...but i find that 15w40 and nail polish remover goes really well with my salad.

i'm going camping tomorrow, and i like reusing containers (well-cleaned containers...). obviously i also like making silly labels with my label maker.

after reading my recent post about my new car one of my austrian friends asked for more details; not exactly surprising as utes like that are not exactly common in austria.

so here are some more photos and phacts.

for comparison here is a photo of my old subaru outback, an average length station wagon ('kombi'). the colorado just fits into the garage, ie. iff i drive up to and carefully bump into the wall. stock length is 4.9m, add about 25cm for the bull bar on mine.

stock weight is about 1920 kg (cf. 1460 kg for my previous car, a forester), but this one lugs around a bull bar, winch, canopy and dual battery; installing a lift kit is very much in planning.

it's got about 120 kW and lots of torque thanks to the turbocharged diesel; agility is...well, this is not a race car. it doesn't feel underpowered but there's the inertia of 2 tons of car.

(as in sewing, not faecal matter)

believe it or not, but i do actually own a sewing machine - and i can even operate it (if not exactly well).

it's a relatively old elnita 150, and the only electric bits in it are the motor and the light bulb. i do admire the mechanical design of mechanical sewing machines: two cam drums, a comb full of cam followers, a few levers, a bunch of springs. in this machine that's enough for 15 different stitch patterns.

however, mine doesn't get used often. today i wanted to prep it for some upcoming fiddly fabric work, only to find out that it would only zig spastically, not zig and zag.

after applying occam's razor to isolate the involved ziggy bits it turned out that the issue was just stiffened old grease and/or insufficient lubrication: one follower lever had gotten too sticky to return properly when released. for zig that one gets pushed but for zag it needs to return under spring tension, which it didn't do reliably.

the solution was trivial; a bit of fresh light oil, some soaking time and vigorous exercise of the mechanism and it's all working again. me happy :-)

following up on an earlier post, here are a few more reasons why 3d printers are both cool toys and useful tools.

i live in a pretty humid climate, and using vacuum storage bags (for things like spare blankets and pillows) is quite important; but the dyson vacuum that i inherited from my daughter has this nice-but-unhelpful clicky connector that sucks because it doesn't suck -- there's no flat interface that you can press against the bag valve.

so i spent a little time on designing and printing a sucker adapter (in PETG because i wanted to do more testing with the material).

or this one, from earlier this week: the built-in cupboard in my hallway has a broken door catch (cylindrical post in the frame, claspy catch on the door) and i couldn't find any even remotely similar replacement at the (sole remaining :-( ) hardware chain.

however, calipers and persistence and one failed test-print later i've now got a parametric model and an actual replacement part that works.

on the last photo you can see my newest mod to my printer, a mk52 (clone) magnetic heatbed. the print surface is PEI on a removable sheet of spring steel, which is held to the actual bed and heater by many strong magnets. when your print is done you take off the steel sheet and flex that, rather than prodding and prying with spatula/chisel/knife.

so far it works pretty well, but the bed is made from PCB/fibreglass and prone to warping. i haven't fully bolted the bed down (like official prusa does it) because i like the ability to level things manually, but i may want to change that later; for now i've setup 7x7 grid level compensation with my smoothieboard clone and that takes care of the imperfect flatness.

i'm now driving what might conceivably be called a chick magnet car -- with the understanding that the magnetic attraction is confined to within the cabin, and the repelling forces work all over the outside. net result: it'll take a woman of superhuman persistence to actually come close...
click here for the rest of the story...

(beagle as in 'beaglebone black', which is not a dog by any means)

almost exactly three years ago i built a minimal-budget online weatherstation for fido, john sinclair's fraser island defenders org. that station was installed at happy valley and it's been working pretty well ever since - well enough that we followed it up a few months later with another station which ended up at eurong.

both of these were build on a shoestring budget, and for the second i used the same fairly yucky 'authentication' chip setup on perfboard and hacksawed 2mm 2x10 pin connector, and everything was housed in a really ugly fashion inside a weatherproof box.

recently fido got a budget together for four more stations. about two months ago we started acquiring the bits and pieces for these stations, and this time i decided to make everything a bit nicer and easier to assemble.

this new litter of beagles will be housed decently: i designed and 3d-printed a custom enclosure that attaches to the back of the weather station console. thanks to the odd geometry it was a bit annoying to print but the resulting four dog houses look good and work really well.

this time i've also decided to 'design for manufacturability' (bwuahaha - translate: i wanted less messy manual soldering and no connector butchery). this meant switching to a different pic microprocessor, an 8-pin PIC16F18313, and making a printed circuit board with keyed connectors to make everything a bit more fool-resistant.

the 16F18313 is a little powerhouse, and i found it amazing how much functionality microchip crammed into this chip (datasheet for the 16F18313: 471 pages. 16F88: 228). at au$1.31 it's also much cheaper than the older PIC16F88 i used for the first two stations, and thanks to freely reassignable pin functions it's much easier to route a single-layer pcb for that processor.

but getting it to actually work was immensely painful: none of my infrastructure would deal with this fairly recent chip. my version of xc8 wouldn't compile for it, pk2cmd would not program it (nor would mplabx talk to my pickit2). lots of swearing and fiddling later i've got a working PK2DeviceFile.dat for that chip, and proceeded with the pcb making.

after a little time with eagle (and a brief detour to build a small drill press for drilling the circuit boards) i made these four boards in the most pedestrian fashion possible (read: using the toner transfer method). soldering on the smt connectors was easy, using solder paste. and everything did work the first time round :-)

in about a month the four stations will get installed on fraser island and i'll post an update when they're live.

you can find the updated code for the new pic and the board design (in eagle format as well as printable image) at https://github.com/az143/davis_weather.