needs to be a gap (2 to 3mm or more) at the rear unglued joint in a web frame, imperitave if the sides of the cabinet are solid wood, this then allows the wood to move, when the side expands this gap will grow, conversely when the side shrinks the gap will close.
The size of the gap is dictated by the weather conditions during construction, (raining/dry) choice of timber, cut of timber (radial/tangential) width of panel.


Pete

Originally Posted by pjt

Hi John,
In addition to what has been said there needs to be a gap (2 to 3mm or more) at the rear unglued joint in a web frame, imperitave if the sides of the cabinet are solid wood, this then allows the wood to move, when the side expands this gap will grow, conversely when the side shrinks the gap will close.
The size of the gap is dictated by the weather conditions during construction, (raining/dry) choice of timber, cut of timber (radial/tangential) width of panel.
Pete

Pete,

I hear you. However, if I leave the rear joint in the web frame unglued with a small gap, do I then glue the rear section (blue in the pic) of the web frame to the panel/legs?

It seems to me that I have received two different suggestions ... but perhaps I am not reading things correctly.

APPROACH ONE.
Web Frame 1.jpg

APPROACH TWO
Web Frame 2.jpg

Looking at the pics, my inclination is to use Option One under most circumstances, but I am here to learn, so comments are welcome.

One more question ... If I make veneered MDF panels with (say) 45 mm stiles and rails rather than edge banding (which I dislike) is it OK to dado/biscuit and glue the web frames to this panel and glue up all the web frame joints; or do I still need to allow for movement?

Below is a day bed made recently. I now see that the 325 mm deep side and back panels are unable to expand where they are glued to the legs. I could, in the future, make the side panels (say) three separate pieces of timber with space between them, or I could make it with the frame and panel method. Any other ideas/suggestions?


Thanks

Originally Posted by John Samuel

Pete,

I hear you. However, if I leave the rear joint in the web frame unglued with a small gap, do I then glue the rear section (blue in the pic) of the web frame to the panel/legs?

It seems to me that I have received two different suggestions ... but perhaps I am not reading things correctly.

APPROACH ONE.
Web Frame 1.jpg

APPROACH TWO
Web Frame 2.jpg

Looking at the pics, my inclination is to use Option One under most circumstances, but I am here to learn, so comments are welcome.

One more question ... If I make veneered MDF panels with (say) 45 mm stiles and rails rather than edge banding (which I dislike) is it OK to dado/biscuit and glue the web frames to this panel and glue up all the web frame joints; or do I still need to allow for movement?

Both methods would work equally well, they both allow movement, a maker might have a preference for one over the other because that is the way it has always done it or machinery is suited to one or the other, I do like ONE because movement is somewhat hidden inside the cabinet, the same as I don't do breadboard ends on tables because movement can become visisble.

With the veneered mdf panels essentially there's no movement so you should be able to glue it all, it then raises the question whether you actually need to glue it all, I like to design such that draw runners are replaceable so I screw these where possible, with a cabinet with sides that don't move the web frame construction is somewhat redundant, as long as the unit is constructed with sufficient stiffening elements (rails/cleats) top and bottom and rails at the front (for draw divisions) all these joints M&T/dowel/bicuit can be glued to the side panel, the draw runners can just be screwed to the sides for ease of replacement down the track, you as the maker though might just choose to glue them on as well.

Again with the mdf method of constuction the carcase can just be made up as a big box with a face frame which then provides the stiffening, it also then depends a little on what you plan to do with the back and whether it provides any strength to the unit or just floats.


Pete

Originally Posted by John Samuel

Below is a day bed made recently. I now see that the 325 mm deep side and back panels are unable to expand where they are glued to the legs. I could, in the future, make the side panels (say) three separate pieces of timber with space between them, or I could make it with the frame and panel method. Any other ideas/suggestions?


Thanks

Where I think the greater issue here is if the panel wants to shrink rather than the expand direction (not too down play the expansion) I would think it may split or a joint will fail, whether it does or not it will be a function of the rate of timber movement species dependant and the cut of the timber, how well the joint was made, type of joint, mechanical locking and glue or a loose fit and lets hope the glue holds (not that I'm casting doubts here John)

To do this in individual boards with a gaps may detract from the overall look, maybe a shiplap or similliar may be ok as this hides the gap kinda, if that design element (shiplap join) was continued as a theme elsewhere, even if it wasn't an actual shiplap but looked like one, but that's for the maker to decide. For this width in solid it lends itself to a sliding dovetail the full width (minus a top and btm shoulder) and for ease of assembly a tapered sliding dovetail where only the top 100mm of joint is glued, then an infill piece fills the open joint below where the board finishes, allow for a gap to the shoulder tho.


Pete