DeskShelves

[MarkDennehy]

Right. So, first off, you can blame Custard for this, I'm reasonably aware of how cack-handed my stuff can be, but he's been trying to get my head to inflate to where my ego won't let me out the door, so here we are, a project thread without the lovely warm fuzzies of the Dunning-Kruger effect to ease the butterflies.

So, first off, meet the client.

Just turned six, now in Kindergarten and headed into Vorschul in autumn (he's going to a german school because the Irish schools here all have priests in them) and while he's starting to learn to read and has the usual amount of books first kids get bought by well-meaning parents, we haven't gotten him any shelves in his bedroom yet or a desk to do homework on, so, that's the project. A first desk, something that he'll mainly play on with toy cars, with a shelf or two for more toys and his books, kindof as training wheels for when he needs a "real" desk for schoolwork in a few years.

So, I want this thing to be fairly robust.I did some mucking about with the growth charts for boys aged 2-20 (I figure this should last him till he's nine, so that gave me a starting height and a final height for the client to design the piece around), the Woodworker's guide to Furniture Design, took stock of the material I had in the shed and had a think about how I wanted to decorate the finished piece, had a look at Paul Seller's leaning wall shelves which I liked the look of but not the joinery (like I said, this thing is for a six-year-old, it needs to be closer to concrete than fine joinery), and came up with a few design criteria/features/whatever:

• Top shelf at 48″
• Bottom shelf at 17″ (it’ll double as a play surface/desk)
• One intermediate shelf about 18″ above the bottom shelf
• Poplar for everything bar the bottom shelf which will be walnut (because that's what I have)
• Stringing on the bottom shelf to inlay things like a racing track for toy cars into the surface as well as things like letters and the like in various places
• Curves everywhere instead of sharp edges and corners for little eyeballs
• A serious set of problems with clamping curved bits into housing joints during construction.

The other thing was that I wanted the sides to angle outwards:

My sketchup skills are rusty - I'm not having a lot of success with curved parts there, so regard those cad things as just a very rough way to try to explain what the joinery problem I was worried about was - namely, if those were housing dados, how the hell would I clamp that during glue-up? So I asked on here, and got a fair amount of good advice, so the desk and intermediate shelves will be attached with sliding dovetails, probably with a locating pin or two for fits and giggles. I know my limits, so there's no way I could cut the sliding dovetail accurately enough (I could make it hold, but I couldn't be sure it'd be tight enough to stop things rocking back and forth in the breeze). So there's a router bit to cut the dovetail on the way and I'll finally dig out the aldi router table and set it up for the shelf part of the joint.

The top shelf, in the meantime, will just be a straight through dovetail, for more rigidity. And I think that at the back of that top shelf I'll put a single board acting as a back plate - it'll extend above the top shelf by a few inches so that it'll be another decorative surface, the back will be in contact with the wall as the main contact surface, and the bottom edge of it will be a french cleat for even more stability. With the thickness of material I have, the shelves will support a nine-year-old climbing up to the top shelf so the thing can't fold sideways and drop, hence the cleat. Yes, it's probably massive overbuilding, but it's in a good cause.

So this won't be a fast build for two reasons : 1) I'm not that fast; and 2) I don't really know what I'm doing But I'll post the photos as I go. So far, the most I'd done was to sunbathe the poplar I had to try to mellow out the green stain on some of the pieces, and sketch the design out in full-size to see what it looked like (it looked too small. Kids. They're a weird scale...)

 

[RickG]

Hi Mark,
All your plans look well organised to a newbie like me. But what I can help with is I know that desk height in a primary school is generally 500mm (20").
Adult desk height is 750mm, btw.

 

[MarkDennehy]

I got a small amount extra done this weekend - this heatwave is not helping things move along and neither was trying to set fire to my shed while making a new card scraper - but it was mostly failure

Interesting shape for the card scraper though, and it actually seems to work which is outright shocking to me.

I did get the shed tidied up from the last few projects and did some odd jobs that had been hanging around, like putting castors on the cart that the bandsaw and sander live on and then realising that I now had removed an excuse not to sweep the floor...

But at least that's done and so the first thing to do was to make more shavings to replace them.

Woodworking al fresco, in tie-died thai fisherman's trousers no less. Look, it's 28C in the shade over here and 38C in the shed after the angle grinder antics, you're lucky I was wearing trousers of any sort after that.

I actually found a walnut board I was certain I had already used, buried under shavings after it had fallen on its side, the sneaky little fecker.

So I left the nice 60" board I had alone for another day, and cut out the 36" chunk out of the 56" board I had which had a broken edge, and okay, the dimensions are smaller than I was thinking. It's not 30" at the back now, it's 24" and the front is working out at around 33". But I didn't want to have to run off to the timber yard to buy more boards if I could help it, so if it shrinks a little and is still usable for a six-year-old, well, I can live with that. Plus the width is reasonably independent of the heights for the shelves.

I did actually fret over this for a while, but I keep telling myself that custom-made stuff isn't machine-made stuff and it's okay if some dimensions shift a bit, so long as the important ones don't. And as soon as I nail down what's important and what's not, I'll be much better at this sort of thing.

The angles of the sides of the walnut shelf there are the same as the angle the sides of the overall unit make with the floor (same setting on the bevel for the sides as for the feet and the top shelf). I hear that kind of echoing of things in a design is good. It's not completely random coincidence at all, honest...

At this point, I thought I'd be clever, and skip the prep where I skim-plane the faces of the boards, because they seemed fairly flat and fairly similar in thickness, and tried to go straight to an edge joint and then prep the face of the entire panel so I only had to prep once. So I cleaned off the worst with the #05 and then tried for the bookmatching thing with the #08 (I love that plane, it's a monster but it's my monster)...

I actually got reasonably close - the joint felt like I'd gotten the sprung joint thing to work when I dryfitted it here. But I was kidding myself - the moment I tried for the glue-up, it started swinging round the midpoint like a top. I tried for a few minutes to get it to sit in, but no matter what I did, it was rocking back and forth on that midpoint by maybe a half-mm or so (maybe less, it was hard to tell with the glue on). So I gave up, broke the joint down, wiped off the worst of the glue before it set and gave up for the evening. Next time (and I'm deliberately not saying tomorrow to keep my expectations low), I'll clean up those edges again, skim the boards faces and get them at least respectable, and then try for the edge joint again.

 

[MarkDennehy]

But what I can help with is I know that desk height in a primary school is generally 500mm (20").

Hi Rick, yeah, I know, but the desks in primary school are for older kids. I honestly thought 17" was waaaaaaaaay too low and I'd screwed up the maths completely, so I had him sit down and checked with a ruler - right now, 17" gives him over 2" of clearance between his legs and the underside of the desk. Now, he could jump centriles on the UK90 graph and wind up taller than predicted in the next few years, it's not like those graphs are carved in stone or anything. But 17" should work until he's about nine, by which time he'll need a real desk anyway; it's a guess, I know, but I needed some sort of figure to aim at and at least this way I get to blame the graph instead of having to stand up and take it on the chin myself when it all goes pear-shaped

 

[custard]

I thought I'd be clever, and skip the prep where I skim-plane the faces of the boards, because they seemed fairly flat and fairly similar in thickness, and tried to go straight to an edge joint and then prep the face of the entire panel so I only had to prep once. So I cleaned off the worst with the #05 and then tried for the bookmatching thing with the #08 (I love that plane, it's a monster but it's my monster)...

I actually got reasonably close - the joint felt like I'd gotten the sprung joint thing to work when I dryfitted it here. But I was kidding myself - the moment I tried for the glue-up, it started swinging round the midpoint like a top. I tried for a few minutes to get it to sit in, but no matter what I did, it was rocking back and forth on that midpoint by maybe a half-mm or so (maybe less, it was hard to tell with the glue on). So I gave up, broke the joint down, wiped off the worst of the glue before it set and gave up for the evening. Next time (and I'm deliberately not saying tomorrow to keep my expectations low), I'll clean up those edges again, skim the boards faces and get them at least respectable, and then try for the edge joint again.

You have my sympathy Mark.

Edge jointing is the single most important joint in many projects, but bizarrely it's largely ignored or at least downplayed. When I talk to hobbyists it's pretty clear that almost everyone struggles to edge joint narrow boards into one wider board, which often leads to hours of frustration and all sorts of poor quality or just plain wacky work arounds. What's more, edge jointing is a mountain that you have to climb fairly early in your woodworking career, and because wider boards are becoming ever more difficult to find it's not practical to postpone edge jointing until you've picked up more experience.

The problem is edge jointing by hand isn't easy to explain, as success is largely determined by the subtleties of hand plane technique. It's like explaining to a learner driver how to pull away from the kerb without stalling. You can say "lift the clutch until you feel the biting point" until you're blue in the face, but it'll take lots of failed attempts before anyone actually gets it. And it'll then take lots more attempts before they "get it" when parked on a hill, or with the wheels sunk into gravel, or any of a hundred other variables that all change the experience. Edge jointing by hand is exactly like that, a frustrating sequence of trial and error before it all finally comes together.

If I get some time later today I'll try and put together a WIP that illustrates some work around options, plus some of the techniques that seem to work for me.

 

[thetyreman]

edge jointing requires unbelievable patience, try eyeballing it mark, you should clearly see any bumps that way, I find it a lot easier than any other method, also if you haven't flattened and trued the face flat first it makes things a lot harder.

 

[delboy47]

"If I get some time later today I'll try and put together a WIP that illustrates some work around options, plus some of the techniques that seem to work for me".

Look forward to that Custard because this has always been a problem for me. Mind you as a hobbyist I'm much more of a bodger but this has always been a bit of a bugbear. Finding the right swearword sometimes helps!!

 

[DTR]

So I left the nice 60" board I had alone for another day, and cut out the 36" chunk out of the 56" board I had which had a broken edge, and okay, the dimensions are smaller than I was thinking. It's not 30" at the back now, it's 24" and the front is working out at around 33". But I didn't want to have to run off to the timber yard to buy more boards if I could help it, so if it shrinks a little and is still usable for a six-year-old, well, I can live with that. Plus the width is reasonably independent of the heights for the shelves.

I did actually fret over this for a while, but I keep telling myself that custom-made stuff isn't machine-made stuff and it's okay if some dimensions shift a bit, so long as the important ones don't. And as soon as I nail down what's important and what's not, I'll be much better at this sort of thing.

As a rank amateur I've got no authority to say this, but I agree with that statement 100%. I often find that the only critical numbers are the overall dimensions; everything else can be adjusted to suit the size of the boards at hand. When I see beautifully CAD'd drawings with everything dimensioned to the nearest 0.1mm, I do wonder how much extra time and effort is wasted in hitting those arbitrary numbers. Naturally batch production is a different game, but I doubt that applies to most of us hobbyists.

Watching with interest

 

[MarkDennehy]

Rather a long week with lots of late nights in the office and zero time in the shed all week till the weekend. The router bit for the sliding dovetail joint did arrive though.

Honestly, looks tiny to me, but we'll see.

Looked at the boards end on, just to depress myself. Fair amount of milling required and I don't have a thicknesser.

Large hollow on the board on the show face side, took three passes with the new scrub plane to get them out.

I say new, it's a lot older than my father.

Then the other side, and if there's a hollow on one side, you know there'll be a bump on this one and this time it was a bit of a mountain. I lost count at pass eight...

Finally good enough. I'll have to go over this again after the panel glue-up, no point getting it micrometer perfect yet.

Sweaty work.

And of course, now that the board's flat, it's about three-quarters of an inch thick so the other board has to match and it's just over an inch. Because of course it is.

Planed the show face flat and out of twist (it almost had none), marking gauge set on the other board's thinnest point then run round the edges of this one, planed chamfers down to the line and then cross-hatched them with chalk to monitor progress easily, and then it was time for cross-grain pushing again.

I hate this bit. Honestly, it's not interesting, it's not challenging, it's just dull and tedious donkey work. I'm really starting to run out of reasons to not buy that Dewalt 735 or something similar (I can't fit a floorstanding machine with a quiet induction motor in the shed, so it'd have to be a lunchbox screamer, and the 735 can take a helical head so I could at least reduce noise that way).

But I couldn't even finish this (it needs another few minutes with the #05) because even with the nicer new scrub plane which is a lot easier on my delicate little hands, they are now missing skin off one finger and have a beautifully situated blister right where my palm rests while typing. This week's going to be a bundle of fun with that. I mean, software engineer. Typing. It's kindof unavoidable in the job, y'know?

sigh

One day, one and a bit boards, no joinery. Doesn't feel like progress.

http://www.stochasticgeometry.ie/2018/07/15/ouchie/

 

[thick_mike]

For me the two most important power tools are a bandsaw and a planer thicknesser, because resawing and dimensioning are the two most tedious and exhausting processes in any project. Once they are out of the way, the fun starts.

 

[MarkDennehy]

So work on the deskshelves has been stalled owing to two weeks of 12-14 hour days at work and then a traditional Irish family seaside holiday.

Yeah, it's a bit like a traditional Scottish seaside holiday in many ways.

I did get to do some window shopping on the trip home though, at the carpentry store...

So not bad as holidays go. Then, bank holiday weekend here, and the latest addition to the shed had arrived (thanks again MrsC)

So once I'd finished mucking about sticking a DRO onto it, I ran the walnut board I was thicknessing in the last post through the 733 and knocked the last mm or two off it to get it to rough thickness. I now love thicknessers.

I will definitely need to improve extraction, mind you.

Then on to edge jointing again, only to find I now had a slight bow in one board (not much, about a half-mm at the worst, so there will be further handplaning to come).

But I got a nice spring joint this time after the plane-both-together approach followed with light planing on either board separately to set up the spring joint.

Let that cure overnight and tomorrow I'll see how bad the flattening job will be. I don't expect it to be awful though. And the edge joint felt *much* better this time round so I'm mildly hopeful it'll be grand...

 

[MarkDennehy]

So the panel came out of the clamps in an okay state.

The clamps actually deformed as they're those cheap aluminium sash clamps and I was right on the last stop; I'll have to stuff those things with wood to stiffen them up after I reform the ends.

The step in the middle was a bit worse than I thought, just over a mm of a step on the upper face.

So out with the #05 and I started flattening it like a workbench top.

I know, it's not doing great things for smoothing but I figured get it flat first, worry about smooth later. So I took it down to being flat, and then got out the #04 1/2 and gave it a quick smoothing. I'm not quite ready to try for finishing level smoothing yet, I just wanted to keep it tidy. And of course the grain in one board is running against the grain in the other so smoothing is fun, but between a bad knot on the underside and some bowing in the boards and not wanting to have a massive step to smooth away, I was kindof out of options.

The chalk's highlighting areas that have some nasty tear-out with actual chunks missing (they're not huge, but they are almost 2mm deep at one point); I may need to fill those with epoxy or put an inlay through those areas to hide it or something, but even the card scraper wasn't helping much. A final smoothing closer to finishing might help, but somehow I don't think so.

The pencil lines are the intended edges to keep the edges splayed out. My idea for how to build this project kindof starts with this part as the core and the other parts getting built up using it as the reference. I haven't made the cut to the line yet, the pencil is just for me to keep an idea in my head of sizes and shapes. I'll do the other shelves and the sides first, the actual stock prep that is, and then I'll do the cut-to-shape work and then cut the joinery (I might reverse those two for the sides somewhat because the curves might make workholding too much fun, whatever about cutting the straight-line slopes for the ends). And there'll be inlay and other fun bits tucked in there too.

Next up will be the poplar shelves for the middle shelf and the top shelf; I don't know if they need much thicknessing, but if they do I might leave them till the weekend and use the Dewalt; the sides on the other hand, I don't think I can run through the Dewalt because I'd have to do it outside the shed to have enough room and it's waaaay too loud for that. And once the pieces are all prepped, I'll think about inlays and the like, and the joinery.

Incidentally, as the desk part is 17" off the floor and for a six-year-old, I'm trying that whole "finishing for the queen" approach with the underside of the desk part:

That got "flattened" with the scrub plane and given a rather cursory smoothing pass with the #05 and not on a light cut either. It's not quite "textured", to use the hipster term, but it's definitely not that smooth. "Undulating" maybe We'll see if my faux-OCD can stand it

Worse comes to worst, I can distract myself with this:

That's what'll happen if you have a Phillips screwdriver on a magnetic bar three feet over the bench and you put a plane back in the till a bit too clumsily. Bugger. Not even in a part I can cut away. I guess I have another site to put some inlay in...

 

[MarkDennehy]

Not abandoned this, by the way. Just work cranking on the organ grinder handle a bit faster of late, and then getting sick (children, it turns out, are basically large petri dishes of germs that spill on you regularly). Anyway, got out the new thicknesser today, flattened the remaining shelves and side uprights by hand on one side (as best I could - hard to flatten something that's just longer than your bench because you try to tap the corners and one's in thin air over the edge of the bench), and ran them through the machine taking them down a mm at a time (and on both sides once I was getting full-face cuts) to an inch thick.

I even remembered the head lock-off mechanism existed this time

All donkey work, but done in a half-hour (bracketed on either side by a good 40 minutes of setup, teardown and cleanup) instead of over most of a week's free evenings. And obscenely loud to boot. Oh well. Done now.

And of course, that's just rough thicknessing. There's more refinement and shaping and joinery and inlay and finishing to come yet, but this was the donkey work that this *** wasn't looking forward to. The rest should be more enjoyable

 

[AndyT]

I'm watching with interest. I like threads like this, which are honest about the amount of time it can take just thinking about stuff, sorting out the boards, jointing, thicknessing. Much more realistic than "I knocked this up in an hour and it was all easy" and more satisfying too.

 

[marcros]

Worse comes to worst, I can distract myself with this:

That's what'll happen if you have a Phillips screwdriver on a magnetic bar three feet over the bench and you put a plane back in the till a bit too clumsily. pipper. Not even in a part I can cut away. I guess I have another site to put some inlay in...

A damp cloth and a steam iron over it and I think that will come out. Certainly worth a try.

 

[MarkDennehy]

I'l certainly give it a go, but I think the very tip might have punched the fibres down too sharply and may have severed enough to stop that from working

 

[custard]

It's astonishing how effective steaming can be.

Here's a Cherry board with a nasty dent. I can't remember when I took these photos but the red semi circle is the symbol that means it's the rough sawn component for a drawer front, so my guess is this matched other drawer fronts, and if it couldn't be rescued it would involve a pile of faff to replace,

Looked at closer,

A spritz of water on the dent, give it a moment to soak in, then a damp cloth on top. Set an iron to "cotton" but switch the steam facility off (or drain all the water from the iron),

Bingo!

If necessary give a second or third application.

 

[sundaytrucker]

I was thinking this too.

Looking forward to seeing this come together.

 

[MarkDennehy]

Not much time in the shed today, but enough to make a start on the sides. One’s developed some twist since I milled them. Not a huge amount, but noticeable – 3-4mm at the worst point. A lot of that twisted portion might be cut away so I don’t know if it’s something I need to get out of the board yet (or even if the shelf joinery will pull it back to straight). But I’m starting on the untwisted one and I’ll see how we go.

Marked up the cuts for the foot and the wall rest sections using the bevel, and the front curve by bending a long piece of dowel stock (didn’t have anything else that was long enough and thin enough to act as a batten). And then started making a reference edge from the back edge. The sides are long enough to make this awkward…

There’s barely enough room to do this. Really, there isn’t enough, but by being fiddly enough about it you can just about get away with it. Then, with the reference edge established, out with the bevel and the marking knife and time to mark off the final line to cut to for the feet and wall rest. First though, need to take off the last half-inch or so of board, give me another reference edge square to the back edge.

And cut off and then plane to flat.

This is where having an apprentice can help (incidentally, for anyone who was wondering, this is why I have a Record #03).

Once we’ve passed QA, it’s time to mark out the wall rest with the knife and bevel and then cut it (using a short batten to help guide the saw because the cut’s awkward due to the lack of space).

And then a few swipes with the #04 to clean up the cut surface when done. Do the same on the other end for the foot and now it’s time to check. I don’t have a five-foot-tall try square, so I’m just using the door frame of the shed.

So-so. There’s a gap at the wall rest.

Thing is, I’m pretty sure my door frame is as square as a rhombus, so I’ll check this with the bevel and might tune it slightly, but I don’t believe the door frame is a reference surface you’d trust too much

Yup, bevel says the wall rest is slightly out. A few swipes with the #05 will sort that, but I’ll check against the actual wall it’s going up against before I do that.

Had to stop there, but tomorrow there’s a few more hours and the next step may be to cut the curved section at the top. Maybe. I’m not certain if I want to do that yet, the parallel sides might make workholding easier. Have to think about that one…

 

[MarkDennehy]

So I decided to go ahead and shape the front curve today, since it was going to be awkward to cut and might be noisy so it’d have to be a weekend job really.

Why is it awkward? Well, the plan is to make the first rough cut with the bandsaw and then to get down to the line with the compass plane and the board is long enough that swinging it around through the bandsaw inside the shed would be awkward and would probably need the door open and the plank sticking out at some point.

See what I mean? There’s not enough room for a cat to stand in here, let alone be swung around. But, with much swearing and cursing and with the door open and the plank sticking out for at least half the cut, it got done.

I even cut the top of the side to be parallel to the floor again while I was at it.

You will notice the artisan scalloped edge, and I’ll have you know that it takes a lot more effort to create such an artistic statement than it does to just cut a straight clean boring line.

Or something.

Anyway, it’s just a rough cut, so out with the #05 to knock off the absolute worst of the knobbly bits and then time for the fun tool.

The Record #020 is such a a natty little tool, and while there are limits to how curved you can go, within those limits it’s great. I don't use mine enough I feel because by the time you get to the size of work where it's comfortable, you're right on the edge of the size of work where you can't get it into the shed But you can prise it from my cold dead hands, it's too much fun when it gets going. I mean, it's not *quite* a spokeshave level of fun, but it's very close.

Mine is a little worse for wear in appearance. Everything works, I had it apart, cleaned, oiled and resharpened everything, but it’s had some light surface rust over the summer and the enamel’s long gone (and I still haven’t figured out a good way to restore that with the kit I have or can use). But anyway, it’s more than good enough to do the job and between that and the spokeshave I soon had a smooth curve instead of a decaying sine wave.

And I hate the look of it completely. The curve itself is more or less okay, I got rid of almost all that ugly damaged bit, but that little flare-out at the point where the desk will be just doesn’t work. It looks wrong in several different ways. So I ran the plane over the entire edge to get a nice single smooth curve instead of that little flare-out.

I’m more or less happy with this. I was thinking of making that curve into a bow rather than a sweep, so that the front edge would be almost vertical at the foot (that’s what those black lines there are, they’re not spalting, just me trying to see what it would look like). I’m not sure about this though. I’m worrying about the width of the sides and the strength of the piece if I start hacking off that much, but I might just be getting paranoid. And there's a certain symmetry - at least on this side - between the curve I was thinking of and the flow of the grain lines.

Anyway, I’ll leave it at that for now and maybe think about it again later before I cut the sliding dovetails (which will be the point of no return for the shaping I think). Next job, start to transfer the shape from this side to the other one…

I’ve already gotten the foot done (and it matches its counterpart well) and a reference edge planed on the back edge, but I haven’t finished the top straight-line cuts yet and then there’s the second front edge to shape. That should be interesting. I haven’t changed the compass plane’s setting so in theory it’s going to be grand…