Servo and Motor Controller

bm011

An RC servo motor controller is what I’m currently working on.  Here’s a copy of the schematic in PDF format.  In a previous blog post I recounted my effort to build a robot controller and incorporate it into an Arduino clone design (see that post here).

[Read more…]

Robot Wheel Design

Robot wheel prototype

Robot wheel design has been on my mind lately.  We’ve been asked to take some robots down to a fundraiser for the Discovery Museum in Sacramento.  I want to build something new for their event and was thinking about a robot that looks like a regular household item until you activate it.  I’ll probably use a lunch box (my favorite chassis) or a brief case.  But the design I have in mind needs to deploy its robot wheel before moving.  Basically it needs to stand up and then move, and when deactivated it should sit down and just look like a regular object.

To do this I wanted a hinge mechanism that I could position electronically.  Think of it like an aircraft wheel that is stored in the wheel bay when not in use.  A hobby RC servo seemed like a good fit for the design.  On the way to work I picked up some brass at the local hardware store and at the end of work I took a couple of hours to cut/drill the brass and solder pieces together.   The prototype seems to work, and should be light weight.  I stuck a knobby RC truck tire on this, but will use something lighter in the final design.

It took me a couple of tries to get the mechanics correct.   I’ll do a mechanical design of this robot wheel on Google sketchup, since I need to fit four on the robot, and having some defined dimensions would help me do that.  But overall it was a successful effort considering I spen just a couple of hours trying to put this together myself.  I really like using brass since it is easy to work with and you can braze it with solder and a heat gun.  It’s a poor man’s weld.  I’m also quite proud of the hinge that is really just one brass tube that slides into another slightly larger brass tube.

RS232 to RS485 Conversion

rs232_485_2

RS232 to RS485 conversion is somewhat common in industrial applications.  As part of an R&D effort I wanted to see if I could put together a small design converts RS232<—>RS485 with a minimal footprint.  Here were the goals of the design.

1.  Convert logic level signals to RS232 signals (assumes a master, like a microcontroller is attached).

2. Convert logic level signals to RS485 signals (assumes a master, like a microcontroller is attached).

3. Convert RS485 to RS232 and back without requiring a master unit to control the Drive Enable (DE) or Receive Enable (RE#) pins on the RS485 converter.

[Read more…]

The Web can buy beer

Yesterday, the world-wide web turned 21.  That is not so say the internet turned 21.  The internet was developed in the ‘60s.  I’m talking about the Web as we know it:  web pages, “www”, visual interface, and hyperlinks.  It was developed as a side project* by Tim Berners-Lee while he was working at CERN.  As it was developed in Switzerland, it’s probably not going to go buy the Solutions Cubed brew of choice and local favorite:  Sierra Nevada, but I hope it does celebrate in style.  Allow me some space as I get all old-fogey engineer.

I still remember the first time I saw the “web”.  It was 1993, so the web was all of 1 year old.  My business partner (he was just a college buddy at the time) took my into the Unix Lab in the EE Department at CSU Chico.  He loaded something up called Mosaic.   We surfed the web for a bit.  I’m not sure what we surfed to or looked at.  I am sure we soon grew bored and went and had some Sierra Nevada.  I would never have thought how I would do engineering would change so dramatically over the next 20 years.  We used to have to look up everything in paper data books and hopefully you could get your hands on them.  Now it’s all al click away.  All of the web-based businesses we use such as DigiKey and Advanced Circuits, would not exist as they do now without the web.  I am glad that Time Berners-Lee decided to ask for help from his initial audience in populating the web.

*What is the deal with “side-projects” rocking the world.  I am a big fan of Google’s efforts – especially Google Earth.  And a bunch of how the Bell Labs was run was basically with a bunch of side projects.  But I have to figure that Tim Berners-Lee wins the award for “most-useful side-project of all time”.  Almost all of what makes the present day so cool is directly built on the web:  Netflix, Amazon, Robot Videos, and all the rest.  I wonder what Berners-Lee was thinking about as the ceiling for his invention when he was working on it?  Certainly not this.

Robo Cup 2012

Here is an interesting article about the international Robo Cup and the advances that have been made in robotics that can be tied directly to the competition.  It also brings up some issues related to the reason for “play” engineering.  Basically, the thesis is that competitions such as this, the fire fighting robot, and the micro mouse competitions, among others provide a proving ground to solve engineering problems in a concrete manner.  In addition, the tasks themselves are fun which allow for non-technical and semi-technical people to appreciate the events.  I fully support using “play” engineering as a proving grounds and idea factories to solve real-world problems.

A Soft Jumping Robot

soft_robot_jumping

Japanese researchers have created prototypes for “soft” robots that can jump, crawl,  and roll (even uphill).  A detailed summary of their work can be seen in this pdf file.  The prototype uses Shape Memory Alloys, or SMAs, by Toki Corporation.  By applying a pulse-width modulated signal the alloy is heated and deforms to a predefined shape.  Multiple SMAs allow the overall circular form of the robot to be deformed allowing rolling and jumping shapes to be created.  A more 3-dimensional version, where multiple circles are used to create a spheroid shape, can also crawl.

Most mobile robots use wheels or legs to create ground locomotion, so this concept is a unique departure from those ideas.  Like a lot of research ideas the actual practical applications are limited.  Currently SMA are not very energy efficient and have quite a bit of hysteresis in their pre-formed shape.  This means your power supply (batteries) and controller may not be able to reside on the mobile shape and that the shapes you form may have a lot of variability.

But imagine thousands of miniature strips of SMAs each receiving tiny currents from a central controller and you’ve pretty much modeled a human joint.    You can also imagine how useful robots based on this technology could be in places where energy is readily available and gravity is not a force to be reckoned with.  Satellite technology and space exploration might benefit from the ideas proposed by these researchers.

New futuristic computer technology

A while back, I made a big deal about logic gates made from crabs.  In retrospect that accomplishment was not especially great – specifically the OR gate.  If two halls of crabs converge into a single hall, an OR truth table is pretty easy to envision.  The AND is a different story. . . .  But, I’ve since moved onto cellular computing.  Advances along these lines are necessary if Moore’s law is going to keep chugging along.

The new advance was developed at Johns Hopkins University and is the creation of AND and OR gates using protein cells.  The drawback right now (other than the lab-only setting, limited logic functions, and specialized cells) is that the logic assertion takes on the order of one or two seconds.  Talk about a race condition!  Another cool thing, is they extended their research and made their outputs fluorescent  — better than an LED on the output.

The craziest thing about this, is that Douglas Adams called this development (including eons-long run-times) 30 years ago.  I just hope the eventual answer is not 42.

A Robot Head and Shoulders Above the Rest

My initial reaction to this robot was that it was kind of silly.  It reminded me of a robotic Flowbee.  And the sales pitch voiceover in the video made me wonder of Godzilla might come crashing through the roof.  But this robot really is a piece of quality engineering.  Imagine designing a robotic system that touches a human being.  That’s not a trivial design.  Add to that the need to touch a human head, while the human is in a vulnerable position, and cleaning something as tangly as hair (yeah I know tangly’s not a word, but this is the internet, anything goes).   Then you have to design it for the variety of human head sizes, and hair lengths.  All-in-all this is a difficult problem to solve, and a pretty substantial feat of engineering.  Not sure the case was made for why I shouldn’t just wash my own hair, .

Book Report–The Idea Factory

Wow!  I am not going to dwell too much on the writing or the methods of the reporting (by all my measures, everything was just fine), but story behind The Idea Factory is amazing.  I am too young (if that can be said about a guy on the wrong side of 40), to remember any of the big inventions and breakthroughs detailed in the books, so most of the information in this book was new to me.  I am suitably awed.  Basically, from about 1920 to 1970 Bell Labs was THE driver for technology development in the United States.  Here is a small breakdown, as I remember it:

  • Bell labs was the research arm for AT&T (long distance company) and Western Digital (phone equipment manufacturer).  It was semi-autonomous, but was funded from the revenues from AT&T and Western Digital.  The company itself was a monopoly regulated by the US government.
  • It was setup to do basic/fundamental research, development engineering, and manufacturing engineering.  Because of the steady revenue stream from AT&T and Western Digital, Bell Labs could spend literally 1000s of man-years and billions of dollars to develop a technology from a basic idea through a manufactured product.
  • At its height, Bell Labs employed over 10,000 people.  On the engineering side, most of the employees were recruited from the cream of American universities (MIT, Cal Tech, Stanford, etc), and were, themselves, the top of the class.
  • The majority of the employees, worked in one location, at first in downtown New York, and later in New Jersey.
  • The building that everyone worked in, was specifically designed to foster innovation.  People’s offices were purposefully mixed up so that a physicist could be next door to a mechanical engineer and a mathematician.  In addition, the offices were off of a 700 foot long hallway.  It was generally accepted that everyone had to work with their office doors open.  That way if someone had to talk to a colleague, they were forced to walk down a long hallway and interact with all sorts of people.  Ideas would naturally germinate from this simple architecture.
  • The management of Bell Labs (specifically Mervin Kelly) fostered a research environment where the research department was free to work on whatever they wanted.  Eventually something would come of it.

Here is a list of some of the things that labs invented during its run:

Without these contributions, you would definitely not be able to read this blog post.  The people behind these inventions were fascinating and the stories of the development were interesting throughout.  I highly recommend this book for anyone that wants to know about those that came before.

Solar Tracker–Revision 2

Yesterday we sat down and came up with the initial ideas to upgrade the original Solutions Cubed Solar Tracker.  Below is the initial block diagram for the new design.  This is the rough pass for what the new tracker will be able to accomplish.  Basically, we will use the same 2-axis mechanical system with one full-size consumer solar panel on it.  We will be able to then roughly follow the sun based on date and time along with the known installation latitude and longitude.  From there, we will use a light sensor to fine tune the panel positioning.  This is essentially an extension of what we have already done – the big addition will be the battery charging circuit.

While a typical battery charger is fairly straight-forward (we’ve done a few in the past for clients), a charger for a solar panel throws in a few more wrinkles.  Specifically, you want to make sure that all of the power from the panel makes it to the battery.  So you can’t simply step down the 35Vdc to a charging voltage of +14Vdc, as the power in the headroom is lost to the gods.  An actual step-down transformer is necessary – basically we need to make a switching power supply.  Cool.

To control all of this we are going to dive into a Propeller from Parallax.  We have not used one before, but it looks real interesting.  While it lacks a bunch of on-board peripherals, the eight, independent processors will make splitting the work between engineers pretty easy.  More updates in the months to come.