This will be a long answer as it touches on a few different aspects of woodwork. There are two different fundamental approaches to cabinet making. The newer and currently dominant approach, based around accurate machine set-ups to produce interchangeable components, and the older approach, based around the special tricks and techniques for
fitting individual components together. I'll start with the newer method.
Whenever I sell someone a piece of furniture I try and take advantage of having a willing customer with an open cheque book, and sell them something else at the same time. As often as not that "something else" is a large mirror using the same timbers as the original commission. Until recently I had a picture framer's Morso guillotine, the big floor standing version which incorporates a double mitre knife and an accurate length stop.
This machine went a long way to both speeding up and de-skilling the operation. But last year I decided to re-lay my workshop and sold the Morso, so I've had to figure out a way to produce perfect, flawless mitres using my remaining equipment. Here's what I'm talking about, mirror frames from 75-100mm wide components in solid timber
There are several separate elements to first quality mitres, and I'll tackle them individually.
First is cutting a precise 45 degree mitre. How precise is precise? The most accurate angle measuring tool I have is a Mitutoyo Protractor that reliably reads down to five minutes of arc, which is one twelfth of a degree. I've established this is
just at the margins of acceptable accuracy, for zero visible gaps I'd still like it a bit more precise, but I can work with this. It's not far from the 0.1 degree that you mentioned. You can pretty much replicate this precision with a test mitre on two scrap arms, assemble them and check with a backlit square. I see Martin you're using a 300mm Starrett combi, personally I prefer to do this with a 600mm blade fitted, but a 300mm will do. Dropping the 45 degree shoulder section of your combi square onto a cut mitre won't give anything like the same level of precision. The key thing is getting that test mitre tightly fitted and being really rigorous in your examination. If there's a minute gap don't kid yourself it's not there, keep adjusting the saw fence until you've removed it. To put this in context, I find it fairly hard to achieve the required accuracy on a Festool Kapex mitre saw, the adjustment mechanism just isn't fine enough. It can be done, but it's a long job and if the machine gets moved it goes out of the setting. It's a bit quicker on my sliding table panel saw, but it's still at the limits of achievable accuracy, and I'll often resort to shims of masking tape on the fence to make the final adjustments. Surprisingly the quickest route I've found is using an MFT table with dogs and a track saw. The price you pay going this route is setting the lengths.
Which takes us on to the second issue. Getting each pair of opposing arms
precisely the same length. This is harder than it seems, you cannot measure the length by the way, it's impossible to achieve the required consistency from a ruler. The first mitre cut on any arm is easy, but the second cut, when you're referencing from a previously cut mitre, is a pig. The reason is your reference end finishes in a feather edge, this can either disappear in a gap behind the stop, or you can dub over the feather edge if you apply too much pressure. A variance of a few tenths of a mill will result in gappy mitres. The best solution is to cut all the mitres at
one end of your four arms, then make a false stop with a matching 45 degree angle for cutting the mitres in the opposite ends. This is how the length stop on the Morso is designed and it works. The only thing you need to be careful about is blowing out any dust that gathers by the angled stop.
The third issue is jointing. It's not really a problem with sheet goods because they're more stable, but with solid timber it's a real issue. The key thing is this, over time the timber will shrink and swell, and when it does it will try and open a gap at
either the inside or outside edge of your mitres, eventually the glue will start to fail and a tiny gap will open. This gap will quickly fill with household grime and will look very shabby. If your mitre is to look good for decades to come then you need to find a joinery solution that stretches as far across that mitre as possible. It's easy to pop a Domino or a dowel into the
inside corner, where there's plenty of meat on the bone, but the
outside corner is harder. The solution I use most is a biscuit, but a shaped spline would also work well. I aim to get the joinery to within 3mm of that external corner to provide support right the way across.
The fourth issue is cramping during the glue-up. With heavier components (say over 50mm wide) you can
just about get away with a band cramp. But beware, when you tighten a band cramp it's trying to turn your rectangular frame into an oval, in other words it's attempting to open up the
inside edge of the mitres. Many people see that happening and respond by tightening more, when the correct remedy is to slacken off the pressure. Picture framers use an underpinner; as well as driving in framing wedges this machine also grips the joint within the rebate and pulls it tightly together. I was very fortunate in finding a long discontinued Morso product that is a specialist and heavy duty corner cramp. In the absence of such a tool the best alternative is temporarily glued on cramping blocks. Mike references cramping blocks in this current forum thread,
i-hope-this-works-t106982.html
Cramping blocks are a giant faff, but hey, achieving first quality work is just one giant faff after another! So only you can decide if you're up for it or not. Another option is specialist corner cramps (they're not expensive and there's loads on the second hand market) or picture framer's "springs" and then tacking the joint together while the glue sets and filling the holes. One final cramping tip, the mitre joint is essentially end grain to end grain. If you're using PVA it helps to apply a first coat of slightly thinned PVA, leave that for a couple of minutes to soak into the end grain, then apply a full strength coat and cramp it up. You'll get a much stronger joint this way, but UF or polyurethane glues are even better still for end grain.
So these are the chief elements to achieving first rate mitres in the contemporary workshop, where the overarching assumption is that quality comes from producing precise and interchangeable components. But in woodworking there's
always another way, and our forebears, without power tools, needed to find one. There was a thread recently about a chap struggling with his mortice and tenon joints, the solution was scribing the tenon shoulders. It's disappointing that this approach,
fitting your joints together by hand, is slowly dropping off the radar of many woodworkers. It requires hand tool skills, but where's the satisfaction in woodworking if not in that? Andy T hinted at this approach with his suggestion of running a plunge saw across the joint, but that particular remedy won't work as it'll just open up bigger gaps elsewhere. Finely tuning your errant mitre with a block plane would still produce gaps elsewhere, but at least they'd be smaller!
Many years ago I studied under a well known cabinet maker and restorer called Bruce Luckhurst, he showed me an alternative approach to fitting mitres that I still use today for inlay work. Incidentally, he also showed me the tiny marks on antiques that the original makers had placed there to assist with assembly, I guess they were usually in pencil and were subsequently removed, but every now and again there were minute imprinted dots or notches, they were needed because all the components, the stiles and rails etc, were
not interchangeable, they had all been hand fitted together so needed to be glued up in that one unique order. The secret to Bruce's method for fitted mitres was working in a particular sequence, a sequence that ensures you're always fitting a component
between two previously cut mitres.
You draw up a rod on a sheet of scrap giving the inner and outer edges of your frame. You cut all the components a couple of mill over length and already mitred to the best of your abilities. You can use a shooting board like the old boys, or a really good modern option is a disc sander. You're not trying for a precise length so a disc sander is an excellent tool for this, you'll find you can make a fence from a bit of straight scrap and cramp it to the table with a couple of G Cramps, from there you can play around, tapping the fence with a hammer and checking until you've got a really precise 45 degree angle. By the way, I can find that precise angle faster on an inexpensive disc sander with a bit of scrap for a fence, than I can with a Festool Kapex or a £20k panel saw. You then tape down (or use hot melt glue or whatever) two
opposing arms of your frame onto the rod with the mitres precisely aligned to your pencil marks. You then
fit the arm
between these two fixed references. With practise (it takes a modicum of skill) you can creep up on the joint, making minute adjustments as you go, planing or sanding off tiny amounts here and there, until you arrive at a perfectly gap free joint. You then number these joints on either side so you can reconstruct them exactly at the glue-up, and progress around the frame in a similar fashion. This is how I achieve perfect mitres on inlay work like this,
