We took a bit of artistic license with the design, but it is based on the screwdriver of the 10th Doctor. (see http://www.thinkgeek.com/geektoys/cubegoodies/8cff/)

The piece was turned out of aluminum and the final version has LEDs inside for lighting. Machining took about a week’s worth of afternoons, leaving me with a day left to make my costume :p

Build photos after the bump!

The piece started from a piece of 5/8″ aluminum round stock. It is made in three pieces (left to right in top photo):

• Head piece
• Collet
• Barrel
• Tail

Head Piece:

Starting the head piece:

Cutting the decorative features:

Aligning the head piece in the mill to cut the windows:

Cutting the windows with the mill. The brass tabs are to protect the turned finish:

Finished head piece. A glass bead will be glued to the top later. We chose to forgo the “crown” on the top to save some time.

Collet

The collet piece attaches the barrel to the plastic tube. I don’t have a good picture of it by itself, but this pic shows it fairly well. The collet is attached to the barrel with two screws:

Barrel

The barrel was the first piece to be started and the last one to be finished. It holds the batteries and LED that lights up the head piece. Most of the machining operations – boring, cutting the slot, drilling and tapping set screw holes, etc. – were done before turning the outside to avoid damaging the finish. The most interesting aspect of this piece is the mandrel I made to hold it when finishing the outside:

The mandrel consists of two aluminum cones and a center shaft made from a 1/4-20 bolt. One end of the shaft was held in the lathe chuck. The barrel was held between the cones (tightened with a nut and lock washer) and the right end of the mandrel was supported with a live center. This was the first time I had used a center with a project and the setup worked like a charm!

Tail Piece

At the bottom of the barrel is the tail piece. This holds the batteries in an provides some artistic decoration. I didn’t want to spend the time trying to thread on the lathe, so the tail piece is made in two parts – an outer aluminum shell and a threaded insert cut off of a 1/2″ bolt. The bolt piece has a 4-40 hole on one end and a small screw holds the two together.

Conclusion

All in all, I really enjoyed this project. Unfortunately, the friend who I was building this with moved to Seattle recently, so I can’t go back and take a nice set of glamour shots. I’ll be sure to do that early for my future projects… If you have any questions or want pointers on your own Sonic Screwdriver, feel free to leave a comment below!

Planning:

Seeing as how [Evan's] rep rap is still being built and there was no time for drafting besides, I couldn’t go the 3D printer route. It was easy enough to start from scratch and I had some components lying around

I thought about a couple of lighting sources – discreet LEDs, a cellphone backlight, surface mount LEDs, and EL wire – but discrete LEDs seemed easiest given the timeframe. The backlight would have also been cool, but would have required a color gel. As for power, I was originally going to power the LEDs with a watch battery, but I wanted a way to switch them on and off and also to be able to replace the battery easily. A single watch battery was more than sufficient to light all the LEDs, though:

Construction:

Using PVC pipe for the case was an easy choice given my past success with sensor node enclosures. I cut a 0.75″ wide piece of 3″ PVC and painted it silver. The tube around the outside is a piece of 0.125″ ID vinyl hose. I bent a piece of brass rod to fasten then ends.

To light the prop, I used eight high-intensity blue LEDs with 1K resistors around the edges of the case just like the demo I made for OSHW ’11 back in September.

The silver paint did not stick very well at all to the PVC, so I had to re-glue several of the LEDs. Using primer would have been a good idea.

The positive leads were long enough to reach to the center, so I just soldered them together there. A piece of copper tape helped bridge the gaps. The ground bus was wire-wrapped.

To make the “arc tube,” I fastened magnet wire to the vinyl hose with copper tape and wrapped it around a few times. This was a perfect use for an old spool of wire that had come undone and gone all “rat’s nest.”

To diffuse the light, I glued a piece of printer paper to a lens cut from a scrap of polycarbonate(?) sheet.

I made the little center bit from a steel washer, driller alternating holes for decoration:

The back of the piece was corrugated carboard. I glued some aluminum foil to it to reflect light towards the front.

To wear it, I just glued to the piece to a shirt and poked it through a gap in the buttons of a top shirt. I had to pin the top shirt just underneath to add some extra support, but it worked pretty well:

The wires lead to a 3 AAA battery case in my front pocket. Popping out the middle battery was a perfect “switch”!

BOM:
The whole build was completed in less than 5 hours, with a minor re-glue after an hour of walking around with it. Out of pocket, I spent $1.20 on resistors at Radio Shack. The total bill of materials was less than $10.

• 3″ PVC x 0.75″ – $0.50
• 3″ strip of aluminum foil – $0.07
• 8x blue LEDs – $2.50
• 8x resistors – $2.00
• sheet of paper – $0.05
• magnet wire – $0.01 (?)
• copper tape – $0.01 (?)
• Battery case – $3.00

I was super excited to get a shiny new kit of Maker Beam today, so I thought I share my first impressions of it. Read on for the details!

Maker Beam is a “Mini T-Slot” aluminum extrusion. It is very similar to other products like 80/20, but Maker Beam extrusion is 10mm x 10mm. The small size makes it perfect for small projects like enclosures, small fixtures and small to medium robots.

One of the big attractions of Maker Beam is that it’s T-slot profile can fit a piece of cardboard, plastic, or a circuit board. Since this was the main reason I wound up buying it instead of Microrax, I was extremely disappointed to find out that it doesn’t actually fit pre-made circuit boards that well. The t-slot is roughly twice as wide as the thickness of a standard PCB (shown below next to a Maple Pro), so boards will not fit in snugly. There is an indentation in the bottom of the slot that is designed to cradle a PCB, but it’s too far back to hold parts that have components near the edge (eg. Arduino). Proto/perf boards and cardboard from boxes fit nicely, though.

The kit comes with a nice assortment of pieces and lots of hardware for putting things together. As numerous trolls commentators on Sparkfun point out, the angle brackets are marked in Tau radians, but there are only three angles to choose from, so it shouldn’t be a problem to pick the one that fits where you want it to.

Of more importance is the selection of length of pieces you get. There are plenty of pieces in the kit and they’re sized appropriately to match the included brackets. There is about 5.78 meters of extrusion in the kit total. You could probably get a lot more for a lot less, but the kit does come pre-cut, which is worth a lot in terms of time and quality of cuts. That said, the ends of the pieces are unfinished bandsaw cuts. For $130, it would have been nice to have finished ends. Or at least de-burred ends…

Another problem with the Maker Beam system is the choice of capitve nuts vs. captive screws. In Maker Beam, the screw heads slide into the T-slot, leaving the nut and a bit of screw hanging off the connector. Personally, I don’t think this looks very nice in a , though I imagine that it would have been much more expensive to get custom nuts to fit the profile. Someone on Sparkfun suggested that you could drop a project and damage the screw threads, making it impossible to get apart. I doubt that (worst case you strip out a screw and have to toss it), but I am concerned about the screws getting snagged on stuff like carpet and wires…

All in all, I think that Maker Beam is promising. The profile looks nice and the kit fits together well. If you have some specific needs that can be met with a kit of Makerbeam (like I did), then I’d get one. I don’t know if I would go so far as to recommend it just to have around, but I do wish that makerbeam.com would come up already and start selling bulk extrusion. I think I’d buy a dozen meters just to keep around for custom cuts when needed.

This is incredibly tedious and error prone, as you might imagine. Fortunately, automated testing of circuit boards is an extremely common problem, so someone invented Pogo Pins – fancy little pins that are spring loaded and have tips that fit into circuit board holes.

I set out to make a fancy test jig for the Cirque Lisa so that we could program and test them *before* installing the screw terminals and without having a crazy mess of wires everywhere. I did the machining and jig assembly, then [Evan] soldered the test board together. Here’s the result:

Pretty awesome looking with all those pins, no?

Both Sparkfun and Adafruit have tutorials on making pogo pin jigs if you’re interested in making your own. Both shops also sell them, but the ones in this jig came from Sparkfun:

• http://www.ladyada.net/make/pogojig/
• http://www.sparkfun.com/tutorials/138

Read on for pics of the build process and a cool video of what I think I need to build next ^_~

I started the build with the stabilizer plate. This piece of plastic will help keep the long pins straight in the jig:

During assembly, one of the ATmega pads lifted on this board, so it was used as the base for this jig. The board was bolted to the plate and used as a template for the pin holes:

The drill press had too much vibration to make these tiny holes accurately, so I used a milling machine instead:

Perfect fit

Test fitting all of the pins. A welding tip cleaner came in handy to clean out a few of the holes:

Cutting some nylon spacers that will fit between the board and the alignment plate:

I used a scrap of 0.25″ aluminum to make the base for the jig. The slot in the middle isn’t functional, it just came with the scrap. The holes are being tapped for the 4-40 all-thread that will hold the jig together. The rods will be held in the holes with Loctite:

Assembling the jig:

(Almost) finished product. The alignment rods will be cut down from where they are in this photo:

Something to aspire to:

Long story short, there were a lot of improvements to be made. To get around the limitations of Eagle, [Evan] drew up the second generation of boards in KiCAD, taking advantage of the human-editable XML file format to lay out parts like the screw holes and terminals with precision that would have been difficult to obtain using only grid snapping. KiCAD also allow the use of the TopoR auto-router, which generated the sweet looking organic traces.

The new motes also omit the LED on the XBee disable pin. This pin is pulled high to tell the XBee to sleep, so having the LED on it wasted 5mA or so during the phase when we would be trying to *save* power. We chose different resistors for all the other LEDs, which saved a few more mA.

Overall, the MKII design is quite nice and we’ll be using these motes as the basis for some upcoming aquarium monitoring projects.

Tune in next time for a build of the test stand that will be used to program future MKII mote boards!

-Michael

Read more for a tip on programming ATmega328/P chips with AVRdude and Arduino

PS – An issue we did run into was a subtle difference between the ATMega328 and ATMega328P. The “P” stands for “pico power,” and there are a couple extra registers on the 328P that aren’t available on the regular 328 that do things like disable the brownout detector to save precious uAs in super low-power conditions. We used regular 328s for the first round of boards, but ran into a problem because AVRdude only knows about the 328P (device signature 0x1E 0×95 0x0F) and croaks when it sees the device ID for a 328 (0x1e 0 x95 0×14).

The easiest thing to do was trick avrdude by changing the device signature in /etc/avrdude.conf for the ATmega328P to that of a regular 328. The two chips are otherwise identical as far as programmers are concerned, so the other parameters are fine. This worked great for reading/burning fuses, but the Arduino IDE has it’s own version of avrdude, so it needs to be changed there to burn the bootloader. The good news is that the Arduino bootloader has its *own* device signature, so once it’s on the chip Arduino will think it’s a 328P and it will work fine with unmodified versions of the software.

In this installment of the Mote Build Logs, I’m making enclosures for the first generation prototypes. I wanted to keep the enclosures round to mirror the design of the mote and make them clear so you could see all the blinky lights. The cases are open around the edges to allow easy access to the screw terminals. They’re designed to be “backpack rugged,” not weatherproof. One is basically a shell so I can toss a mote in a bag and not worry about it getting banged up.

Read on for details on how they were built!

Cases

Each case is composed of two halves made out of 0.25″ acrylic. The material was first sliced up into square on the table saw, then rounded using a bandsaw and circle jig.

Since it was just that much more cool looking, I decided to hollow out the bottom half of each case so the mote could sit lower and be that much thinner. The mote boards didn’t come pre-cut, so I decided to use the mill to round them out as well. Since I would be cutting 15 pieces total, I made a nice jig out of aluminum to use to bolt the circuit boards and cases to the rotary table for cutting. Below is a case blank bolted to the jig plate:

Case blank being milled:

I decided after the first prototype that it was smarter to mill the case blanks *before* rounding them off so that I could use clamps in the corners. The screws in the middle worked great for alignment, but had to be removed on-by-one during hollowing, which took a lot of time.

The mounting holes in the bottom half of each case were marked with a center punch, drilled, and tapped for 4-40 machine screws (by Josie, my lovely assistant <3)

After the first prototype, I also milled a groove in the top to help seat the lid against the screw terminals and let the mounting screws catch a couple extra threads in the bottom. Holes were also drilled in the top so that the terminals could be accessed with a screwdriver:

A test assembly. The screw terminals on this version of the mote couldn’t be placed with precision in Eagle – so they’re not perfectly round, which would have been easy! They had to be lined up by eye, instead, which didn’t exactly work every time. Some had to be widened, but they all wound up usable enough for a first prototype run, though:

Circuit Boards

Since the circuit boards were already soldered (not a good idea, BTW), I raised them up on a piece of MDF so that the terminal pins would have somewhere soft to dig into, providing some more support. They vibrated a lot, but since the bottom of the boards wasn’t flat, using the mill was by far safer than the bandsaw.

Before milling:

and after:

Results

Here are a couple shots of the finished product:

Complete with a fancy sticker on the back! This first batch was built for my research in the Networks and Systems Optimization lab at NMSU (http://nsol.cs.nmsu.edu):

Next time…

We get a new version of the circuit board designed, manufactured and assembled

(I don’t have a pic of a bare board, but this one has only SMD parts)

To solder them, we used a $20 drugstore toaster oven controlled by a Reflow Toaster Controller from Sparkfun.

We tried some solder paste, which worked horribly and bonded many pins on the first ATMega together. For the rest of the boards, [Evan] added some extra solder to the pads and fluxed the boards. Then I placed components, using the flux to stick them to pads. A quick trip to the oven and they were soldered. A couple boards had a handful of components that had to be touched up by hand – mostly LEDs.

When the SMD components were soldered, we soldered the minimum number of screw terminals and loaded the Arduino bootloader. Below is a pic of a board wired into a USB Tiny AVR programmer. The crystal pads on these boards were a bit tricky and one board got bricked once the clock fuses were burned, but [Evan] busted out with l33t assembly skills and set up another AVR as a clock source so we could rescue it.

Here’s a mote board fully populated with screw terminals and XBee being programmed with FTDI breakout board:

In the next installment, I’ll be cutting the boards round and making some fancy cases!

The Good

The DC-24 is quite a nice little vacuum. The big selling point of Dyson vacuums, of course, is that they “never loose suction.” Whether this is true or not, I can’t testify too, but it seems to work just the same whether the dust bin is full or not. The cleaning wand is sturdy and the included furniture tool works very well picking up hair from my short hair cat. The wand is too long to use comfortably on furniture, but the included brush tool attaches directly to the hose, which is long and flexible enough to get the job done. The DC-24 is also very light, which makes it easy to move around the apartment when doing furniture cleaning.

The ball lives up to it’s hype. This vacuum is hella easy to steer.

The brush head, though, is probably the best part of the DC-24. The brush is driven by a separate motor inside the head, which means no belts to break. It also means that the brush can be switched off to clean hard floors. There is also a circuit breaker that switches off the motor if you run over a shoelace or something and jam up the brush bar. This should happen less than with other vacuums, though, since the brush head is designed to ride very low to the floor (“auto adjust to different carpet heights”) and is small enough to keep out most cords, socks, etc.

The Bad

For all it’s (over) engineering, the brush head is also the DC-24′s weakest point. If you or someone you live with has long hair, forget about it. Long fibers get wrapped around the brush and clog it, requiring constant frequent cleaning. Hair also works it way into the bearings on either side of the brush, which will cause overheating and damage to the brush assembly if not constantly maintained. Which is difficult, since the bearing is all but impossible to remove, leaving tweezers and patience as your only option.

The Ugly

The dustbin on the DC-24 is small, requiring emptying pretty much every time I vacuum. This isn’t a huge problem, but it’s smallness makes it hard to empty and the process usually involves me having to reach inside the dustbin and pull out a wad of lint and/or hair. Dirt also tends to get statically charged and gather around the ball, which is annoying to me because I’m the kind of neat freak who would spend so much on a vacuum cleaner >.>

There are 2 filters and the vortex cone thing to clean, vs. my old vacuum’s single filter. Also, Dyson has apparently trademarked “Clear bin”, which is ridiculous.

Buy This Vacuum If…

You have a small apartment, a 50/50 mix of hard and carpet floors, furniture that needs vacuuming, and, most importantly $400 to spend on a vacuum.

Don’t Buy This Vacuum If…

You, or your significant other has long hair.

Instead get a…

Cheap Eureka Maxima or other bagless vacuum.

Would I buy it again?

Probably not – it’s nice, but not *so* much greater than much cheaper ones.

Context – other household vacuums I’ve used:

• Eureka Maxima (mine, bagless)
• Hoover WindTunnel of unknown age (my parents’, has bag)
• Unknown brand of canister vacuum (owned by a friend with a tiny apartment)

Once we made the breadboards, the next step was to prove that we could etch some boards and put surface-mount components on them. Here are some of the results:


Blinky LED Board



The Blinky LED board was designed in Eagle and etched using pre-sensitized copper board. Exposing a mask to the pre-sensitized copper was easy and clean and didn’t require the use of an iron which we didn’t have. This particular board turned out to produce a positive image with NaOH as the developer, which was handy because NaOH is extremely potent (1 gram/200ml of warm water develops several boards) and cheaply available as a supply in soapmaking.

We made this board to prove that we could indeed solder components of the size that would be required on the motes. It also complimented the breadboards as a fun demo in class. The microcontroller is an ATtiny2313 that I got as a sample.

HMC 1051 breakout
The HMC 1051 is a single-axis magnetometer from Honeywell. The dual-axis version (HMC1052) is the magnetometer mounted on the SBT80/TelosB mote package used in my undergraduate thesis project.

The sensor was attached to the board using solder paste, which was melted by placing the prepped board on a “special purpose reflow surface” – a.ka. an old waffle iron. Solder paste seemed to work pretty well with the large components, since we didn’t have the ability to make stencils application was fairly imprecise.

“Kitchen Mote”
The so-called “kitchen mote” was made to verify the eagle design for the final sensor node boards. Unfortunately, the pad spacing on the processor and several of the smaller passive components was too small for our etching methods, so we couldn’t actually test it.

1. Use PDF - In most cases, PDF is the most appropriate format to use. When sending invoices, resumes, letters, memos or anything that does not need to be edited by the receiving person, save as a PDF. A PDF will look exactly the same no matter what device or operating system it is viewed on and nearly every computer on the market has some kind of PDF reader installed. iPods, iPhones, Android devices, and Blackberries can all read PDF. Opening a document in Word or Open Office is very likely to corrupt the formatting of the original, which is distracting and can result in misinterpretation. Fonts are also preserved in PDFs, eliminating the annoying boxes and question mark symbols that are all too common in emailed .doc files.
2. If you really need to send an original document to be edited, don’t send .docx, .pptx, or .xlsx files unless you know for certain that the other person has Microsoft Office 2003 or later. Unless you work with someone and everyone in the company uses the same version of Office, err on the side of caution and use .doc, .ppt, or .xls. If you have super fancy formatting in a document, it might not be preserved, but it rarely happens. If you have fancy formatting that really needs to be preserved (like a layout to be printed), use PDF. Better yet, try using a cloud service like Office Live or Google Docs to share the document where both you and the other person can edit it.
3. Don’t use “Forward as Attachment” - Using this option sends the forwarded email as an attached file. Unless the recipient of the forwarded message uses the same email client as you do (Outlook or Thunderbird, for instance), these attachments are basically unreadable to the recipient of the forwarded message. The raw email file is sometimes human-decipherable, at least, but embedded images can be corrupted completely
4. Scan responsibly - If all you have is a paper copy of a document or you need to scan a copy of a signed doc, please scan responsibly. Most scanners have a “Text” setting that optimizes the image for contrast so that it compresses smaller and is easily readable. Also, don’t scan in color unless you have to – a lot of paper has a color tinge to it and this will be picked up by the scanner and wastes ink at the other end if the image were printed again. For similar reasons, try and crop out black areas outside of the border of the document. No fancy image editing software is required – Microsoft Paint does a superb job at small jobs like cropping and resizing
5. The final rule of attaching documents is: Don’t send attachments - If the message is a document or memo, just put the message in the email body. More and more email is being read on smartphones and those people who use Blackberries and such heavily likely won’t have access to a computer to open a .doc or have the patience to download a PDF. Plain text with simple formatting is the most readable format of all. If branding matters, try HTML email, but keep in mind that many email clients block images by default as an anti-spam measure.

Above all, keep in mind that presentation is everything – if you send out a document that is difficult for the recipient to open or work with, expect the content to be overshadowed by the annoyance of opening it.

As an extreme example to underscore this point, all students at my university were sent a letter and a form today asking for nominations for a faculty award. I’m guessing that a digital version of the university letterhead was not available. Thus, the person who sent the letter typed it, printed it out (on official letterhead), had it signed by the provost, then scanned it and sent out to students as an image embedded in an email. The message body had absolutely no content, just the image, making it all but impossible to read on a mobile device and even highly inconvenient to read on a computer proper; not to mention probably ranking high on a lot of spam filters. The accompanying form had also been printed out, scanned, and sent along with the cover letter as a PDF attached to the email, thus making it impossible extremely inconvenient to fill out electronically and email back as suggested in the cover letter.

So much for the “paperless office”… XP