And after a fast skim.
Not worked in oak very much yet, this should be interesting…
And after a fast skim.
Not worked in oak very much yet, this should be interesting…
Part of the tidy-up in the shed meant getting the scrapers off the little holder they were in and up on the wall…
And then the second thing was going through every single plane and chisel on the wall, going over all the unpainted parts with 600 grit wet&dry paper and WD40 to take off the surface rust, then rubbing briwax into the steel as a barrier against rust (renaissance wax would be better and some’s on the way). I’ll have to do this every weekend for a while to get the barrier layer set up properly.
I hate condensation. Gah.
There may be a dehumidifier in the shed’s future.
Honestly, the plan was not to collect planes.
But through random chance the first few I bought (I bought a retiree’s toolkit off ebay as a starting point) were all Record planes. And over the course of a few months I learned, as everyone does, that hand tool woodworking was effectively killed off during the second world war, though it had been in decline for a while before then. Machines and power tools took over for wood fabrication; and after that point, tool companies could no longer compete by producing the best tools because workers were no longer competing on how fast they could do the job to a set standard. Nobody needs to spend a weeks wages to buy a saw that can be sharpened more so you can cut 5% faster, when they can spend that much and get a saw that cuts 500% faster. The demographics of buyers changed markedly and suddenly a tool that would stand up to occasional weekend use and not cost a week’s wages was the thing that was in demand (and thus was born things like Black and Decker). Older tool companies either changed (like Stanley) or went bust or were bought up in mergers (like Woden) and so there’s this quality curve that takes a slow or a fast decline at some point after the end of WW2 for every hand tool manufacturer. Older tools are substantially better than modern ones from companies like this; you have to go to the modern artisan makers like Veritas or Lie Nielson to beat the vintage stuff. And that’s seriously expensive by comparison so most people go vintage, at least in the beginning where I am.
Now Stanley were the main brand for almost a century (and are still around) so if you go to ebay and try to buy an old hand plane, they tend to pop up first. But there were so many, identifying them is an expert’s task – is this one a bargain that’ll work like a charm for decades or is it only fit to be melted down for scrap, not even worth the postage? So instead, I just kept looking for Record because they didn’t dive down the curve for a few years later than Stanley and they’re relatively easy to date even from ebay photos.
So my #4, my #4½ and my #5½ got joined by a #7 and a spokeshave and a shoulder plane and a rabbetting plane and so on, all from Record’s catalogues as I thought I’d need them. And then one day I turned around and aw, crap, I have a collection of the sodding things.
Well, now look where I’m stuck.
Every slot filled, with planes for getting things straight, getting things curved, carving slots or rebates into things, and so on. But two empty slots got build in there, for the last two obvious things I’d “use” – a large Record #8 jointer plane and a small Record #2 smoothing plane.
Came up on ebay and I couldn’t resist. Old too, this is from somewhere in the 1945-1950 range based on its design, frog, blade and handles.
Just that bit larger than the #7, so handier for some things. A tad rusty though and the paint had worn through in places. Well, sod collecting, I want a usable tool so…
Then out with the newly mounted wire wheel on the bench grinder and off with all the accumulated rust (and I tidied up the #6 while I was at it, as someone had painted that thing almost purple in the last while and it annoys me).
A bit hard to read but “Best Crucible Cast” there indicates that this is one of the better blades Record ever made. And the sharp corners suggest this is laminated but I’d have to grind the bevel off to find out and I’m not that curious.
Then out with the heresy Hammerite paint and sprayed both the #8 and the #6 (I may redo this if I ever find a decent source of the original Record colour, but I seem to recall it was oven enamel and that’s a bit of a faff).
Then we had a hurricane and work had a Thing and so it was left drying for several weeks all told. Yikes. So this weekend, back out to the shed, tidied up a lot, reassembled both planes, spent a while taking rust off damn near every tool (the shed’s damper than I thought and the felt in the tool racks now counts as a Bad Idea) and oiling them (WD40 at least) and finally…
See that little spot? There? That one?
Yeah, that’s kindof my little joke. It’s for a Record #2 plane.
They never made a #1, Stanley did that and these days those go for five-figure sums to collectors. But Record’s #2 still goes for stupid money, when you can find them at all. They didn’t make as many as they did of the workhorse #4s and #5s or even of the fancy #7s and #8s and when the second world war happened they went out of production in a hurry – and when there’s a war on, lots of metal gets melted down to make gun barrels and the #2 was a small smoothing plane for fine work; nobody really had a use for them that a #3 couldn’t fulfill so off to the crucible a lot of them went (same happened to the Stanley #1s). Oh well. They’re not as crazy as the Stanley #1s, but the last time I saw one on ebay, it sold at the asking price of almost €400 within 12 hours of being posted.
Yeah, that slot’s staying empty for a while longer I think 😀
But at least the shed’s been tidied up a tad so I can start doing things again…
So I’ve said once or twelve times that I needed to do something about this stuff in the shed:
The little air compressor proved useless for shellac and stuff so it’s moved indoors to the lab for future mucking about with airfix models with junior and the like. The dremel, well, that’s handy for stuff so it still lives there, but the grinder was just a pain in the fundament. It’s on an MDF base with a small cleat on the bottom for the Black&Decker Workmate to grab onto (or the vice on the bench these days). So I finally got round to fixing that up and getting it out from underfoot.
Smaller base, cleat at the front end to act like a bench hook and a holdfast at the back keeps it nice and stable (the base tends to cup a bit, I don’t think it’ll last forever). Pulled the stone on one side (the finer grit stone because if I need that grit, I just use sandpaper for the scary sharp method or I use the diamond stones – I’ve never used that fine grinder stone since I got it) and put a wire wheel on there to help clean up rust off some older tools. The wire wheel wouldn’t fit in the guard, so I lost it.
If that cleat looks familiar, it’s because you’ve watched woodworking videos on youtube in the last decade. It’s a french cleat with a matching cleat on the wall.
So now it’s out from underfoot and positioned perfectly to remove an eyeball as you enter the shed. Who really needs to look to the left though?
So not every piece of woodwork has to be fine and finished to be fun…
It’s surprising how much fun 5-to-13-year-olds can have with hammers and nails so long as someone older is there to do the branding and the sawing 🙂
But dear grief, it’s rough work and now I feel I need to remake a doll’s school locker properly in something nicer than badly warped pine and poplar offcuts 😀
8-pin microcontrollers are pretty neat things. I’d only used the larger PICs before, I didn’t see how these could be all that useful. And the 16C74s could do more… twenty years ago. These days, the 8-pin PICs have caught up with the older models in internal features, if not pincount. It’s really quite remarkable:
Same amount of program memory as the one from PICrat (actually, the modern version of the chip from PICrat, with the fancy flash memory), the CPU’s faster, it has more RAM, it has an internal EEPROM for persistent storage, same number of digital comms, one less CCP (but 32 fewer pins, so okay), half as many ADC channels (again, 32 fewer pins) but higher precision on those channels, same number of timers, a new comparator, a wider range of operating voltages and it takes less power (much less, it turns out).
Yes, it’s an 8-pin microcontroller; most of us forget that these shiny PCs are only a tiny, tiny, tiny fraction of the deployed computers in the world. These PICs probably outnumber them, and Moore’s law is in full swing on that side of the industry still.
So anyway, a fortnight or so back I’m watching a Big Clive video about making up a DMX cable tester with a PIC12F635…
…and I thought I’d like to play with a 12F so I bought a tube of them (they’re a lot less than a euro apiece) and a programmer. I splurged and got one of the higher-end 12Fs, the 12F1840, and then got to wrestling with resurrecting a 20-year-old toolchain.
GPASM, it turns out, has improved considerably, and there’s finally a CLI for the PICkit3 (I remember it coming out and the sheer daftness of the dropping of the PICkit2’s pk2cmd program that rendered the PICkit3 useless on launch to anyone running linux at the time). It only took a few hours of rereading manuals and poking things randomly before getting the flashing LED “hello world” program to work.
So, what’s next? Well, apart from the really neat WS2812 array in the vice there, which has a fun data protocol that’ll need a bit of programming work to hook up to (there is a C library for it, but I’ll have to rework that in assembly for my toolchain), and ignoring this idea I have for an interface between a sensor and an I2C link to a raspberry Pi that I have for another project, there’s also a little box dropped off by the postman to one side here. See, ebay’s one of those places where you can get fun toys for next to nothing if you look and are patient…
70 5×7 LED matrix displays for €17. 70! Mind you, I’ll need a few 74 series parts, some darlington arrays and a bit of wiring, but this could be fun…
So twenty years ago this year I graduated, and just after that (and before I started chasing a PhD), I did some work on a project to create a platform for building a micromouse robot. Not a speed run winner or anything, but just a bog standard, easy to use, off-the-shelf black box platform.
This was 1997, remember, these things technically did exist but they were research project level expensive and finicky and whole companies were making good money building them; Arduino is still six years out at this point and it would be years before you saw the level of available support hardware it enjoys today. Lego Mindstorms is years away and Lego’s RCX doesn’t come out till the following year. The BASICStamp is the most popular solution to beginner hobbyists and it costs the guts of $100 at the time if I recall correctly. You had PC104 based systems available but that was about as close as you got to a Raspberry Pi and your average PC104 was about as expensive as your average desktop PC. So for beginners you pretty much had Microchip’s PIC or its newfangled Amtel clones as your highest bang-for-buck choice of small simple microcontroller. There were older choices still around of course, the 8051 and 68HC11 will probably still be there after the cockroaches die from radiation poisoning.
So at this stage if you wanted to build a micromouse and were a beginner with just about a screwdriver to your name in terms of tooling, you were going to have to roll the whole thing yourself from the ground up, sourcing parts from dumpster diving and with zip-ties, the battery wires and the circuit boards themselves being major structural elements, as with the one Ian and myself built, Brandon:
So the idea with PICRat was to be something like the dozens of kits you can now buy off the shelf as a beginner. Unfortunately as a commercial idea it was horribly badly conceived – timed just as a wave of fully developed, well funded alternatives were about to land on the market (and in fact create the market), developed without any facilities or tools or manpower beyond one graduate so new he hadn’t even finished the graduating ceremonies yet, and done without any real plan or oversight or direction other than “go build a robot that a PC can connect to by serial port so students can do a micromouse project in a college course without having to do all the mechatronics legwork”.
Yes, serial port. RS232C baby. USB 1.0 is a year old at this stage but nobody is using it, and USB 1.1 (the first actually used version) is a year away. WiFi? Good luck, 802.11a only comes out this year, 802.11b next year and it’s years before that costs less than a grand in hardware. We have radio modems at this point and yes, that is exactly what it sounds like and it was an actual improvement because at least with those we didn’t have to go to ComReg to get experimental radio licences to use them…
Anyway, the spec I was given sounds like it’s complete enough… if you know the square root of nothing about building a commercial product (let alone a robot). And I knew just barely under the square root of nothing about commercial products at the time so I dove right in 😀 File that under “learning opportunities seized with both hands and all available teeth”. These days I’ve learned enough to back away slowly from the crazy person suggesting that I suffer for six months before the project gets dropped and its failure blamed on the nearest lightning rod (ie. the one guy working on building in three months with no tools what commercial companies were taking twenty man-years in fully equipped labs to produce).
Not a fun time. However, I came across the code for the project this weekend while playing with a PIC12 chip (more on that later), and hey, historical value at least and I’m pretty sure I still own the IP (no, seriously, it’s a whole thing I’m not getting into here, but shambolic is a pretty solidly descriptive term of that time), so I put it up on github for anyone who’s interested in how you hook a PIC 16C74 up to a PC and act like a state machine controlled over a serial line. By the way, that’s 16C74 – that’s how old this is, you had to burn the code into an eprom in the chip because the F series (that are all you can get these days) with the fancy FLASH memory were an order of magnitude more expensive…
Of course, it wasn’t just code, it was an inch-thick (2.5cm for us metric types) manual to learn, from scratch, for a non-von-neumann architecture (which college had failed to mention as a thing that existed in computing) and in assembly because the C compilers for 8-bit microchip MCUs were absent then and thin on the ground even now if you’re not able to drop a few hundred euro on one, and of course there was some wire-wrapping hardware fabricobbling:
And yes, this was before digital cameras were retailing at a cent apiece for the sensor, how did you tell? That’s a scanned photo from a normal film camera (I think we were doing a lab equipment audit at the time or something). It’s not an Instagram filter, it was taken with black and white film because that was still cheaper to develop than colour…
Top one’s the first prototype. I think the bottom one was the second prototype (it’s been twenty years and I can’t find my notes…)
It did eventually get used, by another new graduate equally haplessly roped in to do another robotics project without tools or anything else such a thing needs. He built his own board using the design and parts from PICRat and that was the end of the hardware:
And thus ends the saga of about five months of my life way back when. It was a lesson I didn’t learn at the time, which nobody explained to me at the time, which I’m now sure my supervisor at that point either didn’t know himself or didn’t know how to teach it, a lesson I have since learned the hard way a few times over. I guess that’s part of going from being an engineering graduate that knows the technology to being an engineer who knows engineering 😀
For those new hands who don’t have anyone to pass it on, let me restate it here (but anyone in that field can tell you as well) — build as little as possible yourself, and focus everything on that little. As Red Whittaker once told two rather new researchers who were wholly out of their depth and sinking at the time, you can do anything; you just can’t do everything.
So fathers day was coming up and dad’s just finished the first year of a law course so lots of desk time and books involved. And then I noticed this on accidental woodworker:
Well, a desktop shelf that’d fit in the corner and leave room to hide pens and such underneath should be about right, and I had a nice piece of sapele…
…yeah, no. Turns out sapele is a total pain to plane with a normal handplane because of interlocking grain. By the time I’d resawn it from 6/4 down to just under 3/4inch thick and then flattened the resulting planks I’m down to just a shade over a half-inch thick and that just doesn’t look right for a shelf. So I abandoned the sapele to future box-making duties, and ordered a toothed plane iron to deal with the remaining sapele in my store.
And a new lighter mallet for finer work (the lignum vitae mallet is great but was a bit heavy for working on things like half-blind dovetails).
And then I hauled out one of the last planks of walnut I had to hand, skimmed it and rough-cut it from 9×48 to 9×23 and 8×24, losing an inch of length to kerf and clipping off a rough end on the board and losing an inch of width to a bit of live edge on one half of the board. I also grabbed some oak I had and crosscut it in half, then laminated the two halves together to one 3×3 block; and then cut a diagonal line across the width of the board to give me the two feet – the thing you stand on, not the archaic unit of measurement – roughly 3×2 at the front and 3×1 at the back for a nice gentle slope so that the spines of the books are visible when sitting or standing at the desk.
Then on to the sides. I took the 8×24, cut that in half after I had it flattened and thicknessed down to about ¾”, marked off which corners I’d cut off for that sloped look and then marked out both for a stopped dado for the shelf (which would act as a sort-of-half-lap joint as the shelf would have two shallow rebates on either end to fit). I was planning on using the cut nails I got from Dictum a while back as a design feature so I didn’t make it a sliding dovetail, the nails would hold well enough.
I cut down the walls of the dado as much as I could with the saw and then cut out the waste and the rest of the dado with a chisel, and used a router plane to tidy it all up.
Next though, have to attach the feet to the uprights. Ralph used biscuits for his, but I don’t have a biscuit joiner (or room to use one or store one) so I just cut mortice and tenon joints, complicated slightly by the joint being sloped – the mortice is deeper at one end than the other and the tenon is trimmed as well to avoid chopping through the feet when making the mortice. This gives even more rake to the uprights – it’s not a huge amount, but it does give that slightly steeper angle without needing either the joint or the angle of the feet to take the full angle. It’s a little nicer and it means the spines of the books are at an angle that makes reading the titles easier if you’re sitting or standing at the desk (Dad’s desk is one of those electric standing/sitting desk things).
Not as hard to cut as the curved tenons in the cot a while back. And I did think of using my new morticing chisels but the one in the size I wanted to use has a handle that is literally falling off, and it weighs two pounds and looks like a railway spike. It’s more for deep morticing through a few inches of oak. For a blind mortice like this that’s barely an inch at the most, the firmer chisels are the better choice. Also, holdfasts. Best morticing workholding ever.
And with those fitted, time to sort the back rails, also from some oak I had.
They’re a bit thin, but that’s okay, this is for a desktop so “brick shithouse” isn’t really the design aesthetic I’m looking for here. I was thinking of doing mortice and wedged tenons here, but the walnut was a bit narrower than the sapele board I started with (so pushing the rails back means being able to hold wider books), and walnut’s far easier to carve, so I switched to the idea of using half-blind dovetails so the sides were kept fairly clean-looking. Those went very smoothly (it’s walnut, you’re basically cheating using it for dovetails), the only difficult part being the layout (because you basically need the whole thing assembled to get the final shoulder lines for the last two joints). But you’re only cutting four dovetails in total so it’s fast work.
With that done, that was the last of the joinery, and all that was left was shaping, small touches and fettling and finishing. I took the front corners off with the small ryoba and planed them ganged together to keep the sides matching, then took a small gouge and did some end-grain detail stolen again from Brian Halcombe:
Getting a little better, but still nowhere good as his are. I think I need to practice sharpening my gouges more 😀
I also took the fretsaw and my new preston spokeshave to the shelf to do some shaping work on the two ends (and got out the gouges here as well for the little bit of remaining endgrain). The shelf is a bit thick to avoid (a) sagging in the middle from the 60kg design weight (books weigh more than most people think, the standard loading is about 30kg/ft or so); and (b) thicknessing the board down too much because it turned out one side had the corner of a horrible knot in it and it was a complete pita to plane. But by planing a slope into the underside of the shelf in the last inch or two before you see it, it looks a lot thinner at first glance than it actually is – neat trick learnt from Richard Maguire’s end table videos.
You’ll notice I’ve also drilled pilot holes for the cut nails here. I bought a few boxes of those from Dictum a while back and haven’t had a chance to use them yet. They look quite nice:
For those who don’t see what the fuss is about and haven’t spent sixty hours listening to Christopher Schwarz, Roy Underhill and every youtuber with a table saw ranting about these things, they’re what nails used to look like for thousands of years until someone invented the cut wire nail a hundred years or so ago and found he could make nails that were objectively worse than the existing product in every single way and yet still be successful, so long as they were cheap. The race to the bottom is a very, very old game…
Anyway, these things look decorative and hold things together much more securely than the round nails we use today, but more than that, when the wood swells or shrinks and moves with the seasons, the nails flex with the wood which screws can’t do. This stuff lasts so long they’re still finding shipwrecks from the roman empire where the nails are holding them together.
Anyway, that’s my story about why it was okay to pay six quid for a hundred nails and I’m sticking to it 😛
And then I knocked the arises off the other edges with the preston spokeshave (that thing is rapidly taking over as my favorite tool), and with that done and everything suitably handleable (and the front of the feet turned from blocks to a less aggressive shape), it was time to do the final fitting and fettling of joints.
It was fairly painless this time. That shelf doesn’t rock or tilt 🙂 I’m a bit pleased by that, the layout was a bit hard because the wood’s not perfectly flattened (that stupid knot on the underside looks pretty but was a pain to work). And then final finish planing before the shellac. At which point I made the happy discovery that not only do we get some lovely colour to the walnut and some lovely medullary rays in the oak, but the walnut is also lightly figured. Which was a nice unexpected surprise, but is unfortunately most prevalent on the underside. Doh.
Well, only one last thing to do before applying finish…
Turns out, a real blowtorch versus a chef’s blowtorch isn’t even a contest. A chef’s blowtorch will do creme brulee, whereas one of these weapons of mass destruction will just burn through the sugar, the creme, the dish and most of the table. They’re excellent 😀
I did nearly burn the wood though. Mental note; branding irons only need to be that hot for flesh, for wood you want them less hot. Had to take the block plane to the endgrain to clean it up a little in the end. But that was that, and now on to the shellac. First a quick test…
And we’ll go with garnet for the walnut and lemon for the oak, with maybe a last coat of blonde? Brush on the first coat, then wipe on the next three, knocking back with 0000 steel wool in between each coat.
See what I mean about that walnut? Shame that’s the underside really.
And after a few days (I just did a coat every evening after I got home):
And then it was time for assembly and glue-up. That was a lot more straightforward than I expected too because the nails effectively acted as clamps. So out with the hide glue and I didn’t even need the hot water this time to warm it because today was around 26C in the shed, the warmest day of the year so far. Glue into the stopped dados, then seat the shelf in the dado and knock it firmly home with the deadblow hammer, then drive the three nails to within 2-3mm of being fully seated; turn it over and do the other upright. Then glue in the back rails because the uprights slope inward slightly and the tension holds everything in place. Drive home the nails fully, glue the mortice and tenon joint for the feet together, stand it up and clamp the lower back rails (the upper rails couldn’t be clamped because of the slope on the far side of the upright at their level, doh. Maybe I should have drilled the dovetails for smaller decorative nails, it’s certainly a historical thing to do that).
Left everything to cure for a few hours, then painted the bottom of the feet with titebond over the shellac, and sat them down onto some nice green felt (don’t want to scratch the desk) and let that set up for a few hours before trimming off the excess with a sharp knife. And that was it, all done.
The brand looks nice actually. I was afraid it’d be a bit out of place, but it seems to blend in discretely.
Kicking myself that that figure is on the underneath of the shelf…
But those nails do just look the part, don’t they?
Obviously I need to buy more Lost Art Press books 😀
And yup, there’s the original sketches and notes as well. Not an Ikea design 😀
Thanks again to Ralph at the Accidental Woodworker blog for the idea, it worked out pretty well.
So, lots of people who make stuff in wood (going back a few hundred or more years) made a mark on the stuff they build (assuming it’s good enough). Usually in an out-of-the-way place, even one where you’d have to disassemble the piece to find it; like signing one of two faces before you glue them together.
Sometimes it was a simple stamp made with a metal stamp and a few taps of a hammer:
Sometimes it was a paper sticker, and sometimes it was branded:
This wasn’t done out of the same sense that triggers graffiti by the way – some guilds from the 17th century onwards made it mandatory to mark every piece a workman made. Not every guild, but enough that it was considered commercial rather than odd or vain.
Well. Not vain for them 😀 For me, it’s pure vanity, but sod it, it was fun. And these days with 3D printing and CNC milling machines, it’s no longer a very expensive process involving difficult custom forging, it’s dropping €30 to €40 on a guy on etsy and giving him a design and waiting for a package to show up in the post three weeks later from Hong Kong. Living in the future has some benefits 😀
So here’s the design:
Not terribly big or fancy (by today’s standards – I suspect you couldn’t readily make this back in the 18th century). M for me, big C for herself indoors, in a venn diagram with a small c for junior in the middle, a common baseline and a random tangent forming an acute angle. So it’s uber-smarmy-symbolic is what I’m saying 😀
But who cares? I have my own branding iron 😀
Simplicity itself to use – hold by handle, point blowtorch at head until hot enough, press onto wood and rock back and forth slightly, quite gently, and you leave a mark in about ten seconds. I don’t think it’d cope with any kind of high rate of production, but that’s fine, I don’t produce stuff very fast 😀
Yes, I’ll brand that sapele, no I won’t be doing that today. Have to finish working on it first.
That’s not too bad, it’s almost discrete 😀
Now I just need to buy a bigger blowtorch. The creme brulee torch isn’t cutting it 😀
Finally got the last of the tools off the bench and onto the wall by building a small holder for the drills from a hardwood offcut and some recycled plywood. And managed to get a rebate and some stopped dados in as practice 😀
The shed is now finally looking like a proper workshop. I mean, it’s not done yet – the western saws need to go up on the wall too (another magnetic tool rack for them I think, and mount them below the japanese saws – lidl didn’t have any left so ebay it’ll be), and I do need to do something with the grinder and the airbrush, which will probably mean french cleats of some kind and probably redoing the mounting board for the grinder. But for now, it’s not too bad.
Yeah, okay, there’s a huge mess to the right hand side. That’s a bunch of boards rough-cut down to components for some small tables I’m going to build, and they’ve accumulated crap on top of them as any flat space does. I’ll have to tidy up the shed during the week.
But at least now I can work on something that’ll actually leave the shed instead of living there until I recycle it 😀
Oh, and a new moisture meter and thermometer as well. No huge reason for one, I just came across it while buying something else on aliexpress and it cost just under two euro shipped so what the hell.
Stochastic Geometry is Stephen Fry proof thanks to caching by WP Super Cache