Eye-Fi Mobi Wireless SD card demo and review

I show how the Eye-Fi Mobi Wifi SD works and give my impressions after shooting three shoots using the card.


Lite Brite LED clock published in Make Magazine and a new writing gig with EmbeddedRelated.com

I'm on page 135 of this issue!
I'm published!

In August of last year, I was offered the opportunity to write an article for Make: Magazine regarding my Lite Brite LED clock.  I wrote and submitted the article and even got a nice check from O'Reilly publishing.  A couple of weeks ago, over a year later, I receive a mysterious manila envelope in the mail - my article was finally published!

I eagerly found my project in the table of contents, then flipped to pg. 135, expecting to see my Lite Brite LED clock in all it's glory.  I found a beautiful, full-color picture of someone else's Lite Brite LED clock.  Make actually built their own Lite Brite LED clock, presumably from my instructions.  It was strange looking at someone else's Lite Brite clock.  Until then, mine was the only one I had ever seen.  After I thought about it, I decided it was really cool that other clocks had been built from my idea.

Going to the article, there was just a quick intro to the project, then a link to the website for more info on building the project yourself.  I did get a nice bio in the bottom right.  All in all, it was a total win - I got a full page in the premier publication for the maker movement. 

It was also neat to see my name in the front of the magazine under 'Contributing Writers'.  My name was printed alongside the likes of Limor Fried, Massimo Banzi and Forrest Mims.  Just the Entrepreneur of the year for 2012, the founder of Arduino and the best selling electronics author of all-time.  Fun!

A few weeks ago I was also offered the opportunity to write for another electronics blog - EmbeddedRelated.com.  Interestingly, the current bloggers of EmbeddedRelated are all electrical or software engineers, with the majority of them having either a Masters of PhD in engineering.  Not really my peer group or intended audience, but I guess Stephane (founder of EmbeddedRelated) wanted to bring a little hobbiest flavor to the site.  You can see my EmbeddedRelated.com content here.

I am going to try blogging at EmbeddedRelated for awhile for a couple of reasons.  Writing for EmbeddedRelated will allow me to make a little extra money and reach a different and bigger audience.  I will still blog on MeanPC.com, but for the time being it will probably be more editorial writing than technical writing.


Are LED light bulbs worth the extra cost?

Cree LED lightbulb vs. GE incandescent light bulb

Are LED light bulbs worth the extra cost?  How long do they take to pay for themselves?

I saw that Home Depot is now carrying Cree LED light bulbs.  I know that Cree is probably the best LED manufacturer in the world, so I decided to investigate a little further.  Not long ago, the price of LED bulbs was just too high to even consider, but prices have come down.

You can currently buy a 6 pack of Cree 9.5 watt LED bulbs for $74.82 at Home Depot's site.  These bulbs are dimmable and rated to last a whopping 25,000 hours.  We'll put these bulbs up against some GE 60 watt soft white bulbs from Wal-Mart.  $5.29 for a 4-pack.  So, the LED bulbs are $12.47 apiece for a 25,000 hour bulb and the incandescent bulbs are $1.33 apiece for 1,000 hours.

It's clear to me that over a 20 year period the LED bulb will save a ton of money for us.  But most of us are a little more short-sighted than that.  Especially when it comes to paying a lot of money for something that used to be a trivial cost.  Since we are so impatient, let's see if the LED bulb can pay for itself in just one year.

Let's look at a heavy use case scenario first.  I have a home office which I work from for 40 hours a week.  I like to keep the room nice and bright during my work hours, and I cannot stand fluorescent bulbs.  I run three 60 watt bulbs in a light fixture and a 60 watt desk lamp for 40 hours a week.  Let's do the math.

Here in southern Louisiana, residential electricity averages around 12 cents/kilowatt hour.  According to data at the US Energy Information Administration, this is consistent with national prices.  

Incandescent calculations - first year (and every year)
60 watts X 4 bulbs X 40 hours = 9600 watt hours or 9.6 Kwh per week.
52 weeks X 9.6 Kwh =  499.2 Kwh for the year
499.2 Kwh X $.12 = $59.90 energy cost for 1 year
In a year, we will consume 2 4-packs of these bulbs, if we are a little lucky.
8 bulbs = $10.58 in bulb cost
1 year incandescent bulb cost is $70.48

LED calculations - first year
9.5 watts X 4 bulbs X 40 hours = 1520 watt hours or 1.5 Kwh per week.
52 weeks X 1.5 Kwh = 78 Kwh for the year
78 Kwh X $.12 = $9.36 energy cost for 1 year
In a year we will use 4 LED bulbs
4 bulbs = $49.88
1 year LED bulb cost is $59.24

In the first year, the "expensive" LED bulbs will actually save us $11.24.

But wait, there's more!
After the first year, these LED bulbs are paid for and it's all gravy from here.  Let's look at a 10 year cost comparison.

Incandescent bulbs - ten year cost
10 years X $59.90 energy cost/yr = $599.00
In ten years, we will consume 20 4 packs of these bulbs, if we are lucky
80 bulbs = $105.80 in bulb cost
Ten year incandescent bulb cost is $704.80

LED bulbs - ten year cost
10 years X $9.36 energy cost/yr = $93.60
In ten years, we will still be on our initial set of Cree LED bulbs
4 bulbs = $49.88
Ten year LED bulb cost is $143.48

After ten years, the LED bulbs have saved me over $500!  That is starting to add up to some serious money.  This scenario is a high use case, but some of you that have kids might save even more than this if they are bad about leaving the lights on.  Outdoor lights can save even more money than my example.  Just two 60 watt light bulbs on your porch for 12 hours/day, 7 days/week.  Try the math for yourself...

Depending on your local weather trends, you could also save a nice chunk of money on your cooling bill.  Those 60 watt incandescent bulbs - each one is a little heater.  The LED bulbs generate almost no heat.

There can also be those bulbs that can be really hard to change.  If you have vaulted ceilings, it can be a major PITA to have a bulb burn out 20 feet above your head.  I've also noticed that incandescents that tend to get turn on/off a lot burn out more frequently.  I'll bet we go through 3 or 4 bulbs in our closet even though it is only on for about 10 minutes a day.

LED's are a no-brainer, right?


Powering cell phones, tablets and laptops with internal combustion engines

Since 1980, personal computer processing power and harddrive storage has roughly doubled every two years - as predicted by Moore's Law.  Which technology has been stagnant in that same amount of time?  Battery performance.

The current battery technology, Lithium-ion, was invented in 1979 and brought into mass production by Sony in 1991.  According to this article at Battery University, Lithium-ion batteries are only twice as powerful as the Ni-Cd batteries that preceded them.  Lithium-ion batteries are expensive to manufacture and their storage capacity degrades relatively rapidly.

So why hasn't battery technology kept up?  In fairness, battery performance was not on the radar before the current boom of smart phones, tablets and laptops.  For years, computing was typically restricted to desktop computers.  Cell phones were just used for phone calls, and could go days without a charge.  A spirited game of Snake just didn't take too much out of the Nokia.

Recently, cpu manufacturers have shifted their focus to put a premium on designs that are energy efficient for the new landscape dominated by mobile devices.  At this point, battery performance is the limiting factor in mobile processor design.  

Fuel-cell technology has been right around the corner for about 15 years now.  An article from the MIT Technology Review in 2001 had a headline that claimed "methanol-powered micro fuel cells are racing toward market".  The only commercial product I've seen so far is the Horizon Fuel Cell Minipak.  The Minipak is about the size of your hand, a full inch thick, and costs $99.  The Minipak stores 2 watt hours of energy.  The typical cell phone battery stores over 5 watts of energy.  At this point, mobile fuel-cell technology is not even close to viable.

While cutting grass the other day, I started thinking about internal combustion engines.  I cut a yard that is almost an acre with a 6 horsepower self-propelled mower.  It takes a little over 2 hours for me to cut the yard each week.  I can cut the yard 3 times on just 2 gallons of gas!  The reason internal combustion engines and fossil fuels are still around is that they are so efficient.

The 1 inch square microdrive, released in 1999
Do you remember the IBM microdrive?  In 1999 IBM released the 170mb Micro-drive.  Increasing demand for digital storage for digital photography and music drove this market to look for a different solution than the compact flash technology of the time.  

Remember, this was 14 years ago.  IBM managed to fit a hard disk platter, spindle, motor, read/write heads and hard disk controller into a package that was 1 inch square and about 1/5 of an inch thick.  It wasn't until 2006 that solid state compact flash storage capacity surpassed the microdrive and rendered the mechanical technology obsolete.  I though the microdrive was amazing then, and I find it equally amazing a decade later.

If IBM can fit a full harddrive in the space of a compact flash card, then why can't a micro internal combustion engine be developed to power our phones and laptops?  My lawnmower engine can produce 6 hp, and it's about a cubic-foot in size.  6 horsepower is the equivalent of 4470 watts.  A cell phone uses around 2 watts. 

Why not just make an internal combustion engine to run our phones, tablets and laptops?  A phone would need an engine 2000 times smaller than my lawnmower.  Why doesn't IBM just aim their magic shrink ray at a lawnmower engine and fit it, and it's fuel, in the space of a compact flash card?

The micro-rotary engine, developed at Berkeley in 2001
It turns out this was attempted in 2001 at the University of Berkeley.  In a research project funded by DARPA, research scientists at Berkeley attempted to build rotary engines the size of a penny.  Theoretically a great idea - the rotary engine has much fewer moving parts than a stand internal combustion engine.  

According to the documentation of the Berkeley research team, insufficient manufacturing tolerances resulted in engines that had poor combustion due to 20% leakage.  These motors scale fine performance-wise, but as the size decreases, tolerances for the miniature machined parts decreases also.  The paper does not state the precise reason for the insufficient tolerances, so it is not clear if the tolerances were not achievable in general, or if THEY were not able to achieve the tolerances with the funding and equipment they hand on hand.

At the end of the findings, the researchers state that "the results obtained to date with small-scale engines indicate that there are no fundamental phenomena that would prevent the development of the micro-rotary engine."  Further, "The challenges facing the fabrication of the micro-rotary engine include manufacturing tolerances, wafer alignment, assembly, and materials. Fabrication of the micro-rotary engine has begun at U.C. Berkeley."

DARPA pulled funding from the project the same year this paper was written.

Imagine a laptop that can run for a week on just a few squirts of fuel.  If IBM could manufacturer the amazingly tiny and precise microdrive 14 years ago, surely a micro-engine that makes almost no noise and produces negligible emissions should be achievable.  Surely this technology is in somebody's skunkworks.


New Arduino robot kit unveiled at bay area Maker Faire

Arduino Robot

I was surprised to see the announcement of the Arduino Robot yesterday.  I guess I haven't been keeping up with Arduino news very well lately. Julian Horsey at Geeky Gadgets actually got his hands on one, as they were available for sale at the bay area Maker Faire.  

There is a link to the Arduino Robot at the Arduino site right now, but the robot won't be available through normal channels until July.  You can at least read a few of the specs and get a look at the hardware.

I haven't seen an official price on this robot yet, but I've seen rumors on Google Plus that the price will be $275.  I'm really hoping this is just a rumor, because that seems extremely overpriced.  

I do think an official Arduino robot is an awesome idea.  I've always thought a curriculum based around robotics, or more specifically, Arduino robotics, would be incredible.  A robotics curriculum could be tailored to encompass math,science, computer science, physics and shop - all in one.  If you have groups design and market a project, you could even throw free enterprise into the mix.

I'm not sure this Arduino robot is the best design, based on what I've seen so far.  While there are some things I really like, there are probably more things about the Arduino Robot that I think could have been done much better.

Things I like - the LCD panel, control pad and built in speakers are really nice for input/output.  I haven't used one of the Arduino branded LCD TFT's before, but it looks like a pretty good screen, and it has a nice library backing it up.  I also like the 3 pin headers going all around the robot.

There are two major negatives to this robot, however.  The micro controller and the the motors.  

The micro controller is an Atmega 32u4.  This is a fine controller, and is the same chip used on the Arduino Leonardo.  I would have much preferred to see Arduino use the Atmega 2560, the same chip used on the Arduino Mega 2560.  The Atmega 32u4 only has 32K of program space vs. 256K on the Atmega 2560.  There are also a ton more input/output pins on the Atmega 2560.

I initially thought that the Arduino Due would have made a better processing backbone for this robot, but the combination of being a 3.3V system and having sketchy library support would make it a bad choice.

I would normally consider the Atmega 32u4 to be a fine choice, but by the time pins and program memory are taken up by driving the motors and LCD screen - is there enough left over to have fun with?

One other major flaw I suspect are the drive motors.  I have no idea what motors they are yet, but from photos they appear to be cheap toy-class motors in white plastic housings.  Maybe they perform and survive better than they look like they will.

If this is a $100 robot kit, then my above criticisms are not really warranted.  If this is a $200+ kit, then I think Arduino could have done a lot better.  Until we hear what the actual price is, and I find out how good the motors are this is nothing but conjecture.

An Arduino Robot is a great idea, I'm just afraid this might be sub-par implementation.

The greatest thing about Arduino never has been it's hardware though.  It's always been the unified support of it's community.