Laser Cut Scratch Cat

Wyolum got me a laser to expand our distributed prototyping capability. I’ll have more on that later, but I made a little trinket for Scratch classes that I’ve been teaching.


I had difficulty finding a vector file good for lasering, so I made this one. Nice handout for Scratch classes

The official SVG version has overlapping paths that rely on fills to hide some of the shapes.

I traced a bitmap version in inkscape and did a little cleanup. It’s still not perfect with some double lines.

I’ll try to do a better trace to improve but this one cuts ok.

Update: done


You can find the SVG over at

Note that the Scratch cat image and logo are trademarks of the Lifelong Kindergarten Group at MIT. Check out Scratch at

Holiday project: Dreideltron 5775

The Boston JCC  called me looking for maker activities for an upcoming Hanukkah event, and I thought of this little activity.

The copper tape is pretty unreliable though, and it may not be suitable for a big crowd where there may not be time to debug. I’d be curious to hear of other people’s experience doing paper circuits and the like. This was working perfectly for a while and then it stopped. I suspect the adhesive (which is supposed to be conductive), just stopped conducting. Perhaps I just need to buy higher quality copper tape?

From Blink to Bot: Introduction

JoebotJoe McDermot (leader of the Boston Robotics Meetup) discovered many members of the Boston Robotics Meetup had never scratch built a robot, so he sourced some cheap components from China and led a group build session. Joe did a fantastic job, keeping the price to $60, for a robot with 4 motors, Arduino clone controller,  3 ping sensors and IR control. (Joe’s a modest guy, but I call it Joebot)

After building it, a lot of the members had trouble with the programming, trouble I’ve seen when people attempt to do slightly more complex projects on the Arduino, after doing the basic Blink, and other examples.

In this series of posts, I will attempt to show how to coordinate several activities in an Arduino program. While the robot is an excellent challenge, the lessons here can be applied to any project where you have to “simultaneously” read sensors, control actuators, etc.


  • 4 DC gear motors with Wheels
  • Acrylic base
  • L9110s  BLUE motor control board (2 channel, 2 wire control per motor)
  • IR receiver
  • 2 – Ping type distance sensors
  • Arduino Nano V3 Clone
  • Breadboards, wires, etc.

One note of caution here, many of the Nano clones used counterfeit FTDI chips, and were bricked by the new windows driver. DCCDuino, is actually not an out and out copy and uses a different USB-Serial chip, and works well.

JoeBotLabledNext post we’ll work out how to handle motor control.


Paper Circuits Workshop

I led a paper circuits workshop at the Auburndale Community Library, and I thought I’d share a few thoughts in case you’d like to try it on your own. By the way, all the photos in this post are by our host Dana Hanson, a good friend and one of the key volunteers that keeps the Auburndale community library going (with no municipal support!)

I let the library know that it would be ok for all ages, but if kids were under 7, a parent should stay to help. We got lucky in that we had a couple of parents who either got it right away or picked it up very quickly and they were able to help tremendously.

The first I saw this idea was on the MIT High Low Tech site and they have some good pointers. They use surface mount LED’s but they are quite difficult to tape, so we modified with regular LED’s.

Setting the basic rules: Lithium batteries need to be recycled (and don’t eat them…), What is a circuit (relating to circles…) watch out for short circuits, polarity of LEDs. This circuit is very forgiving, the internal resistance of the coin cell (we used CR2032) means you don’t need a series resistor to limit current for the LEDs, and also means that short term short circuits don’t destroy the battery immediately or cause a fire!

Some Clown

We set up a station to hand out materials at the circulation desk. Next time, I think I might divide it into several stations. The most time consuming part was letting them pick which LEDs to use. You could either limit it to one kind, or spread it out a little. You could have people just pick them from bins on the table, but sometimes people take too many. (It wasn’t a problem here….) My son’s Mason and Grant assisted me both in preparing and in debugging people’s circuits, and were indispensable.  25 Kids with about 80% of them needing help would be too much for one person.

Critical help

One of the mothers came up with a good way of making the connections more reliable add copper tape under the leads of the LEDs as well as on top.

Hard at work

Beautiful artwork

One of the book racks made a nice display for the finished work.Gallery of creations




Free and Open Source Data Service from Sparkfun


Sparkfun launched a free cloud data service for your devices. It’s limited (rolling last 50MB) but also open source so if you want to roll your own you can do as much as you want!
This is an example of how doing a service and open source can benefit you. As they say

Our hope is that you buy a SparkFun widget to connect your next beehive.

It looks easy to use, and besides making it easy to put your own sensor data up, data from all other users is public, making it available to data scientists and hobbyists.

I’m looking forward to trying it out, and if I do I’ll share here!



NovaBooth – Improving the Open Source Photobooth

NoVa SnapAs you may have read here previously, we (the Wyolum Gang) created a photobooth for the Open Hardware Summit, for the purpose of customizing the e-paper badges we made for the conference attendees. This processed the pictures into a small black and white image for the e-paper badge, and saved it onto the badge’s micro-sd card.

I was headed to help out at the Northern Virginia Maker Faire, and thought it would be fun to update the photobooth to take full color pictures, upload them to the Internet and offer to email them to friends and relatives.

The email message and logo files are easy to add and customize.

For basic construction, visit the original post, but download the new software here:


The fabricate directory has the laser cut files, arduino for the AlaMode Program, and scripts for the python photobooth code.

Edit to customize the email subject and message. contains the authentication information for the google email and posting accounts. You’ll need to set up application specific passwords for this on your google account. You can use the same account, or separate.

Wireless keyboard, had to add a powered hub.


  • External powered hub was a pain.
  • Proto-screw shield was too heavy and lifted off
  • Some of the nuts came loose in travel.

Photobooth 1.5

To solve the first problem, I determined to replace the non-powered hub in the photobooth with a powered one.  I tried to add power to the unpowered hub, and this worked at first, but then took out the power supply and made the raspberry pi flakey too.

Scratch that, I ended up using a small belkin powered hub. I y-connected the power to it.
I noticed that  a convenient orientation put  4 ports right next to the edge,  so I cut a hole in the box to expose them.

Luckily AlaModes ship without shield headers installed, so I replaced the AlaMode and protoshield with an AlaMode that was directly soldered to the button, led-strip, power and ground.

I updated the AlaMode’s photobooth program directly from the photobooth. Apparently the new AlaMode’s pullups weren’t as strong, so I added a 5.7k pullup to the Button Pin.


Instant Party for Arduino Day!

You Do It Electronics in Needham asked me to help out with their Arduino Day celebration on Saturday March 29, and I wanted to make a nice Arduino demo that uses parts that they sell, relatively simple, and Fun! You Do It also sells DJ and disco equipment, so I thought what would be better than Arduino controlled Disco lights and Music?

Here are the Parts I chose:
Sparkfun Motion sensor (to trigger the Instant Party)
Sparkfun MP3 Shield
Sparkfun proto-screw Shield (to make it easy to hook up.)
Power Switch Tail – turn on the disco lights

You’ll also need a micro-sd card and some music. I chose some Creative Commons Techno from Soundcloud:

First, we need to find out what pins are in use in the MP3 Shield so we can figure out where to hook up the motion sensor and power switch tail.
The easiest way to do that is by looking at the schematic, or this page at Sparkfun:
MP3 Shield hookup

Digital Pins 5 and 10, and Analog Pins A0-5 are all free.

I used the analog lines because they are close to the power and ground pins for both the power switch tail and the PIR motion sensor.

Because the PIR output is open collector, I use the internal pullups on the Arduino.

Git the code here:


A Brief detour into Cryptocurrency

dogecoinI attended a meetup of HackBoston, run by Abby Fichtner (HackerChick) from the Harvard Innovation lab. Project 11’s Reed Sturtevant wanted to learn more about Cryptocurrency, so he started experimenting and shared the basics with us at the meetup.

Most people have heard of BitCoin, but Reed shared that it’s so hard to mine now, it’s very difficult to experiment without putting in real money. He found DogeCoin as a fun, and lighthearted alternative. Often used as a +1 or like tipping system for blogs and on social networks.

Pre-work was to set up wallets before the meeting, as they take time (about 5 hours) to synch with the network.

We each recieved 2000 dogecoin at the beginning of the meetup. Then we learned the basics of how to use addresses (essentially the destination of coin transfers. We transferred money to each other.

While doing this Reed went over the basics of cryptocurrency, which I won’t cover here, but the wikipedia article covers it pretty well.

We also created accounts at which lets you send DOGE via twitter. (feel free to tip @osbock)!

Intro to Mining

If you read the wikipedia article, you know that transactions are verified by a cloud of contributers performing cryptographic checks distributed across the Internet. “Miners” are rewarded both through the creation of new “coin” (which slows) and the distribution of transaction fees. Without special hardware, your chance of winning a race to solve the puzzle that is a cryptocurrency transaction is minimal, so mining pools were created to distribute the work and the reward.

We created accounts on the mining pool site, created workers (essentially addresses within the pool for your mining processes to communicate) and downloaded and configured mining software.

I recently got a nice “gaming” laptop from Lenovo, because it offered very good price performance, plus it has an Nvidia graphics card which I wanted for computer vision experiments.

Because Nvidia graphics cards contain many parallel processing cores, and Nvidia has released an API that allows their use for general purpose parallel computing, they are often employed for mining.

The Experiment

Following instructions from, I set up cudaminer (cuda is what Nvidia calls their processing cores) and another instance running the pooling cpu-miner to take advantage of the native cpu cores of my laptop’s i7.

Each mining process (cudaminer and pooling-cpu-miner) were configured to report to a “worker” setup at

I then mined overnight to determine:

  1. The kH/s (thousand’s of hash per second) capability of my laptop
  2. How much doge/hour I could mine
  3. Electricity cost of running the computer. (determined with a P3 Kill-a-watt)

Using an exchange rate, I converted into dollars to see if I could mine more doge than the electricity cost.

The Results

The cuda miner was reporting an average of around 80 kH/s and the cpu-miner around 54 kH/s

In 14.07 hours I mined 261.229 Doge for 18.57 Doge/hour, or $.0147/hour (at an exchange rate of .00079$/Doge.) Power usage was approximately 1.4 kWh, which I calculated to cost $.19.

Projecting out to a month, at  134 kH/s, I’d earn a whopping $.95 (and lose a lot of the computing bandwidth of my laptop!)

I could invest in a fancy graphics card and dedicate a computer, but I imagine it would cost even more electricity.  So, I would say  overall, a negative return on investment.

This apparently became the case for bitcoin as well, until people developed specialized hardware that gives a high H/s rate at low electricity usage, first using FPGA’s and then specialized ASICS (Application Specific IC’s). Doge uses a different Hash algorithm than Bitcoin, so you can’t use the now ubiquitous ASIC mining equipment optimized for the SHA 256 algorithm. There’s a new ASIC to be released soon that supports Doge’s (and litecoin, and others) Scrypt algorithm, and by extending my calculations, the cheapest of these devices ($119) would pay for itself in 3 or 4 months.

I’m going to try to get one to see. Of course, all these calculations are based on assumptions of stability (exchange rate, mining rate etc.) which are completely unreasonable. Still it’s fun to try!