24 November 2015
A new project that I'm working on requires a two axis gimbal that holds the item in the stationary while the rings around it move. It's made from 3 layers of Birch ply with the middle layer placed at right angles to the outer layer to keep as much strength in it as possible. The whole thing can only be 9mm thick when flat so I had to embed the bolts into the middle of the wood (bit hard to see now) and the rotating side of the gimbal uses an acrylic insert to minimise the friction while rotating. It's definitely functional but needed a few more tweaks for my intended purpose.
23 November 2015
21 November 2015
I was asked to make a magic Santa key, personalised for my friends children. Ta da. Not sure there's much else to say, it's 3mm mdf painted with metallic acrylics and it's a good job our kids aren't old enough to be reading my blog yet :)
19 November 2015
Bit of a slow day today, few bugs going round our house and extra children after school, the only thing I managed to get done was this batch of fairy doors for the local school craft faire. I added a few new elements into the mix (mushrooms and stockings etc) and changed the blue out for purple. I've got enough elements now that no 2 doors will ever be the same which is a nice touch.
18 November 2015
The laser is very good at whipping up a batch of gears for steampunk decorations. I've done gears before but it's actually quicker to draw new ones than the find the old files. (svg here)
17 November 2015
Last in request and completion. This spoon loitered at the bottom of the todo list so it's not a major surprise I overlooked it while doing the other trophies. It's a standard engrave for the logo and low power cut for the text, I am quite surprised at how well it came out though consider it is all a curved surface. Clearly a few mm in either direction of focus doesn't make a vast difference, I doubt the same could be said if I was trying to cut all the way through.
16 November 2015
15 November 2015
13 November 2015
12 November 2015
10 November 2015
Tonight I've made some proxy bases by request. When you have multiple models in close proximity during a match there is a chance they can bump into each other and become scratched. You can replace the model with one of these proxy bases that keeps track of location and angle for the model. I've also done some objective tokens and some flag markers (not pictured) to complete the set.
08 November 2015
07 November 2015
I wanted some price tags for my craft faire, as everything can be laser cut these signs were a prime candidate. The text was done using the Hershey Font Generator now included in Inkscape, it makes text as a series of single line strokes which has a very neat effect when cut on the laser.
06 November 2015
04 November 2015
Star trek style Com badges for a group of friends who were also going to be playing Artemis. Instead of pretending to be Humans of Earth, they were the Citizens of Klunge and they have rather interesting shaped space ships. Maybe the next time you see the crew of the TNG ejecting their warp core at the last possible moment you'll have a little chuckle too :)
03 November 2015
02 November 2015
01 November 2015
31 October 2015
Halloween means costume making and now I have 2 kids that naturally extends to them. Eli wanted to be a scary robot, I ended up settling for Retro robot because we have cardboard boxes and vinyl pipes floating around the house. Topped off with egg boxes, bottle caps and yoghurt pots, I wouldn't say this build was challenging, time consuming or detailed but I think that's the main selling point. Eli was dead keen to help at all stages and there was an awful lot of love for it once we were out and wandering around the streets.
but probably not enough to satisfy @MarkPhelan, Just a quick Halloween silhouette to decorated the garage doors. A long chain of neopixels provides some changing colour back lights.
30 October 2015
We were inspired by the Glowforge laser cutters ability to continuously autofocus it's z axis. A camera mounted in the head of the laser detects a laser dot being shone onto the surface of the material and from there it can determine how far the lens needs to be adjusted to set the focal height for the correct cutting distance.
The Nintendo Wii Remote has an optical camera in it that is used to detect up to four points of infrared light. The hardware automatically identifies these points and feeds back XY positions through a bluetooth connection. The cutting laser on a laser cutter is an infrared beam, as it cuts through the surface of the material there is a moment where it is reflected off the material and the Wiimote is able to detect the location of the cut. The location data is fed back to a laptop and by comparing this point against the initial 'in focus' point we're able to detect if the Z axis needs to move up or down and by how much. The laptop sends data to an arduino that is connected between the laser cutters onboard Leetro controller and the Z axis stepper motor driver. The Arduino controller is the same as we have used on previous projects to control the Z axis. The result is a laser cutter that has the ability to remain in focus throughout the duration of a cut.
The project was written using C# and Brian Peeks Wiimote Lib, which made it incredibly easy to connect to the Wiimote using just a few lines of code. The source is available here but be warned it is a horrible mess of gaffa taped code written for a project that could never be more than an interesting hack.
- Our laser cutters move the whole bed up and down to adjust the focal height. There is a lot of mass to move and it gains momentum so it isn't able to adjust the Z axis as fast as required. This is why it's so noisy as it tries to keep up with the requested position.
- The camera in the wiimote is capable of detecting light sources 60 times a second, the laser cutter is not able to process this information fast enough so there is some lag between the height change and the actual movement.
- The Wiimote is only able to detect the focus height while the laser is actually cutting. If the laser stops cutting and moves to a new area with a different height then the system will take a second to adjust to the new height once it starts cutting again.
- The increased mass on the cutting head prevents the laser cutter from reaching it's top speed of 50mm/s, and even if it could the response time from the system would not be fast enough to cope with changes that quickly.
26 October 2015
Every so often a project comes along that would have been the envy of my childhood self. This was one of those projects - I was asked to build a 8 lane pedal powered Scalextric. This meant putting together Scalextric track designs and testing Scalextric cars.
I was really pleased with the end result, which was built within a 10 day timescale. This blog post shows the design stages.
22 October 2015
I am falling behind on this years large projects, I've been working on projects 9 and 10 simultaneously and although neither of them are ready yet this one is slowly coming together. I'm making an Bar top arcade cabinet for myself based on the Clarkcade cabinets. I just got the buttons into the front panel to see what it would look like while I'm waiting for some vinyl wrap (to hide those screw heads).
21 October 2015
Cyclops Pedal Power, based in Leeds, kindly invited me to work with them on their pedal power equipment.
They wanted to record the power and energy from their three-bike set-up, and have a system that can be expanded for more and more generators.
We had a three day workshop based in the fantastic Pedallers Arms working on monitoring voltage, current, power and energy.
20 October 2015
A quick and simple speech bubble to be held aloft in wedding photos. The inner paper is actually a mask because the bubble will eventually be chalkboard black with a white edge to make it stand out. Guests can then chalk on a new message as they wish. Sounds like a fun idea for wedding photos.
19 October 2015
My friend Steph does some wonderful pen and ink drawings which she posts up to Facebook occasionally. Somehow I missed this one way back in February but as soon as I saw it I knew it would make an awesome lasercut. The artwork is all hers, I just vectorised the lines, by hand, to create some fancy line art. Automatic conversion would make 1 line either side of the pen marks in the drawing and I knew it had to be a single line per stroke to look right.
It's made from 2mm ply wood, Mahogany and Oak and backed with 0.8mm birch, I think it would make a nice brooch. Next time I might do the chin in Oak too.
18 October 2015
I was asked if it was possible to engrave Wax candles in the laser. So I grabbed a pillar candle from the nearest shop and whacked it under the laser using the rotary attachment. We gave it several attempts slowly turning up the power and down the speed and the answer is 'No' wax just melts and bubbles away without really engraving.
I'm sure if we have a silver or gold candle, those are only coloured on the top few layers and you could probably cut that away but plain colour candles don't work.
If you are a serious British maker with a project you’d like to promote or commercialise or even if you Continue reading
16 October 2015
14 October 2015
So, before I dive in to this, a quick history lesson for those of you that aren’t fully up to speed on your rare 80’s vintage 8 bit computers. In the beginning, there was Clive Sinclair, and he invented the Sinclair ZX80. From the ZX80, the ZX81 and indeed home computing was born. From the ZX81 came the ZX Spectrum. Ok, so Clive didn’t invent these single handed. He had a team working for him, which included Steven Vickers and Richard Altwasser. Somewhere between the initial design of the ZX Spectrum and its launch, Vickers and Altwasser thought there was a better way to do things, and left to set up their own computer company.
Jupiter Cantab was formed in 1982, and Vickers and Altwasser developed the Jupiter ACE. This little computer was a weird mix of ZX80 (vacuum formed white case), ZX81 (black & white low res graphics, 3k memory) and ZX Spectrum (rubber keyboard, small speaker). The one fundamental difference, however, was that it used FORTH instead of BASIC as it’s operating system. FORTH was much more efficient than BASIC, and would revolutionise the home computer market…
Although history tells us it didn’t. Cantab went bust after 18 months and was bought by Boldfield Computing, who went on to sell off the remaining hardware in 1985. Approximately 8000 units were manufactured.
When I was offered an original unpopulated Jupiter ACE by John Fletcher, it was too good to turn down…
The PCB itself was typical of those from 1982. All the tracks had been laid out by hand, with different size pads (depending on the stencil they had to hand?), solder mask on the underside only and no silk screen. There was quite a big scratch on the underside, although this checked out to be ok, and the board was in otherwise pretty good condition. Unlike the ZX81 or Spectrum, it used standard off-the shelf components (Except the ROM – more on that later)
The Jupiter ACE has 3k of RAM, (1k user space and 2k video memory) uses 6 2114 ICs, which are pretty rare now. Luckily, John had a few of these spare. I had pretty much all the other components, with Nottingham Hackspace donating a couple of 74LS logic chips and EPROMS and a small order to RS for the remaining items. As a variation to the original, I decided to use chip sockets throughout – I didn’t need to be as penny-pinching as people were 30 years ago!
Getting hold of a definitive circuit diagram wasn’t particularly easy. Several schematics exist, along with high resolution photos of original models – however, they all contradict each other in some way or another. Most differences are fairly trivial, some are improvements and some, well, are just different. I mainly went for info from Jupter-Ace.co.uk Grant Searles website and instructions from a Hungarian rip-off of Grants info
After some of the essential bits were put together, it was time for a few tests
Even without any RAM installed, the video circuit could be tested. By looking at the output on a scope, it was clear that video signals were being generated. Below is a shot showing a couple of lines and the blanking signal. Things were looking good!
The original Jupiter ACE used an 8k ROM comprising of 2 2532 4k chips. These are nigh on impossible to get hold of these days, but have a very similar pinout to the industry standard 2732, although it requires a little modification to the PCB tracks. It seemed a shame to have to cut these, but there really is no neat way to get around this. The other problem I ran in to was that I could only find a copy of the ROM as a single 8k file. I assumed it was simply split in half with 0x0000 to 0x1FFF on one chip and 0x1000 to 0x1FFF on the other, but couldn’t find confirmation of this. (In case you’re using this blog for reference – that is, indeed, the case!)
Things didn’t work though. The screen just filled with garbage! I wasn’t sure if this was a problem with the ROMs I’d burned, the memory, any of the old components I’d used, my soldering or anything else. After comparing signals with Johns fully functioning Jupiter ACE, it was tracked down to the clock to the CPU not being a strong enough signal to get through. This is amplified by a transistor with a couple of resistors which is one part of the circuit that’s different depending which source you look at. I fixed this with a small pot so I could tweak the gain to the sweet spot.
This gave me some success, although often only for a few seconds before the computer reset itself. Much head scratching and logic analysing went on before this was diagnosed as a power supply issue. I’d been using a ZX81 power supply, and this is rectified down to 5v by a 7805 onboard the ACE. If I bypassed this and just fed a regulated 5v from a USB adapter in, it worked fine!
The modulated UHF signal looked ok on my TV, but I hooked up a simple composite interface, and this gives a much sharper picture. So far, so good…
As you can see, the keyboard is made up of PCB tracks that need a metalised rubber keyboard to make the circuit. Original Jupiter ACE ones are not available, and I don’t think I could put anything together that would both work and do the old thing justice. So, instead, I’m working on something to offload the keyboard function to a regular PC keyboard. Look for more posts on this as it gets done.
13 October 2015
I was asked if I could make the Derby Mini Maker Faire badges as shown in the attached drawing. I think I made a pretty good job of it :P
11 October 2015
10 October 2015
09 October 2015
|Pendula held in the framework to get the right string lengths|
|The whole flat pack kit|
08 October 2015
Recently, though, I’ve been trying to streamline my use of the shell in Emacs. Here’s my Halloween Emacs tips.
Added support for multiple shells
The first, glaring problem with shells in Emacs is that you can only have one.
I added a function that creates a shell based on the remote host of a buffer. It also names the shell after that host, so if I’m working on
/example.com:/some/file, I’ll get a shell called
*shell:example.com* (instead of just
*shell*). This allows you to have multiple shells open simultaneously, one per host. [Credit: EmacsWiki]
(let* ((tramp-path (when (tramp-tramp-file-p default-directory)
(host (tramp-file-name-real-host tramp-path))
(user (if (tramp-file-name-user tramp-path)
(format "%s@" (tramp-file-name-user tramp-path)) ""))
(new-buffer-name (format "*shell:%s%s*" user host)))
(shell (if host new-buffer-name buffer))))
Added F9 shell toggle
Then I added the F9 global key binding to toggle between the current buffer and the related shell for the current buffer’s host, creating that shell as necessary.
(if (string-match "^\\*shell[\\*:]" (buffer-name) )
(global-set-key [f9] 'my/shell-toggle)
Handle shell termination
There are several ways you could handle terminating the shell. You could, just for shell buffers, turn off the prompt that asks you if you’re sure you want to kill a buffer with an active subprocess running in it, and then use C-x k. But this would kill bash rudely, and bash would not save your command history. You could tell bash to save command history after every command. But this leaves executed commands between different shells all intermixed in the history file and shells would no longer have their own individual session history. That doesn’t appeal to me.
So I opted for another method: detect when the shell terminates and close the emacs buffer. This way, you can hit Ctrl-D (same as you would in a terminal window), bash saves its history, and everyone’s happy.
(set-process-sentinel (get-buffer-process (current-buffer))
(defun my/shell-process-sentinel (process state)
;; show shell status and kill the buffer automatically
(message "shell: %s" state)
(string-match "exited abnormally with code.*" state)
(string-match "finished" state))
07 October 2015
|Cut the outline from the greyboard|
|Cut down the lable and stick it to the greyboard|
|Put the sticker back into the hole|
|Cut the individual tokens out|
06 October 2015
I was walking home from school with one of the other parents, she was complaining that people didn't know where her front door was and keep going round the side of the house. 10 minutes later I returned with a sign for her, 250mm long and engraved from HIPS plastic so it's weatherproof. Lasers are awesome.
I've been working with light artist, maker and NESTA fellow Jo Fairfax, (of Jo Fairfax Studio) on an interactive wall prototype.
He needed to have control over loads of motors which all move circular disks a set distance.
I designed a stepper motor controller unit to control up to 12 motors. These can be repeated to produce very large arrays of stepper motors.
The final effect is amazing, as shown in the videos here.
05 October 2015
|We're going to need a bigger Baymax|
04 October 2015
03 October 2015
A second mechanism for Vale LRP and another file that has been floating round the internet. This Iris is a neat piece of engineering that has been built by dozens of people. I rejigged it slightly to put some better gears around the outside but other than that it's the same. Instead of using nuts and bolts to hold it all together it is all made from wooden parts, each pivot point is made from a 6mm wooden dowel. (svg here)
02 October 2015
I needed to make a dohicky for Vale LRP and I wanted something quick and easy so I trawled the net and found these non circular gears of Thingiverse. A bit of steampunking later and I had something neat that would fit into the genre. I've included my files here which are a bit of a mess but at least they will open in inkscape now (svg here)
01 October 2015
29 September 2015
This project was remade from the first prototype and it was ready for my Cambridge workshop on the 13th, it's just taken me an additional 2 weeks to actually write about it. The principle is very much the same, I have fully enclosed the range with perspex to make it better at keeping the ammo in the range but also added a few other features.
While in operation balls were getting stuck behind parts and we were having to remove the framework to get a hand in there and free parts up, I removed most of the natural ledges to limit this from happening and now balls only get stuck behind the targets. I made the rear third of the perspex screen an access panel. It can be lifted high enough to get a hand under and the balls can be moved as appropriate.
When not in use the catapult range is largely empty space, there is a big volume that could be minimised for transport and storage. The rear wall and lid can be removed, the two perspex side walls are hinged so they can be folded into the range. All the targets go flat so they aren't in the way. The whole range can be stored vertically and carried under one arm which should make it easy to get in and out of maker faires. Come and try it for yourself at the upcoming Derby Mini Maker Faire
28 September 2015
27 September 2015
Not every project can be a success and there is a popular mantra that you should fail and fail fast. You learn from failures and the sooner you get them out the way the sooner you can move on to the right ways. I don't have a lot of failures, I have a lot of projects that get put on hold until I know how to progress them, some of them are on indefinite hold at the moment. Tonight's project has been on the cards for a few months but it failed and relatively quickly so I thought I would write about it as I'm no longer working on it.
A lot of Chinese laser cutters come supplied with the Leetro controller MPC6515 or MPC6535, both use the PAD03 display panel to communicate information to the user. The system has a fair few quirks but the one that really bugs me involves the Z axis adjustment. You press 'Z' to move into the height adjustment mode but pressing 'Esc' does not get you out of the mode, you have to press 'Z' again. The idea is simple though break into the loop between the panel and the controller, catch the 'Esc' key as it is pressed and turn it into a 'Z' key press so that the controller leaves Z mode.
With everything connected and the data relayed back to the PC I turned the system on. Data was successfully passed through the Arduino to and from the controller, the system functioned normally despite the Arduino in the loop. Pressing buttons on the panel created additional data on the screen so it should have been easy to spot the key presses. I switched the panel into z mode and started pushing buttons. Only the 'Up' and 'Down' buttons were actually passed back to the controller, none of the other buttons were.
The conclusion is fairly simple. The panel understands that it is part of a laser cutter system, it's not just a dumb terminal reflecting information from the controller. It is the panel that switches the display into 'Z' mode and it is the panel that prevents it from coming out again. The project was a bust but I learned more about the system in the process and that is useful in itself.
The catapult has been around for a long time at various scales and versions, I recently sold a dozen to someone for a wedding and realised that the instructions were lacking a little. I redrew them attempting for a new words Ikea style, they're not as clear as they could be but I think they're acceptable and I'm pleased with the artwork anyway. (pdf here)
26 September 2015
16 September 2015
07 September 2015
06 September 2015
I'm falling behind with projects, possibly due to having 3 maker faires and 1 LRP event in a 5 week period meaning my August disappeared. This was a spur of the moment project for Brighton Mini Maker Faire at the weekend just gone. It went well though and the kids really loved it so it's now undergoing a complete redesign to make it stronger and sturdier, which I'll need in time for my Cambridge Hackspace workshop next weekend.
I've taken my mini catapults to maker faires for quite some time now but they tend to get overlooked especially when we tell kids not to fire ball bearings at us. This time round I thought it would be cool if we had something they could actually shoot at, we could then cut mini catapults on the laser and hopefully sell a few more of both. I built a small range on a slope with 10 targets to shoot at, the idea being that the balls would roll back to the kids and a small motorised mechanism would stand the targets back up again. Testing and construction was swift and I had most of a range to take along with us to Brighton.
The Ensuing Carnage
Brighton Maker Faire was good as usual and pretty busy (although maybe less busy than the previous year). We had children swamping our stall from the moment the catapults were working 10am until they called time at 5pm. The whole range took a battering and I learnt a lot about how these things are going to break. Most importantly it was a success and that is prompting me to remake the range.
- MDF, this was probably my biggest error. when you glue 2 surfaces of mdf together the glue only sinks into the mdf a little way. If you put any serious force into this the mdf just delaminates and the whole thing comes apart. This was the fault behind, targets falling off, hinges being torn from the board and the whole 'hill' coming loose. The remake will be made from mdf, but all the joint will be made from sheets glued at right angles to each other.
- Gravity. The balls are only 2 circles held 90 degrees apart so they don't roll very well. The slope could do with being a little stepper to encourage the return back to base.
- Gaps. A massive mistake on my part, I forgot to close the gap under the hill so after the first wave nearly all of the balls were stuck under the hill. There were also ball sized gaps in lots of other places that should be eliminated, in front of targets were the worst because if a ball wedges there then the target won't stand up again.
- Netting. The cage to hold the balls in was made from netting around a pipe framework. IT wasn't a particularly close fit to the board so balls were disappearing under the board all the time. Also every time we tried to move the cage it snagged on the board to it was a pain in general.
- Balls. I took 50 balls, this was actually sufficient, I've probably got 30 left over. 100 balls would have been better though. It also would have been better to release them 25 at a time throughout the day, kids get through all the balls present super fast. Less balls just slows them down a little which is less effort for us to free them up etc.
New wood is on order, I've booked a slot on the Nottingham laser to cut all the bigger parts I require and I'm hoping to get this whole thing remade in just a week. Watch this space.
04 September 2015
German candy company Katjes has joined forces with organic confection entrepreneur Melissa Snover to create a “The Magic Candy Factory”, a 3D printer that creates custom gummy candy. In addition to having an awesome name, the device enables users to design the gummy candy of their dreams using an intuitive iPad-based interface. The candy is heated during the printing process and then hardens as each layer cools to retain the 3D design.
“While the ingredients are confidential, Snover confirms the candies are lactose, gluten and gelatin-free, and entirely vegan. Delicious designs can be printed from seven different colors and 10 different flavors including mango, apple, blackberry, and extra sour flavors.”
Watch this video to see the mesmerizing Magic Candy Factory in action:
[via That’s Nerdalicious!]
Aw man, someone cleverer than I actually created the thing I was thinking about.
03 September 2015
Desert Fury, pretentiously draped over a Macintosh
Desert Fury, close-up
Desert Fury, before blocking, draped over a table
Desert Fury, as it is being blocked
This is Desert Fury, my new scarf, knitted in some beautiful DK superwash Merino called “Furiosa”.
It took me awhile to figure out what to do with it, but then I found the perfect pattern on Ravelry - Here Be Water Dragons.
This pattern is perfect - it’s complicated enough to keep me interested, but easy enough to do without having to repeatedly check the pattern, especially on the bus. The yarn was so soft and beautiful to work with, the pattern went so smoothly, and it was an absolute delight to knit up.
I did change the pattern slightly - you’re supposed to repeat the main part 11 times, and I knew that’d use up all my yarn, so I just repeated it 5 times and had a little left over to make something neat. And it still ended up being ridiculously long. I blocked it on 5 of my 60cm square mats, and I thought that I’d have to pull out a 6th, so we’re talking 3 metres of awesome scarf.
This is definitely going to be my go-to for when I need to knit a scarf for someone.