Superpositioned

Another RCX-based balancer

Matthew — Mon, 13 Nov 2006 08:38:00 -0800

I’m a sucker for interesting robots. Especially if they are built out of Legos. Everybody loves the Legway, but you have to admit that the concept is overly simplistic. The real challenge is building the system around the limited resources of Lego’s RCX.

Now meet another balancing robot courtesy of Maurits Kooiman on Lugnet. This one balances on one wheel and apparently uses two RCX bricks and four sensors along with three tires to maintain balance on the axis perpendicular to its unicycle. The robot balances on the main axis by directly powering the wheel. The axis perpendicular to its unicycle is balanced by the set of rotating tires. The fourth tire seems to be a freely swingy pendulum, but it may in fact be controlled by yet another motor. Check out the video!

Finally a new layout!

Matthew — Sat, 11 Nov 2006 16:56:00 -0800

I officially declare this site back in action! I been ignoring it for too long. Over the past week, I have updated Typo and designed a new layout. This is not the completely final version, though. The header/logo and advertisements need some work. The previous layout performed very well in that department. Hopefully this layout will make up for the difference by attracting more repeat visitors.

As you can see, the new layout is significantly better on the eyes. Unfortunately, it has some issues with Internet Explorer 6. Hopefully I can figure out what is causing the extra space to appear. Here is a quick list of the updates:

Windows Vista Typefaces: Vista includes some “cute” fonts. If you have them installed, then the new layout uses Calibri. The best way to describe it is “semi-serif.” It actually replaces Times New Roman as the default font in Office 2007.
Fluid width and resolution independence. Go ahead, resize your text. The sidebar automatically resizes as well! Of course I also added fluid width. (Fixed width is useless.)
The colors are kinder on the eyes. I grew tired of deep red and flashy magenta.
Better readability around the ads. I might regret this one, but at least you can read through the copy without getting a headache now.

Watch for a new post detailing my upcoming endevors!

Generating Electricity from Sugar

Matthew — Thu, 01 Jun 2006 09:48:00 -0700

The Digg.com headlines for today include research about bacterium that produces electricity by consuming sugar. The Escherichia coli bacteria actually excrete hydrogen after consuming sugar. An electric fuel cell is then powered by the hydrogen.

The team fed Escherichia coli bacteria diluted caramel and nougat waste. The bacteria consumed the sugar and produced hydrogen, which they make with the enzyme hydrogenase, and organic acids. The researchers then used this hydrogen to power a fuel cell, which generated enough electricity to drive a small fan.

They are also using the bacteria to recover palladium metal from the spent catalytic converters of older vehicles.

Are you logical? (Digital logic puzzlers)

Matthew — Wed, 17 May 2006 12:16:00 -0700

Clive Maxfield recently posed the Black Box Brain Boggler in the Logically Speaking column of May’s EE Times. The original article was incorrect and posed an overly simple problem. In fact, the real problem is much more difficult, but possible none the less.

At first, this appears to be deceptively simple. We start with a black box with three inputs–A, B and C–and three outputs (see below). The outputs, which we may name !A, !B and !C, are the logical inversions of the inputs.

The challenge is implementing this black box with only two inverters, a bucket of basic gates, and without a hard coded binary 0 or 1. The bucket of gates presents two levels of difficulty. The former being far easier than the latter. In fact, the first bucket should be easy for any digital design student to solve.

The bucket of basic gates contains AND, OR, and XOR gates. Remember, you cannot connect any of the inputs directly to a binary 0 or 1.
The bucket contains only AND and OR gates. The solutions are extremely complex, but do exist!

He recently posted the clarifications and some purposed solutions on DesignLine. If you are curious, here are the spoilers for challenge one and two. If you enjoy proofs, there are some more logic puzzlers for your unsatisfied brain.

Capactive Touch Sensing

Matthew — Mon, 15 May 2006 22:20:00 -0700

My last project brought up the subject of capacitive touch sensing. You may not have realized this, but capacitive touch sensing is currently on the forefront of electronics. Mechanical touch sensors are known to wear and ‘push-less’ sensors are just cooler. In case you did not know, the scroll wheel on an Ipod uses an array of capacitive sensors along with the infamous touch lamp.

For the engineer in you, Planet Analog has a lengthy overview. Basically, you design a circuit that is highly dependent on the value of a small capacitor. The capacitance in a finger then causes a significant and detectable change in the circuits output.

The doorknob touch alarm functions in just this manner. It does its job, but is not as precise as more complex circuits. For more schematics, check out Discover Circuits. A recent article on DesignLine proposed that this interface be adapted to mobile phone keypads.

Doorknob touch alarm

Matthew — Tue, 25 Apr 2006 03:14:00 -0700

I recently found a doorknob touch alarm schematic while browsing Discover Circuits’ archives. The project was originally intended as a present for my brother’s dorm room, but a bad capacitor and the lack of a proper oscilloscope caused delays. It has not made it off the breadboard, and it probably will not until his next semester. The circuit contains a few basic elements, an flip-flop based oscillator, a set of delays, a flip-flop as a sensor, and the audible alarm.

The schematic

Download the schematic

The oscillator

The first section of the circuit is an oscillator based on a flip-flop. Clock and D are both grounded while Reset is tied high. Hence, the output Q will only be high if Set (node 6) is high. When the output is low, the transistor Q1* is cutoff. This allows *node 6 to be charged with a delay relating to the system of impedances R1*, *R2, R3*, and *C3. Once the voltage at node 6 triggers Set, the output changes to high and Q1* is opened. *Node 6 then discharges out through the capacitor. Once node 6 is low enough, Set is no longer triggered and the output is automatically reset (because R is tied high) to low and the process is repeated.

The screen capture below shows node 6 charging and discharging as the blue trace. The yellow trace is the output at node 1. You can see that the output turns high when node 6 reaches the switching threshold of the flip-flop (about 1.8 volts). Right afterwards it spikes up due to feedback through C2*, but quickly starts discharging. The oscillator switches off when *node 6 returns below the 1.8volt switching voltage. Feedback through C2* draws *node 6 to ground before the process repeats itself.

In order to change the period of oscillation, adjust the value at C3*. If you would like to make the pulses longer, adjust *C2. The circuit works best right where it is at, though.

The delay and ‘sensor’

The output of the oscillator is divided down two paths. The time constants of the two delays are nearly equal and can be adjusted with the sensitivity potentiometer. The path to node 11 is the Clock input of the flip-flop, and the path to node 9 determines if there is an alarm or not.

In the capture above, node 9 high than the the clock. Hence, the flip-flop stays high when the leading clock edge triggers it to lock. When the doorknob is touched, your body absorbs some of the charge and node 9 charges slower. This can be seen in the capture below. When the clock edge rises, node 9 is not high yet and low value is locked into the flip-flop.

The alarm

The designer uses an audible buzzer in order to relay the alarm. This is also my intent for the circuit, but I use a LED in my photos because you cannot see sound. They are both attached to the inverting output of the second op-amp (Q-bar) because it is high when the alarm is triggered.

There is an endless number of uses for this circuit, but I will just name a few crazy ideas:

Using the intended buzzer for your hotel or dorm room. (This is a bit more impressive than the old sock trick.)
Connecting the output to a relay that triggers the doorbell for you house. Just make sure to put it in parallel with your standard doorbell switch. That way you can still hear the Fed-Ex man. (This one has a major cool factor when someone opens your door.)
Tying the output into a security or home automation system. You could have the lights turn on as soon as you touch the door handle to scare the dog away from laying on the door.

Photo Gallery

The future of Job Interviewing?

Matthew — Thu, 20 Apr 2006 07:35:00 -0700

The Four Dees of Analog is a story written by ADI Fellow Barrie Gilbert about a job interview circa 2025. It is crazy to imagine being interviewed over live HDTV quality teleconference with noncontact stress monitors live at the site. The story goes on to describe the differences between simply collecting information and aquiring knowledge by relating and applying that information.

Of course, this is an article in Analog Devices’ Analog Dialogue so the Sci-Fi aura quickly fades into an advertisement for analog design. In the surrounding of a digital world, the story is a great reminder that analog instrumentation is still all around us. Additionally, it highlights the benefits and challenges of analog circuitry.

PhDs and mathematics

Matthew — Tue, 18 Apr 2006 07:28:00 -0700

Stolen from Now to offend some maths students on Mark’s Things*:

What is the difference between a Ph.D. in mathematics and a large pizza?

A large pizza can feed a family of four…

Getting started with the Atmel AVR

Matthew — Fri, 14 Apr 2006 13:58:00 -0700

There are two major microcontrollers in the hobby domain, the popular PIC and Atmel’s AVR line. PICs are tried and true in both commercial and hobby implementations, but the Atmel AVR offers affordable programming solutions, a free development environment, a free assembler and a stable gcc toolkit that work across the entire AVR line.

Affordable Programmers

The AVR is well known for how simple it is to program. To start, you only need a few resistors and a parallel port. This will burn your code onto the microchip, but if you plan on pursuing larger projects you will need a more complex programmer with ISP support. This will enable the extra features and debugging support.

The ultra low-cost AVR programmer uses an LPT port and a few resistors, but requires special software.
A simple serial programmer for Linux uses several resistors and two LEDs. Programming is done with uisp.
More complex schematics for ISP programmers that can be programmed via AVR Studio or AVR-Dude.
A usb programmer that supports ISP.

A free development environment

AVR Studio is available directly from Atmel at no cost. It has an integrated simulator and programming software. However, there is only an assembler available. You will have to purchase more software for high level languages.

The gcc-toolkit

If you want to use C/C++, then the gcc-toolkit is available for the AVR, free. WinAVR comes loaded with a gcc, binutils, the avrdude programmer, simulavr, and more. On Linux, you will need to download the packages for your specific distribution or compile your own cross-compiler.

A great community

AVR Freaks is a site dedicated to the AVR line of microcontrollers. There is a large archive of tutorials and a very helpful forum.

Charging batteries with Solar energy

Matthew — Mon, 10 Apr 2006 08:25:00 -0700

Depending on the application, charging batteries can be complex process. Charging methods range from constant voltage to pulsed and random charging. Once power is being delivered back into the battery, you have to know when to stop charging!

Once a battery is fully charged, the charging current has to be dissipated somehow. The result is the generation of heat and gasses both of which are bad for batteries. The essence of good charging is to be able to detect when the reconstitution of the active chemicals is complete and to stop the charging process before any damage is done.

Typically, common household batteries are charged with a current that is kept constant and relieved when the batteries reach a predetermined potential. However, solar cells typically generate a constant voltage of 0.5V and a varying current that depends on the amount of collected light. As such, a consant voltage charging model is easier to implement. I found two respectable tutorials on building your own charger:

AA Battery Solar Charger
Clean Power’s Solar battery project charger

The first solution uses a diode to stop the batteries from discharging when there is no sunlight. I highly reccomend including this protection. Unfortunately, neither project implements a charge limit. You have to remove the batteries and test their charge with a multimeter. A shunt regulator is the simplest way to regulate the upper limit.

New Oscilloscope: Instek GDS-820C

Matthew — Tue, 04 Apr 2006 15:32:00 -0700

If you know me on a more personal basis, then you know that I have been experiencing some issues with several projects/articles in the works for this site. Up until now, I did not have my own oscilloscope. Usually I am able to make due, but recently it has been making things harder than it is worth.

After some research I happened to find a deal on a new Instek GDS-820C DSO. The offer was impossible to refuse and cheaper than the GDS-805C. Besides the obvious increase in sample rate, the GDS-820C comes standard with USB and parallel ports along with the RS-232 port, but GBIP is still only an option. These digital storage options will allow me to post pretty pictures on the blog!

This oscilloscope samples at 100MSamp/sec and is capable of 150mHz operation. They claim an ‘effective’ sampling rate of 25GSamp/sec when viewing repetitive signals. In comparison, Tektronix base model samples at rates up to 1GSamp/sec.

At 150mHz it will be hard to see much if the signals are not repetitive! If you know anything about Shannon’s Sampling Theorem, the effective sampling rate makes 150mHz operation seem… possible. Without this hack, the scope would be limited to half the sampling frequency or 50mHz. I could have paid more for a Tektronix scope, but the hardware would have been limited to 40mHz, mono anyway. This will suit my needs.

Update: If you view a signal sampled at over 100MSamp/sec, then you can tell it is making due with the effective sampling rate. I am not sure if this really matters because I do not have access to signal over a MHz at the moment.

Vex Robotics Starter Kit - $179.99

Matthew — Thu, 30 Mar 2006 05:41:00 -0800

The Vex Robotics Starter Kit is currently on sale cheap at Woot.com. If you are interested, then you better buy impulsively. Their deals last 24 hours, tops. It is exclusive to RadioShack and normally retails for $299.99. There is also a large assortment of Vex Robotics Kits & Accessories at RadioShack.com..

PC Magazine has a review along with the infamous Myth Busters. Their main gripes were a lack of accessories and motors that were geared for speed instead of power. Combined with the weight of a steel frame, this hindered a their stair climbing application.

A rich system of hardware, software, and powerful sensors, along with documentation that doesn’t skimp, makes this a terrific first experience with robotics for anyone. Well-machined. Expertly documented. Quite powerful. Tremendous fun. – From PC Magazine’s Review

Compared to other robotics kits in its class, the Vex system is capable of creating some rather large robots.

Downtime & Lag

Matthew — Wed, 29 Mar 2006 14:45:00 -0800

I’m extremely sorry about all the downtime and lag lately. Dreamhost was trying be on the cutting edge by upgrading to Ruby on Rails 1.1 last night. Unfortunately, Typo does not currently support Rails 1.1. A post in the forums brought me back online using slow CGI this morning.

The good news is that the site is back up and running. Dreamhost has also rolled back to Rails 1.0 allowing Superpositioned to speed along on FastCGI again.

Unfortunately between this fiasco and a borked Typo upgrade earlier this week, I have exhausted what little spare time I had. I promise updates soon, though. My TODO list is full of circuits and projects!

Update: As of April 4, 2006 everything should be working again. Wait.. I am getting the “Typo failed to start properly” message randomly again.

Update: A few minutes later and you should no longer be receiving that message.

The Open Automation Project

Matthew — Tue, 21 Mar 2006 12:31:00 -0800

I am back from spring break. As of today, I should have significantly more time to dedicate to this site, unless school/real work builds up on me again.

Anyway, I just stumbled upon the Open Automation project. This ‘open source’ project has the following mission statement:

To fill the gap between the powerful mobile robot platforms typically used by researchers, and the small rug-roving robots with limited processing power that are popular with hobbyists.

The specific project goals are:

Design a coherent set of modular components (hardware and software) that conform to standards (where possible), and implement the functionality of an intelligent mobile robot. Use pre-built components that are readily available where possible (and when such pre-built components are affordable).
Minimize cost. It should be possible to build a robot for around the price of a PC (target: US$1,500 to $2,000). Consumer grade hardware components are to be used in preference to professional grade products.
Focus on stereo vision as the primary spatial sensor to produce useful space occupancy data. Central to the success of this project is the implementation of a functioning low-cost real-time vision system. The prevalence of FireWire-enabled WebCams and mainboards makes this goal reachable from the standpoint of cost; the difficult part here is the software.

With two firewire webcams for eyes, a mini-itx brain, 4-12 sonar sensors, an I2C speech synthesizer, an LCD, the drive and chassis, and more, the bill of materials is not cheap. However, the prototype is really cute. If you want an overview of how it works, there is a block diagram of the system.

Unfortunately, the project seems to be hibernating right now, and there is not much detail on what to do when you are finished building the it. I would really love to build my own robot, but I would not know what to do with it.

Always be wary of schematics

Matthew — Mon, 13 Mar 2006 08:03:00 -0800

Ladyada has an interesting rant on her new blog. In linear circuit analysis, we tend to ignore the difference between the positive and negative inputs. However, in practice positive feedback results in a bistable circuit. (Also read up on negative feedback.)

Usually everyone ends up remembering this detail the hard the way. While designing a tape head preamplifier she was utilizing this circuit from an application note:

After a few hours of staring at the circuit and debugging and wondering “man why the hell is this railing?” I finally look back at the datasheet and realize: oh its in positive feedback, of course its railing.

Just a reminder that the textbook (or pdf) isn’t always right!

Measuring the speed of light with Chocolate Chips

Matthew — Thu, 09 Mar 2006 23:41:00 -0800

Over the past week, I’ve been really busy with exams and projects. Trying to save time by finding the speed of light on Google, I stumbled upon an extremely interesting article on measuring the speed of light with a microwave. As any decent cook knows, microwaves do not heat evenly. In fact, this article explains their heating patterns are relative to the speed of light!

Understanding how a microwave heats

As we all know, microwaves heat using electromagnetic waves. These waves are at a frequency perfect for rotating water molecules (f = 2.5 GHz). The rotating water molecules create friction and thereby heat.

Two types of electromagnetic waves

Although there are two types of electromagnetic waves, we typically only consider traveling waves. The amplitude of the wave travels forward in position over times. The following animation demonstrates the amplitude of a wave over space and time.

The waves in a microwave are not traveling. If they were, it would be nearly impossible to distinguish any uneven heating patterns!

Standing waves in a microwave

The waves in a microwave oven are standing waves. These waves are stationary in space with an amplitude changing over time.

With this demonstration, it is obvious that particular sections of the chips are heated more than others. In fact, these locations are located half of the wave’s length apart.

The physics of waves

We now know the frequency of the microwave and can presumably measure the length of the wave, but how are they related to the speed of light? Simple. Electromagnetic waves propagate through free space (like that in a microwave) at the speed of light. Therefore, their length is related directly to the speed of light by λ = c / f where λ is the wavelength, c is the speed of light, and f is the frequency of the microwave. Solving for the speed of light, c = λ * f.

Where do the chocolate chips come in?

Chocolate chips are perfect for measuring the distance between melted spots. The heat does not spread as quickly through them because they are not uniform. This means the melted spots will be smaller and you will have more time to measure before they all start to melt.

It is hard to tell from the photos, but there were distinct melting spots almost exactly 6cm apart. Remember, this is only half of the wavelength, so λ = 12cm. Plugging all the known variables into our equation, we get c = 12x10^-2 * 2.5x10⁹ = 3x10⁸. Not bad! The true speed of light is 2.9987x10⁸.

Notes if you replicate this experiment

The chocolate chips only take 20-30 seconds to melt. The longer you have them in, the bigger the melted spots will be and the less time you will have to measure.
This will not work in a microwave with a spinning carousel. In fact, the microwave spins to counteract these effects. Usually, you can just flip the carousel upside down to stop it from spinning. (Thanks Ryan)
If you plan on putting the chips back in the bag, simply refrigerate them. Freezing causes them to stick to the plate.
You can microwave anything that melts. (Cheese or a chocolate bar) However, chips work particularly well.

Photo gallery

If radio goes through paper, why not light?

Matthew — Sat, 04 Mar 2006 19:41:00 -0800

USA Today was recently asked, “If both electric and magnetic fields can pass through paper why not light?” If you think about it, this is a really good question. The simple answer: paper absorbs/scatters light, but not low or high energy electromagnetics. Read the article if you want to learn more.

BigDog: A heavy duty quadraped

Matthew — Fri, 03 Mar 2006 19:32:00 -0800

BigDog is a quadraped robot being developed by Boston Dynamics for possible military usage. It is by far the most advanced quadraped around. The site has more mechanical details and a video! About 30 seconds into the video, he kicks BigDog and it stays balanced!

So far, BigDog has trotted at 3.3 mph, climbed a 35 degree slope and carried a 120 lb load. BigDog is being developed by Boston Dynamics with help from Foster Miller, the Jet Propulsion Laboratory, and the Harvard University Concord Field Station.Development is funded by the DARPA Defense Sciences Office.

A low power, long life LED flashlight circuit

Matthew — Thu, 02 Mar 2006 20:17:00 -0800

Scouring the Internet for information on LEDs, I accidentally stumbled upon a PDF detailing a flashlight made from PVC. For the torch, he biases ultra-bright LEDs with ballast resistors as described in my LED lighting guide. This is a simple solution, but a with a slightly more complicated circuit we can extend battery life by over ten times!

Download PDF instructions for the inefficient version

Electrical losses in the flashlight

In this flashlight, each LED/resistor combination consumes 4.5volts at 30mA or about 135mWatts. The ballast resistor alone consumes 1.1volts at 30mA or about 33mWatts. Therefore, 25% of the power being drained from the batteries is lost in the resistor of each LED chain!!

Battery life

The ‘C’ size batteries in the PVC flashlight have a capacity of around 4 500mAh. If you build your flashlight with seven LEDs, there is a constant current drain of 7 * 30mA or 210mA. Dividing this into our battery capacity, it becomes obvious this flashlight will only stay lit for about a day. This is a far cry from the 50-100 hours claimed by commercial flashlights running on smaller ‘AA’ batteries.

Cutting your losses

As mentioned above, the ballast resistors are wasting 25% of our battery power. Electronic Design proposed a simple circuit to resolve this in a recent article. The front end of their circuit draws less than a milliamp of extra current.

The circuit is best described in two parts: one, the boost circuit function of Q1 and Q2, and two, the control circuit of Q3 and JFET1. Assume Q1 is off. With the battery voltage slightly above Q2’s VVB, a positive Q2 base current [iB = (battery voltage VBE)/RJET1] would flow. Q2 turns on, which switches inductor L1 to ground.

The end result is a 23volt pulse (as shown in Figure 2) across the series of ultra-bright LEDs. At 278kHz, the human eye cannot distinguish the difference between these pulses and a constantly lit LED. This saves even more battery power.

As the battery voltage decreases, the pulses become further apart. The brightness remains indistinguishable until the voltage falls near 2volts. (The circuit does not function well below 2volts) I doubt the PVC flashlight has this efficiency near the end of its life.

The extended battery life

According to Electronic Design, this circuit consumes an current equivalent to about 17mA. Powered by our ‘C’ cell in the PVC, this circuit could run for 265 hours! This is ten times the original PVC design.

Podcast covering amateur radio

Matthew — Thu, 02 Mar 2006 14:14:00 -0800

Amateur Logic has been podcasting interesting projects including the active cantenna in episode 3. Their latest episode is straight from the Capital City Hamfest. There is an ARRL review of the amateur radio effort during Katrina, a presentation about Skywarn, and an inspiring amateur satellite radio demonstration. There is not much technical information, but it is a great introduction to what being a ham is about.

To learn more about amateur radio visit arrl.org or amsat.org.