Squareness tolerances in cabinet making

[YorkshireMartin]

Hello all.

Back again with another question, but hopefully one that is general enough to be of use to other people too.

I'm happy with the accuracy of my combination square. It's a starrett model and having checked it against a couple of different engineers square, i'd say it's pretty much dead on.

I'm wondering what level of squareness is acceptable in cabinet making and how to measure the margin of error. Obviously, we'd all like to get everything absolutely perfect every time, but occasionally things happen and there may be a slight problem affecting squareness, maybe a stray woodchip when thicknessing, things of that nature.

My particular problem relates to the cutting of a 100mmx50mm section of walnut. It was planed square but during use of the track saw to cut it in half lengthways, in the denser grain, I've suffered some blade deflection which has pushed the cut out of square. It's enough to see daylight under the combi square, but I'd say no more than about 0.25mm between the ruler and the surface of the wood at the worst point. The two pieces are at final dimension.

Should I just get on with the job, or scrap it and start over to achieve perfection?

What would a cabinet/fine furniture maker do if faced with this?

Many thanks.

 

[Zeddedhed]

I would say that it all depends where the particular piece of wood is going to end up being used and what relationships/interactions it will have with other pieces.

If its going to be a bearer for a table top for instance then it could probably be used as is without too much of a problem as long as the out of square side is positioned such that it's not mating with other pieces.

If it's due to receive further joinery work such as M&T joints, then I'd probably re-make the part or, if possible square it up and try and re-design for it to be smaller in section (as long as this does not negatively impact the finished design, both from an aesthetic and structural point of view.

I've had occasion to change the size of a table leg for example either due to a machining error or to allow me to use a piece of timber that apart from a defect in the wrong place was otherwise perfectly viable. Unless you are constrained by a specific dimension you may be able to re-plane it and carry on.

 

[YorkshireMartin]

Zedd, many thanks.

It was due to have a dado cut into it and joined to the other leg by a board, creating a full height stretcher for the legs, so I'm worried it would throw the joint out. I could try to use the track saw to square it up relative to the other faces, it's capable of quite thin cuts so it might work I guess.

 

[YorkshireMartin]

Correction, it would throw the joint out, it's just a question of how much is too much I suppose.

 

[Zeddedhed]

Could you not plane it square? Less timber removed that way.

 

[YorkshireMartin]

Could you not plane it square? Less timber removed that way.

I guess I could take a mill off without too much trouble but might need the tracksaw anyway.

I've just noticed that all but one side on ONE leg appears out of square. It's bizarre. The two legs are from the same piece of timber, the remainer of which is still square! I can't work it out at all. #-o

 

[dzj]

I think it shouldn't make a difference.
If the grain pattern permits, perhaps turn the non square side outwards, as 0.25mm is not going to stick out terribly.
I usually refrain from ripping planed stock, as such cuts can lead to warping in the next few days, particularly
if grain density is not equally distributed.

 

[PAC1]

I am struggling with the premise that any piece of timber for "cabinet making" is going to be used straight off the saw? Surely it needs planing at which point 0.25mm is neither here nor there. Two strokes with the hand plane and it is gone. If a machine is involved it does not register. What am I missing?

 

[rafezetter]

(the above is a registered design, no stealing without telling me!)

 

[custard]

I am struggling with the premise that any piece of timber for "cabinet making" is going to be used straight off the saw? Surely it needs planing at which point 0.25mm is neither here nor there. Two strokes with the hand plane and it is gone. If a machine is involved it does not register. What am I missing?

I tend to agree. I often add 0.5mm to my finished thickness dimensions as an allowance for final hand planing/scraping/sanding, so that's what comes off the machine. The exception is actual joints where I normally aim to get a perfect fit straight from the machine. Neither of these are cast iron rules but that's the default position.

The bigger issue I took from the OP's question is about tolerances generally in cabinet making. And they're much, much tighter than many people believe. There's a wide spread view that because wood moves in service and no one cares if a wardrobe is one millimetre over or under in height, then you can be fairly relaxed about accuracy. But that misses the fact that tight glue lines and joint integrity require precision that's almost at engineering levels. Take mortice and tenon joints for example, we aim for a joint that doesn't need a mallet for assembly but will still hold together without glue and resist gravity. If you work in a fine grained and "hard" hardwood like Beech or Maple then the difference between those two positions isn't much more than 0.1mm. Or if you're mitring and want a virtually invisible glue line then you need to be accurate to about a tenth of a degree and about 0.1mm in the length of the components. These are tough targets, but with care and good working practises they're consistently achievable. The key thing is that woodworking tolerances are pragmatic, within reason you don't really care how thick a rail is at the mid point, but you really, really care how true are joint surfaces and reference faces.

One final point, full marks to the OP for getting a Starrett square. Plenty of people skimp on their layout tools in order to spend the money on more glamorous kit, but IMO you're better advised to go about it the other way around.

 

[bugbear]

The bigger issue I took from the OP's question is about tolerances generally in cabinet making. And they're much, much tighter than many people believe. There's a wide spread view that because wood moves in service and no one cares if a wardrobe is one millimetre over or under in height, then you can be fairly relaxed about accuracy. But that misses the fact that tight glue lines and joint integrity require precision that's almost at engineering levels. Take mortice and tenon joints for example, we aim for a joint that doesn't need a mallet for assembly but will still hold together without glue and resist gravity. If you work in a fine grained and "hard" hardwood like Beech or Maple then the difference between those two positions isn't much more than 0.1mm. Or if you're mitring and want a virtually invisible glue line then you need to be accurate to about a tenth of a degree and about 0.1mm in the length of the components. These are tough targets, but with care and good working practises they're consistently achievable. The key thing is that woodworking tolerances are pragmatic, within reason you don't really care how thick a rail is at the mid point, but you really, really care how true are joint surfaces and reference faces.

Or, expanding very slightly, the tolerances required vary greatly from piece to piece, and surface to surface.

One obvious example - when making a picture frame, it doesn't matter if the absolute lengths of the sides are 1mm off, but the lengths of opposite sides must be identical to a much tighter tolerance.

BugBear

 

[Adam9453]

Its an interesting point that most people spend hundreds if not thousands on machines and power tools but baulk at paying £100 for a combination square that is built to exacting tolerances.
I've seen many people try and fail to produce something well purely because they didn't have accurate equipment to lay out, mark out or draw out what they needed to.
Admittedly accuracy to an extent is relative to what you're making, a house is essentially impossible to make to 0.1mm tolerance whereas a jewellery box ideally would be.
Understanding your design and knowing which bits therefore require accuracy makes manufacturing much more straightforward.

 

[YorkshireMartin]

Thanks everyone for the replies. PAC it wasnt used straight off the saw, it had been planed and I'm not sure where the 0.25mm came from as I'd checked for square. It's only one length too, the rest came out fine so I dont think it's the planer. I reckon it might literally be a stray woodchip on the thicknessing bed.

 

[deema]

It would be very unusual for anything being made by a cabinet maker to go from saw to cabinet. For example, most items will be dovetailed, secret dovetailed, blind dovetails. Mortice and tennon or rub glued joints to name a few. In all cases I can think of, the cut end will be left slightly long and then planed back to size. For end grain this would normally be done on a shooting board. For long grain it's normally for a rubbed joint and a fore or shooting plane would be used (no6,7 or 8).

 

[bugbear]

Thanks everyone for the replies. PAC it wasnt used straight off the saw, it had been planed and I'm not sure where the 0.25mm came from as I'd checked for square. It's only one length too, the rest came out fine so I dont think it's the planer. I reckon it might literally be a stray woodchip on the thicknessing bed.

I always remember reading, in the catalogue entry for a very precise toolroom micrometer, a casual reference to "the usual precautions for accurate work", required for the micrometer to deliver its potential.

(IIRC one member of this very forum nearly threw away an expensive and supposedly accurate square that simply wouldn't behave, until he noticed the tiny speck of stray glue on the stock).

BugBear