Fun with a Waterjet at TechShop

I've been eager to join TechShop since they launched many years ago in the Bay Area.  This fall we Austin makers finally got our chance!  It's a great space and is connected to a Lowes hardware store, so materials are close at hand:

You can see TechShop equipment details here: TechShop Facilities  In order to use many of these tools, one first must take a 'Safety & Basic Use' (SBU) class.  As the name suggests, the class is focused on teaching the basics of using the tool in a safe manner.

Many of the early members are blasting their way through the SBU classes as fast as possible - as evidenced by the full Austin class calendar.  I've taken my share, but was most excited for the Waterjet SBU.  A waterjet uses high pressure water mixed with an abrasive media (60k psi with garnet in our case) to precisely cut (via erosion) very thick material.  Getting access to a precise and powerful piece of equipment like this (supposedly costing > $200k) was one of the main reasons I joined TechShop.

TechShop Waterjet Mach 2

The Waterjet SBU is 4 hours long, taught by local expert Lee Lanford, owner of Heart of Texas Metalworks.  Lee was a lot of fun and great teacher.  We spent 2/3 of the class learning the Waterjet CAD/CAM software - Flowcut and FlowPath - primarily through drawing our own designs.   The last 1/3 of the class was at the Waterjet, where we walked through the operation and each student cut their own playing card (below).

Waterjet SBU Exam

After the class was over, I wanted to take another pass at using the Waterjet to reinforce what we learned.  I used Flowpath to whip up a quick design using the 1st comic book emblem that came to mind:

FlowPath-Screen

And here is a preview of the cutting path, in the FlowCut software. The green lines show the traverse path between cuts, the red dots indicate a lead in cut (entry point). You can control all of this as well as the cutting speed as part of your design files.

FP-Preview

I then found some 1/8" steel in the scrap bin, setup the Waterjet and let it run.  The cut took about 5.5 minutes.  Here is a 4x speedup video.

And the result, a nice ornament for our Xmas tree (probably not).

20121110-IMG_4161

The class was a lot of fun.  I now feel very comfortable with an awesome piece of equipment.  Next step - building something useful!

Measuring AA Battery Life for LED Pixel Halloween Costume

As I mentioned in the LED Ring construction post, I wanted to make sure 4 AA batteries would power my 25 RGB LEDs through a good 3+ hours of trick-or-treating. The Adafruit doc provides a good battery life estimate, but I figured I might as well measure it. Fortunately, I had a myDAQ handy, and hooked it up to measure battery current and voltage as shown in the schematic.

Note: the power consumed by the 1N4001 diode is not measured with this setup.

I then wrote a quick LabVIEW VI to measure & log the battery voltage and current every second.  This way I could see the voltage and current drop as the batteries discharged.

With some fresh AA's, I powered up the LEDs and ran the 'RainbowCycle' pattern from the Adafruit WS2801 library, shown below (all 25 LEDs are on driving different colors, this should be a good worst-case current draw mode).

The myDAQ logged data to a CSV file for about 8 hours, you can see the resulting current, voltage and power vs time (click for a larger version).

The graph shows about 5 hours of LED power before the current drops off, plenty for my application.  In total the LEDs consumed about 1900 mAh during the test, which is in line with the original estimate.

Synchronized LED Ring Construction Details

As promised, here is a rundown of how the synchronized LED rings from our Halloween costumes came together.  When we initially designed these Roomba costumes, we knew they would look great in daylight but needed some flair to make them more visible after dark.   LEDs to the rescue!

Design

As all three costumes had a similar cylindrical shape, we decided to build an 'LED ring' that we could then mount to the top of each cylinder.  An LED ring would allow for many interesting animations and could light our way in the dark.  Here is the original sketchup design:

LED Ring Sketchup Concept

Of course, one of the first questions with any portable electronics project is battery life vs. weight.  With RGB LEDs, current consumption varies widely based on the color displayed.  Adafruit has a nice document with some good general rules for estimating LED power consumption.  This gave me confidence that 4 x AA could last through ~3 hrs of trick-or-treating, but to be safe I setup a quick battery life test (which confirmed the estimate).  I'll detail this test in another post.

Finally - blinking LEDs is really cool, but we wanted a little more flair.  Why not synchronize the animations on the 3 rings wirelessly?  We will update the rings every 50 ms, so we would want < 10 ms synchronization, ideally < 2 ms.  Many people may not notice, but it's a fun challenge.  With about 3.5 weeks before Halloween, we started hacking.

There are a lot of cheap RF modules and ICs from TI and others that allow for wireless communication.  Here is a fun project that uses a TI radio: Wireless Friend Finder .  However, with 3 weeks to work we had to keep it simple - so I thought it would be fun to try to communicate & synchronize via Infrared (IR) .

IR LEDs and receivers are cheap, easy to use, and surprisingly effective for communicating at little data over 10-25' (like your TV remote, or the famous TV-B-Gone), and we could hook them right up to the Arduino UART (handy for synchronization).  Of course, like your remote, the IR transmission can easily be blocked by objects between the transmitter and receiver.  Given our low data bandwidth requirements, we felt we could adjust our protocol to tolerate lossy communication.

We whipped up a simple IR TX/RX PCB, sent it off to OSHPark, and crossed our fingers.  The general idea was to allow the 3 rings to share time and LED animation status via IR communication.  The PCB design is shown below, I'll detail this in a future post.

Screen Shot 2012-11-10 at 9.25.43 PM

Materials

The next step was gathering the materials to build the ring, here is what we used:

• Ring structure = $5 Metal wreath frame, 24" diameter (from Joann)
• 1000 4" zip ties
• 12mm RGB LED Pixel Strand of 25 (from Adafruit)
• 4-pin JST receptacle cable (from Adafruit)
• 2-pin JST In line power wire, male & female (from Adafruit)
• 4x AA Battery Holder (from Adafruit)
• 1N4001 Diode (from Adafruit)
• 6-wire Ribbon Cable (from Sparkfun)
• Pro Micro 5V Arduino compatible ATmega32U4 (from Sparkfun)
• 4 Custom IR TX/RX PCBs (detailed below)
• Various batteries, hot glue, solder, header pins, etc.
• Velcro-attached fabric sheath

Assembly

Amazingly, the 25 LED strand was the exact length to fit the wreath frame circumference, LEDs even lined up with the cross members of the frame!  Pure luck, but it made assembly go quickly - I zip tied the 25 LEDs around the ring, making sure to wrap in the same direction for all three rings, and leaving a space for the Pro Micro at the back.

LED Ring Assembly - zip ties

Once the LEDs were attached, I soldered connector cables to the battery and Pro Micro SPI pins.  Pro tip: When soldering headers onto a small PCB, use a foam backing to keep the headers straight without burning your fingers:

Pro Micro Headers

If we didn't care about synchronization, we could attach the Pro Micro and battery pack to the ring and be done.  Unfortunately, we had only just begun.

IR PCB Assembly

The IR PCBs arrived from OSHPark on 10/22.  With a box full of components from Mouser, we soldered them up in a few hours (4 PCBs/ring = 12 boards to solder), and quickly tested the frequency, IR TX and RX functionality.

IR Synchronization PCBs

Now it was back to the rings - after soldering wires to each IR PCB, I used even more zip ties to attach 4 IR PCBs to each ring (~90 degrees apart).  Then can the fun part of soldering these 4 wire-clusters to the Pro Micro headers.  It's not pretty, but it did the job:

Soldering IR connections

I added in a 9V to power the Pro Micro - wired right in to the 'RAW' pin (just in case LED current caused the 4AA output to droop) - and the assembly was good to go.  Then it was a healthy amount of hot glue and more zip ties to secure everything to the frame.

LED Ring Brain Full Assembly

LED Rings

My wife insisted on sewing black covers for each ring, with holes for the LEDs and IR, and a velcro seam at the back.  In the end it looks pretty nice:

3 LED Rings

Roomba Costumes with Synchronized LED Rings

A quick post on our 2012 Halloween costumes - after many late nights, we had an awesome time walking around the neighborhood tonight.

Related posts:

• LED Ring Build details & bill of materials
• LED Ring battery life test

Background - my son loves our Roomba vacuum cleaner.  He loves to watch it work, loves to take it apart and clean it, even loves to watch 10 min teardown videos on Youtube.  As such, we knew we had to make him a Roomba costume for Halloween.  But what about my wife and I?  The virtual walls!

A few weeks back, my wife and MIL got to work crafting an awesome costume.  Here they are in action:

Crafting the Roomba costumes

And here is the end result (below).  You can see that Halloween started out a bit rough, the key player in theme was non cooperative.

A malfunctioning Roomba with two exasperated virtual walls.

However, all was not lost.  Our backup plan involved mounting his costume to his stroller (along with him).

After dark, we turned on the flair - my friend Matt and I built up 3 LED rings to mount on the top of the costumes.  The LEDs are the 25-pixel strings from Adafruit controlled by an Arduino Pro Micro from Sparkfun, powered by 4 AA batteries and a 9V.

Update: I detailed the LED Ring Design and Construction here.

The best part - we were able to synchronize the 3 LED rings via an IR transceiver board we built.  The rings used IR to share timer values and calibrate to a common understanding of time - allowing them to blink in sync!  Here is a mockup of the ring (with purple PCBs from OSHPark) and a photo of our PCBs (9 of 12).

The costumes lit up the night, we heard a ton of oohs & aahs, pictures taken, lots of questions, it was a blast!  I'll post more on how we did it when I catch up on my sleep.

And here we are trick or treating.

Roomba Costume

Fresh Popcorn in a Faraday Cage

RFI is a serious concern (especially for radio telescopes), but so is a tasty snack.  Here is one way they balance these needs at the GBT.  This microwave is in the main control room, which itself is in a faraday cage.  Side note: I need a microwave with a 'Danger' mark.

Toddler Snorricam - Chest Mount First Run

Yesterday at my son's urging we flipped the 'robot' backpack around and got some front facing video with the Snorricam.  It's pretty great, amazingly he ignores the camera and goes about his business!  Here is a photo of the rig in action.

Build details are here: Toddler Snorricam – First Video

Toddler Snorricam - First Video

Ever since I saw this IndyMogul how-to about a year ago, I knew I wanted to build a Snorricam.  For the uninitiated, a Snorricam is a camera rig that mounts to the body of the actor, keeping said actor in a fixed position in the frame, while their surroundings change in the background of the video (example). Of course, I had a twist - I wanted to use it with the most interesting person I know (who can be difficult to work with) - my 2 year old son.  We built a scaled down PVC rig and my wife came up with the genius idea of using his doggy backpack as a mount.  I thought it would take some coaxing, but he was eager to give it a try (but only if we left the tail attached!).  We put it on and he just took off (laughing the whole way)!  He seemed to enjoy running the camera rig into obstacles. It turned out great.

Build Details

I was reminded of this project last month when the Indymogul team posted a new video of a $14 third person rig (video here).  This new Snorricam looked much faster to build as it was made of PVC (Instructable here).

Here is our final, doggy pack mounted rig.  It isn't as secure as the Indymogul version, but it turns out to work well enough for our purposes:

Toddler Snorricam

Build Tips

I bought PVC glue, but the joints were snug enough to roll glue-free, at least for this first run.  For cutting PVC - spend $15 on a ratcheting PVC cutter (below), its faster and cleaner than a saw.  Although I kept him away from the cutter, my son loved putting the PVC pieces together!

Racheting PVC cutter

Parts List

(Amazon links)

• Ratcheting PVC Cutter

• Eddie Bauer Harness Buddy, Brown Dog
• 1/2" PVC from your local hardware store - 6 ft of tube + 6 R angles, 3 T's, 2 45 deg, and one cap

• GoPro HD HERO2
• 1/4-20 (1/4" diameter, 20 threads per inch) bolt + nut for camera mount
• Power Drill (drill hole in PVC cap to mount camera)

Next Steps

His older cousins play soccer, race bikes, and other fun things - so maybe we'll see how the rig works with more action. I'd love to try with a front-mounted rig (to capture his face), but that may be too much of hazard. If we get a lot of use out of it, maybe we can build a fancy rig like this: Videoschool Snorricam

Update

That was fast! We have some front facing video: Toddler Snorricam – Chest Mount First Run

16 x 32 LED Matrix Panel - How do they work?

I found a blog post today that does a pretty good job of detailing the 16x32 LED matrix circuit components and how they work.  The PCB looks a little different than mine (components in different positions), but the circuit appears to be very similar. Here is the diagram they created highlighting each IC, click through for the full explanation and links to datasheets:

Hobby PCB - 16x32 RGB LED Matrix Technical Details

16x32 LED Matrix Power Consumption

Following up on my LED Matrix bringup post - as I am using the Source Measure Unit (SMU) as my 5V power supply, I can quickly measure the 5V current draw of the LED matrix under different clock and pixel data conditions. Of course, these measurements are from one LED panel at room temperature - not a large enough sample set for setting specifications.   But this data can be a rough estimate, useful if you're planning a portable LED matrix application.  (click to expand)

Notice that the LED panel current consumption decreases as clock frequency increases.  This suggests that the LED 'on time' is decreasing.  I'm guessing this is caused by frequency-invariant delays in the LED driver shift registers.

For example, the output of the LED driver shift registers may be updated a fixed 50 nanoseconds after latching the data (for any clock rate), whereas I start loading the next line exactly 6 clock periods later.  As I increase the clock rate (decreasing the period), the time between latch + 50 ns (LED on) and latch + 6 clock periods (LED off) becomes shorter.  I need to track down a datasheet for these parts to see if this holds water.

Less LED 'on time' per time unit should also decrease in the perceived brightness of the display, I'll take another quick video to demonstrate this affect.  I didn't notice this during my initial testing, but I was not cycling through clock rates with a fixed display pattern (and the display is really bright so I wasn't looking directly at it too much!).

Austin Mini Maker Faire 2012

I had a great time at the Austin Mini Maker Faire this weekend.  Although smaller than the events held at the Expo Center in years past, the Austin-focus of the 'Mini' Faire is a great improvement.  Rather than trucking in makers from other cities, this organic mini-Maker Faire approach helps to grow the Austin maker community. There were a lot of great electronics projects.  My son (and many other kids) couldn't get enough of the full size R2D2 (which he called "Big Roomba!"):

Two local FIRST teams had several robots driving around, the AustinEV club had electric truck and motorcycle conversions on display, ATX Hackerspace held a soldering workshop which was packed... and much more (full list here).  I snapped a few photos, and there is a nice photoset from the event on the Austin Mini Maker Faire Facebook page.

[nggallery id=3]

Next week - the big show in San Mateo!