- This project was to build a small-pet hutch and explored:
- how to compare material prices
- optimising the use of stock material
- dealing with irregularities in stock dimensions
- Made-to-measure versus cut-to-spec working
Our eldest cat Kickass (sometimes Kicky, sometimes Kix – depending on how adorable he's being) does like to be out in all weathers; sometimes sheltering under the hedge in order to be on look-out, or else huddled under the neighbour's shed. Having just finished making a luxurious 8 cubic meter pleasure palace for the ferrets we thought that Kix should have a little shelter of his own. We could have bought something new (=mass produced and somewhat ugly) for around £40, but since I'm spending a year with wood I naturally decided to make something myself at around that budget. Here it is completed – it's too early to say if Kix will adopt it or else will be annoyed that this thing has suddenly appeared in his favourite corner of the garden:
As you'll know if you're reading through all of my woodworking articles the cost of the timber is a constant problem. It seems like it is impossible to hand-make anything at all for less than the cost of buying something new. It's very frustrating, making something by hand should not only be 'better' but should be bloody cheaper too. But it never is. Anyway, I'm not ranting on that topic again – suffice it to say my budget meant that I wanted to use the cheapest wood possible.
Wood of course is sold 'per length' or 'per board' – so comparisons in cost aren't immediately apparent. Eg. Battens are available at 48x25mm costing 60p per meter or we can use 25mm ply for a wall panel of the same thickness. 25mm ply comes in 2.44 x 1.22m boards at a cost of £76 per board. If I want to make a 25mm thick wall panel measuring 50x50cm , which is the cheaper wood to use?
I would need 10.4 half meter strips of the 48mm battens, which would be £3.12; or about 1/12th of a plywood board which would cost £6.40. Making the plywood twice the price of the battens. Plus of course I only need 6 walls in a hutch so actually I'd have to buy a full sheet of plywood when I only needed half of it – making this project more expensive in terms of outlay. My frugal nature makes me shiv-iv-iver at the thought.
To compare the prices of different materials I calculated the cost per cubic meter for each by multiplying the 'unit dimensions (in meters)' by the 'unit price'. For the battens this was 0.048 x 0.025 * 1.0 * 0.60p (=£500.50/m3) and for the ply this was 2.44 * 1.22 * 0.025 * £76 (=£1019/m3).
So I decided not to use sheet materials. I would make the entire hutch out of 48 x 25mm battens.
Actually, my intention was to use 50 x 25mm battens, since that's what the website I used said they would be. Now I'm quite used to the weirdness of wood sizing, I know that 2 by 4's are really 1.5 x 3.5's and I 'get' why machined timber will have a 'nominal' and a 'finished' size; but I just don't understand why the 50mm battens I ordered turned out to be 48mm... (worse still, some were actually only 47mm) so I asked the supplier MKM Building Supplies. But they didn't bother replying, which is a shame because I was otherwise well ready to recommend them for their prices and their fast and courteous service. Perhaps we can blame their lack of response on COVID-19? Yeah, let's do that, and recommend them anyway; they do usually over-deliver by a little so that's always nice...
I'm not sure Kix was happy that I raided his larder when it came to balancing the stock on my saw horses
There are a lot of pieces in this project (106 to be precise) because the walls, floors and roof panels are all made from lengths of 25x48mm battens that have been glued together to make boards. There are 58 pieces in the simple panels and a further 25 in the super complex front facia and roof-gables. So although there's a lot of pieces, most of them are super simple lengths that are cut and glued together. There are only 22 pieces that need joints cutting in them for the internal frame. We'll look at those later but I want to talk about cutting the stock first.
In previous projects I've pretty much cut all the lengths I've needed from the stock in a kind of haphazard way – you know, just chopping a bit off each time I needed to. I always over order by around 10% so I haven't had to worry too much about making efficient use of the stock. But there have been times when one piece went wrong so I needed to cut another length, only to find that although I still had plenty of unused wood available, none of it was long enough for the piece I had to re-cut!
So this time I wanted to be sure that I made most efficient use of each piece of stock material.
In practice it's a relatively simple matter of always cutting the longest piece that you want from the shortest piece of stock that you have - but with 106 pieces needing to be cut into one of 21 different lengths the scope for error was quite significant. So I made a spreadsheet to help me plan the cuts. Feel free to download my CUT-PLANNING SPREADSHEET if it's any use to you (maybe use the DONATE button if you do download it).
There are on-line tools that help in planning how to cut pieces from stock, so if you don't like spreadsheets take a look at this free online tool from 'cutlistpro'. They have a pay-for downloadable app version but I don't see any advantage of it over the online tool. Buy it if you like the tool and want to support them.
I think my spreadsheet is a little superior to that online tool (I would though, wouldn't I?) because it lets me name each piece. Naming the pieces isn't any help in terms of managing the cut-to-length plan, but is very helpful when cutting the joints and when assembling the pieces.
I've used a three factor nomenclature in order to suitably identify each piece; describing in turn:
Which way the piece faces, relative to the front elevation: left, right, front, back, floor or roof.
The position of the piece: to the left, to the right, at the front, or at the back (for vertical pieces); or else as upper, mid or lower (for horizontal pieces).
The purpose of the piece: upright support, horizontal support (skirt or strut), vertical or horizontal facia, floorboard or roof board.
It maybe sounds complex, but put in to practice it really simplified my thinking about each piece; I could look at the construction as it developed and work out which piece I needed next without referring to any diagrams – which were all blowing away in the gusty spring winds it seems we get up here in the moors. Having made part of the left wall frame, for example, I can look at what I am doing and readily think to myself 'oh it must be the left facing back positioned upright support piece I need next'. And because my cut-plan is organised by piece name it was really easy to look up 'left back upright support' and see I needed a piece 55.5cm.
It was actually just a little more complex than that, but you can probably make sense of it – here's a picture demonstrating how I named the pieces (it also shows the internal frame construction I used):
And so here's the full cut-length specification for the pet house:
(click or tap any table row to enlarge)
|front||left, jamb||upright support||2||68|
|Left, right||front||upright support||2||68|
|back||left, right||upright support||2||55.5|
|Left, right||back||upright support||2||55.5|
|back||upper, mid, lower||skirt||3||53.5|
|front||upper, jamb||strut support||2||47.8|
So most of the pieces are used to create panels for the walls, floor, roof, roof gables and veranda.
There are six 5-piece panels for the left, back and right hand walls each being 53.5cm long.
11 pieces make up the roof. The centre rib (which sits upright rather than flat like the other pieces) is 87.5cm and each subsequent piece moving out from the centre is 2.5cm shorter than its predecessor, making the outer most pieces 75cm. The front roof strut support is added last, it sits at an angle (as it sits flat against the pitched roof boards) and abuts the front facia. (I also added a rear roof strut support as a last minute decision).
The two sloping gables are cut from one 3-piece board 53.5cm long. I made these last so I could take an 'as-built' measurement to get the diagonal cut absolutely right. Obviously I had used a right-angle triangle calculator to work out that the angle of the roof would be 20.497 degrees for a triangle that rises 20cm across a span of 53.5cm... but there's just far too much scope for error to rely on that, so I held the gable abutted with the in-position roof and drew the required cut-line in by hand.
The veranda is made from 6 pieces 56cm long. Like the roof it was painted before it was glued so that the stripes would be as sharply defined as possible. In fact I wish I had done the same for the front facia, but doing so would have cost another couple of days in waiting for glue-stages to fully cure. Don't be impatient like what I is...
The front facia panel is the most complex due to the interplay of horizontal and vertical pieces; but before looking at the detail of it I want to make a note on sizing...
Kicky on site-inspection.
The 106 total pieces range in length from 14.4 to 88.5cm, in fact there are 21 different lengths of piece used in the design. When I made the original cut-specification there were only 17 different lengths of piece required. I tweaked the design specifically to minimise the complexity of it in terms of differences between lengths of pieces. I had also ensured that all pieces would be a whole number of half centimetres long – so they may be 56cm, 55.5cm, 51cm, 50.5cm etc... but never 51.1cm for example. When the wood arrived and wasn't quite the right size I had to tweak the design. Not only was the wood 2mm narrower than anticipated, this meant that it wasn't exactly twice as wide as it was thick – so that even though the external dimensions are still sensible (a whole number of half centimetres) the internal dimensions slip about by +/-2mm and this leaves us with awkward lengths like 27.1cm or 14.4cm.
In some regards cutting a piece to 15cm is exactly the same as cutting a piece to 14.4cm – it's just a slightly different mark on the ruler. But in practice it is easier to find the full or half centimetre mark on a ruler than a specific millimetre mark. It is also way easier to hold numbers in your head if they are always 'something point zero' or 'something point 5' – which reduces error for sure.
I could go back and re-simplify the design for 25x48mm wood stock – but then I'd be presenting an untried design. Some of the numbers are awkward here, but all of the pieces do fit together!
To cope with all this trickiness in the wood size I first of all made each of the simple panels (the roof, the floor, etc...) at the initial design length (of 55cm). I then measured these panels and found that the roof was 2cm narrower than expected. The wall panels were also short by 2cm for each wall. The whole construction turned out to be 2cm smaller in each dimension (x, y and z).
At this point I constructed the internal frame, adjusting the design lengths to match these as-built panel sizes.
And then finally I measured the front panel aperture in the construction and cut all the pieces for that in a 'made-to-measure' approach. In general I really don't like taking the made-measure approach. What it means in practice is rather than correcting an individual piece you end up making tiny adjustments to everything that follows, probably making more mistakes along the way which are hard to spot until you get to the end and find that nothing joins up nice and square. In this case though I was dealing with the vagaries of the wood stock, so it seemed the best approach. It was also the last section to be built, so there was no fear of throwing-off the rest of the build.
Made-to-measure versus cut-to-spec:
In general I cut all pieces to the length specified by the design, and all joints are cut as per the design. If I make a mistake I don't then propagate that, by making micro-adjustments to the lengths of the pieces (or position of the joints) that follow. Instead I either re-make the piece or else put up with the error at that point (technically referred to as 'bodging it'; but I try not to bodge it too much). The point is I never let errors propagate, because they end up causing more errors. But there are times when it is better to take an as-built measurement than to rely on the design specification. Namely when dealing with any angle that isn't 'simple' (0, 30, 45, 60 or 90 degrees) or when fitting 'final pieces' – ie. Pieces that have no knock-on effect in the construction.
The rule is:
Avoid working made-to-measure; unless you can't
Here's the actual layout I ended up with for the front panel:
And in case you want to spoil your own small-pet with a completely bonkers summerhouse, here's the full design spec:
But hang on, there's no dimensions shown in this spec! I know, right! But look, it turns out I didn't know the dimensions of my own wood, let alone that of whatever you might be using – so it hardly seems helpful to litter the spec with loads of precise dimensions. Instead I'll use relative dimensions so you can adapt the spec to whatever wood stock you decide to use. I think 25x48mm has turned out to be very, very solid and actually way too heavy. You might be better with something like 19x38mm but you'll have to calculate your own precise dimensions for that from this pattern.
So how do the relative dimensions work?
There are 2 concerns: the lengths of individual pieces; and the sizes of the cut-outs for the joints. Both are absolutely relative to the width and thickness of the wood stock.
For example, the roof is made from 10 pieces abutted width-ways with one (central) piece abutted edge-ways. So the width of the roof is: 10w + 1e.
The design principle is that we are using wood that is exactly twice the width as the edge thickness (and the build is way easier if that is true). Which means we can say the width of the roof is 21e
The back wall width is as wide as the roof plus the left wall thickness (which it overlaps) – so the length of the pieces for the back wall is 10w + 2e, or 22e.
For 25x48mm wood stock the roof is 21 * 25mm = 52.5cm and the back wall is 55cm (design sizes, not my final as-built sizes using my out-of-spec wood stock!).
So that tells you how long each piece should be. A word of caution though, my wood stock wasn't twice as wide as it was thick (goddammittohell) so these diagrams have been worked up after the fact. They ought to be right BUT I haven't built against these design images. So just be careful! If you do happen to build one of these let me know how you got on – I'll update this article with any real world learnings.
We now need to understand the joinery of the pieces. Everything is either butt or lap jointed. I'll assume you can work out the notches you need to cut to make the floor fit when you are making your own pet hutch; the floor isn't lapped into the frame so cutting it to fit is pretty obvious.
So it's just the frame that needs any real joinery:
There are 22 pieces in the frame of 12 different types.
(relative dimensions:absolute measurements for 25x50mm stock)
A: Front LHS & jamb upright supports (2)
3 full-face (2e:50mm) half laps - one each end and one at the jamb height (21e:525mm).
B: Front RHS upright support (1)
As A but the bottom is half-notched (e x e:25mm x 25mm) instead of half-lapped.
C: Front upper and jamb-level strut supports (2)
3 full-face half laps - one each end and one at the jamb (5e:125mm).
D: Front lower skirt (1)
LHS & Jamb quarter laps. RHS half-notch.
E: LHS & RHS front upright supports (2)
3 full-face half laps - lower, mid and upper skirt levels (0, 11e,22e: 0, 275mm, 550mm).
F: LHS & RHS back, with back LHS & RHS upright supports (4)
3 full-face half laps - each end and mid skirt level.
G: LHS upper and mid skirts (2)
Full-face half-lap at back, with full-face quarter-lap abutting. Full-face quarter-lap at front.
H: LHS lower skirt (1)
As G but this skirt extends further forward than the others.
I: Back roof strut support (1)
LHS half notched at back. RHS full-face half-lap.
J: Back upper, mid and lower skirts (3)
K: RHS upper and mid skirts (2)
Quater-lapped each end.
L: RHS lower skirt (1)
As K but this skirt extends further forward than the others.
It's bonkers, right? Definitely not the sort of thing you could ever hope to buy. Hand-made stuff brings with it a great creative freedom; Why emulate the mass-market? I wonder if it was fortuitous that I had so much trouble over wood size in this project. That lead to a very merry dance in and out of the design world and the real world. It gave me a chance to think about the cut-to-spec versus made-to-measure approaches. Which feels like the same concern we have in writing between the 'getting it good' and the 'getting it right'. Ie. Recognising when to us the impetus of the plan to move forward and when to simply wing it.
In my next project write-up I shall look a little closer at the design phase, giving particulate heed to where I feel I should be prepared to just wing it. In the meantime, here's a final picture of the Kixkat Club with no sign of Kicky; who is still studiously ignoring it in so far as I can tell: