As my back is killing my right now, I won’t be going anywhere anyways. So I might as well crawl around my home and measure how much energy I waste. Maybe this will bore the little devils sticking hot needles into my back to death and they go away. Even if walking upright was a major step in the evolution of mankind, it comes at a price…

 

The measurements:

 
Wattages of some other devices:

 
Now many folks may think that 10.5W of standby power isn’t a whole lot, but an average year has 8760 hours. Multiply that by 10.5W and you end up with 92kWh per year, senselessly wasted! According to my measurements that would be enough to have about 30 full loads of laundry washed and dried! If you like to think about it in terms of money, 92kWh would cost somewhere from 14€ to 24€ depending on where you get the juice. That’s definitely enough for going to the movies and maybe having some food as well.

Assuming that I’m “better” than average in terms of avoiding the standby trap, let’s just assume the average guy/dude/dudess/single/singless/family also has a big stereo, maybe an XBox or PS3 + TV, radio, alarm clock etc. etc. – all running on standby power. Let it be 50W total for sake of this thought experiment. Now take a country like the USA (0.30 milliard people, it’s NOT billion dudes!)

 
Short intermission:

See also here. BTW, short scale is crap and the world is metric!

Where was I… right the USA. Again 300 million people, maybe about 100 million households wasting the above assumed 50W. That’s a whoppin’ 120′000′000 kWh a day, 43′800′000′000 kWh a year or the output of a standard nuclear power plant (2GW) running for 2.5 years nonstop.

Fact: it takes about 334kJ to melt 1kg of ice of 0°C. Doing the math one gets: 1 kWh = 3600kWs = 3.6MJ, therefore the very large number from above gets multiplied by 3.6 to give MJ. Dividing that humongous number by 0.334MJ gives about 472 million tons of ice or roughly 0.5 km³, a cube with sides of about 800m. 800 meters are about the same size as the new Burj Khalifa sky scraper, which was completed earlier this year.

Gasoline has an energy content of about 38MJ/l. Wasting said amount of energy is equal to burning about 1′000′000′000 gallons of gasoline. That would take your average 25miles/gallon car 1′000′000 times around the globe.

All of this waste because of “I’m off, really!” a.k.a. “Standby”. Personally I’m more than willing to make the very small effort and switch off my gear when I don’t need it. Awareness to energy waste should be taught at school (and it could be fun too), as well as “Shut off the lights when you leave the room!”.

/rant

Mystery link!

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

I happened to see a pretty simple nitelite my mom likes to use. It comes with a base unit and a glass sphere resembling the planet earth on top of that. It only had a blue LED, and a pretty inefficient one as well. It was pulling 80mA and wasn’t any bright at all. The upgraded ‘thing’ now comes with an RGB LED and an ATmega168 microcontroller as the brains. The metal cap now works as a capacitive touch sensor.

As you can see in the video, as soon as you touch the base it changes color. In normal mode it slowly fades through all colors. When touched, it shows the delay time used for the fading encoded in red and blue. If it’s more on the red side the delay is large, if it’s on the blue side the delay is small. The longer you touch it, the more the delay is changed. I used an uint8_t variable to hold it, so it automatically wraps around to zero once it gets larger than 255.

Here’s a look at the somewhat messy innards:

The metal cap to the left connects to two pins of the ATmega168 microcontroller. Together with 2x 510kΩ resistors it forms a voltage divider. The cap has a small capacitance against ‘ground’ that can be charged through the resistors. The code of the CapSense library measures the time it takes until the ’sense pin’ changes its state from LOW to HIGH (RC-time measurement). By moving your hand closer to the metal base the capacitance is increased. In case the power supply is not grounded (wall wart), the measurement is somewhat limited, as the ground reference is missing. The difference is huge if you do the same measurement when connecting it to your PC (which is grounded). That’s why I had to set the threshold to 2 to get any result. When connected to properly grounded PC, the values reported by the code are magnitudes higher. It might have helped to install a copper ground plane, but that wasn’t possible. Together with the metal cap it would have approximated a parallel plate capacitor.

Closeup of the important parts:

The 1kΩ resistor glued to the ATmega168’s belly is used to bring the wall wart’s voltage down to a safe value. It’s a cheap one labeled to provide 3.5V. When loaded with the old blue LED it supplied 3.8V at 80mA. With no load it spits out 6V, which would kill the microcontroller. The resistor pulls about 5mA out of it which brings it down to about 5 volts, which is safe enough. The 10kΩ resistor is used as a pull-up for the RESET pin, 100Ω go to the red LED and the parallel pairs of 130Ω are used for blue and green.

Here’s the pretty simple code. It requires the Arduino CapSense library by Paul Badger.

A few words of caution:

Please make sure you’ve flashed a bootloader to the chip AND verified that it actually works BEFORE you start the solder job. I had flashed a bootloader, but as it turned out it was the wrong one. ARGH ARGH ARGH. I can recommend the ‘ATmegaBOOT_168_lilypad.hex’ one with a long enough wait phase. The other ones don’t give you enough time for manual upload. Fuse settings for internal RC oscillator: LFUSE:0xE2 – HFUSE:0xDD – EFUSE:0×00

As the chip is mounted belly up, it is VERY useful to draw a mirrored diagram of all the pins you’ll need to use + their function. If you don’t do this, you’re bound to make mistakes. Trust me on this.

Related posts:

1. IKEA SAMTID modification — part 1 I have 3 Ikea Samtid lamps. They’re affordable, useful and...
2. Jyetech AVR digital oscilloscope DIY KIT — Firmware upgrade odyssey I bought the DIY KIT at seeedstudio recently, as I...
3. Using a standard Arduino to program V3 Matrix boards A comment by Andy reminded me of the fact that...

Related posts brought to you by Yet Another Related Posts Plugin.

Almost 20 years ago I was lucky to have a math teacher willing to teach our class ‘all about’ complex numbers. At that time there were some changes to the curriculum and he could have skipped that or chosen something else. I don’t remember what the alternative could have been, probably stochastics (I just couldn’t resist) ;-)

One part of it was analytical transformations. Of course it wasn’t called that way back then at school and we weren’t told what criteria complex valued functions have to meet to be called analytical either, because that needed knowledge about derivatives, possibly integration and a whole lot about a guy named Cauchy. And of course calculating derivatives and integrals was part of next year’s math curriculum.

But we were taught quite a lot about

And of course there’s the famous:

BTW, complex math is the source of the world famous Mandelbrot set.

Here’s what I’ve found in the deep chasms of my mind today:

This little formula acts as a mirror using the dashed circle of radius 1 as the ‘axis’. Funny things happen.

If I’m not mistaken this transformation here is actually not analytical. If it were it should reproduce the 90° angles of the square perfectly. One textbook application of analytical transformations is calculating equipotential lines and electric fields in unpleasant geometries. As everybody should know field lines and equipotential lines are orthogonal. This fact is preserved if transforming geometries of electrodes into a more pleasant shape, meaning that calculation becomes easy or even trivial.

And the last one:

Here’s the files that were used to create the images. Some Inkscape post processing is necessary.

circle_mirror.pl , makeplot_circle_mirror.txt

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

Heating up water with electricity is madness. It costs a fortune and most likely fossil fuels were burned for that before. So why not use my gas/water heater and a simple bucket for getting warm water into my washing machine ? It only comes with a hose for cold water. The tricky thing is hidden in here:

This machine has a temperature sensor. Good. It’s also used by the on-board brain to lock down the lid if the contents are above 40°C. Why 40?

If you just fill the machine with hot water by hand and start it, it will pump out all of it again. To trick the machine, you need to start it and let it suck in some cold water and only then pour in more hot water by hand. If anything goes wrong the machine locks the lid. Of course that’s what happened. Fortunately the lid can be opened by removing power for 2 minutes. Unfortunately this also resets the microcontroller, but I found that one out too late.

Now all hot water will be pumped out again, but that is not what we want. As the machine is still hot inside tricking it by letting it suck in some cold water doesn’t help either. Because of the thermal lock-down the lid will stay shut. The lid must stay open while the machine runs in this case. Of course normally you’d just let it suck in some water and then add hot water by hand, but that only works once and with a cold machine.

Fortunately the machine can be tricked by pushing the lid lock switch by hand with the machine still open. Now you can add hot water. As the machine is hot and its brain thinks it favourable to lock the lid, the lid now cannot be closed anymore. Argh!

Outsmarted by a washing machine. But I’ve got sticky tape, HAH!

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

“Until you decide that NOTHING will stop you from success, EVERYTHING will.”

That one contains a lot of wisdom.

And now some images of snow (lots actually, I’m sick of it already), my new green 5mW DPSS laser pointer from Dealextreme, and some POV experiments. The laser pointer will come in handy once I’m done with my current occupation. I call it “The Bane”.

Riding my bike will be interesting in these conditions:

Losing some photons:

Much better:

Lissajous Figures. The first to calculate the frequency ratio and phase shift will get a little something:

More POV:

And now if you will excuse me, I’ll have to take a hot shower. Otherwise I’ll freeze to death outside. Probably I shouldn’t ride my bike, but at least it won’t be boring. Up till now I wouldn’t have thought that boredom in a more abstract sense could be so annoying – and that is an understatement.

Edit:

On second thought I shouldn’t call it boredom, but a massive shift of interest. The emotional subsystem of the human brain is as strong as ever, thanks to our distant ancestors, tiny rodent like mammals. Or in other words: RATS!

Related posts:

1. Happy New Year 2010 2009 has been a pretty tough year for everybody...

Related posts brought to you by Yet Another Related Posts Plugin.

2009 has been a pretty tough year for everybody it seems. I’ve had a big share of nasty setbacks myself. Let’s hope the next decade of the second millennium has more pleasant things in store for us. Let’s make 2010 the perfect year. Concentrate on what really matters to you and quit being somebody else’s dog. Bite back!

And now some real fireworks !

Related posts:

1. Happy New Year 2009 ...
2. Picture Dump — Jan 2010 New Year is over, time to post something new to...

Related posts brought to you by Yet Another Related Posts Plugin.

To my knowledge Kingelectronics is currently the sole distributor for Macroblock LED drivers in the western world. They have quite a few interesting chips, e.g. the MBI5168. It’s an 8bit shift register with included constant current sink drivers. Just one external potentiometer is enough to adjust the current from about 5 to 120mA per channel. With three of these you can easily drive 8 common anode RGB LEDs and simply balance the colors with the potentiometers. As the outputs are current limited, you can also parallelize them if you need to switch a bigger LED. Proper thermal design required. This chips is ideal for Arduino users wanting to drive somewhat bigger LEDs at a reasonable price. The MBI5168 chips are used in Seeedstudio’s Rainbowduino BTW. Other LED drivers are available as well.

If you find something of use in their list, you can join. But I only accept requests from inside the EU to keep shipping costs down. And as usual, I’m a hobbyist and no big business ;-)

Here’s the math for calculating the price per chip and the final one:

• price of a single chip in US$ as per Kingelectronic’s website.
• price of the chip in € including all import costs that I have to pay.
• shipping costs for importing the whole order (current estimate: 30$).
• shipping cost from Me to You in € (somewhere between 1.45€ and 3.00€).
• 1€ handling. carton, envelope, tape etc.
• number of parts for a certain component (i) (whole order).
• number of parts for a certain component (i) (yours only)
• import tax in % that I have to pay for the whole order (parts + shipping, 19%)
• number of different components
• component (i)
• €/$ ratio on January 14th 2010
• price in € for your parts including shipping from Me to You.

So the chip price is calculated from the original chip price + percentaged shipping costs and all import fees there might be. The percentaged shipping costs are weighted by the chip’s price. If the chip is cheap, only a small part of the total shipping cost will be added. If the chip is expensive, more of the shipping cost will be added.

Example order (no real prices, just easy to calculate):

• Part A – total order 100 pcs – 0.50$ each
• Part B – total order 10 pcs – 5$ each
• Part C – total order 1 pc – 50$ each

Each component contributes with 50$ to the whole order, therefore total shipping costs would be split into 3 equal chunks of 10$. If shipping cost 30$, the price (before tax and import) for Part C would be 60$, Part B: 6$ and Part A 0.60$. I hope that can be considered as fair.

You can chose other parts too of course.

Once again, this is a ‘hobby-level’ order and I won’t advance pay for 100s of dollars and only accept requests from inside the EU. As the paypal fees have gone up again lately, I’d prefer bank transfer for payment, which shouldn’t be a problem in the EU. Thanks to EU regulations, EU bank transfers don’t cost more than normal national transfers.

Cummulative group buy list as of Jan. 14th 2010 including the approximated chip prices in EUR according to the formulas above:

• 130x MBI5168 DIP16 –> 0.69€
• 115x MBI5168 SOP16-150 –> 0.69€
• 025x MBI5031 SOP24-300 –> 1.32€

ORDERED — Any news regarding the order will be posted in the comments section!

BTW, here are a few nice links: MathTeX, DeTeXify, No fake!

No related posts.

Related posts brought to you by Yet Another Related Posts Plugin.

The following graph is taken from the old! ATmega168 datasheet. It is not valid for the newer P or PA series. For the newer series there is also a difference between chips with different memory sizes. As usual, reading the datasheet is necessary.

Here’s the Arduino code used to find good values for OSCCAL using an oscilloscope. The current value of OSCCAL is sent to a terminal. Note that you’ll probably want to use your own pins, as I used my custom boards for this test. So better use pinMode() and digitalWrite() as usual for the LED. But on the other hand, if you already have an oscilloscope, you’ll probably know what you’re doing anyway ;-)

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53

#include <util/delay.h>
#include <stdint.h>
#include <avr/io.h>
#include <avr/interrupt.h>
#include "cxa_guard_acquire.h"
 
void
setup (void)
{
  Serial.begin (9600);
 
  DDRD |= ((1 << PD5));		// PD5 output
  PORTD |= ((1 << PD5));	// RED anodes HIGH. All 8 anodes go to this pin
  DDRB = 0xFF;			// all outputs
  PORTB = 0xFF;			// all 8 cathodes HIGH --> OFF
}
 
void
loop (void)
{
  unsigned int ctr1;		// using an int, so I won't have to check if ctr1 wraps from 255 to 0 in the for loop
  unsigned int ctr2;
  unsigned char osccal_tmp;
  static unsigned char osccal_factory = OSCCAL;	// see "cxa_guard_acquire.h" for a bit more info
  OSCCAL = 0;
  _delay_ms (2000);		// wait some time to let it stabilize        
 
  for (ctr1 = 0; ctr1 <= 255; ctr1++)
    {
      _delay_ms (20);		// wait some time to let it stabilize      
      osccal_tmp = OSCCAL;	// save current value
      OSCCAL = osccal_factory;	// change to factory to make serial work all the time
      _delay_ms (20);		// wait some time to let it stabilize
      Serial.print ("osccal: ");
      Serial.print (osccal_tmp, BIN);
      Serial.print (" - ");
      Serial.println (osccal_tmp, DEC);
      _delay_ms (20);		// wait some time to let it stabilize
      OSCCAL = osccal_tmp;	// restore
      _delay_ms (20);		// wait some time to let it stabilize
 
      for (ctr2 = 0; ctr2 <= 10; ctr2++)
	{			// blink blink for the oscilloscope
	  PORTB = 0;		// all on
	  _delay_ms (5);	// check the LOW period on the scope. should be about 5ms 
	  PORTB = 0xFF;		// all off
	  _delay_ms (5);
	}
      _delay_ms (20);		// wait some time to let it stabilize
      OSCCAL++;
      _delay_ms (20);		// wait some time to let it stabilize
    }
}

For code download click on ‘Projects –> GIT repo’ in the menu above.

I used the “measure the 5ms delay” method, as it was easier than measuring 125ns on my old 10MHz scope. I got values for OSCCAL of 104 and 122 for the two boards. That’s a relative deviation of 17.3% between the boards, using 104 as reference.

Final thoughts:

I still need to re-sync the two boards. Not much gained, but learned something new. It might have been a good idea to use the CLKO (clock output) pin of one board as the clock source for the other. But unfortunately that is on PB0 again. PORTB is used for driving the cathodes of the LEDs already. And I still want/need/crave a new digital scope. Maybe I should get a cheap 50MHz Rigol just so the ‘itching’ stops. I can still buy a good 4CH 100MHz one later.

Related posts:

1. Interactively controlled animation on a 64-pixel RGB LED display OK, enough with this pompous title. I’ll just show how...
2. Experimenting with the PORTC/D connectors on V3 boards: I²C First let’s have a look at the pinout of the...
3. Flashing the bootloader with USBtinyISP/avrdude First make sure the chip sits on the PCB properly...

Related posts brought to you by Yet Another Related Posts Plugin.

I took the schematic directly from the LT1618’s datasheet and modified the LTSpice model a bit. I moved the current sensing resistor from the input side (which has advantages as well) to the output side. That way it can be used as a constant current driver. The “SHDN” pin can be connected to V(in) as well, it doesn’t have to be 3.3V. For lack of a model for the 12V LED bar I just used a few other LEDs.

One interesting an annoying thing that showed up during testing is that the switching operation of the chip seems to confuse my power supply’s voltage regulation. More decoupling necessary. Luckily a battery doesn’t care about such things. The brightness of the LED bar was stable although I operated the converter without a current sensing resistor. I just couldn’t find 0.22Ω anywhere. So I will skip the LC output filter I had planned to add.

This circuit here operated a 12V 3W LED bar at 270mA. Input voltage was 4.3V and it was drawing a current of 1.35A. The chip was getting a bit warm to the touch.

It was also really nice to see how it regulates. Set a current for the LED bar and change the input voltage of the circuit. As power should be more or less constant, it automatically adjusted the current it needed.

Here’s a crappy video. I really should get a camera with autofocus that also works in video mode, not just _before_ you start the clip.

Just for the record: the LT1618 is a versatile chip, but I’ve just managed to kill it. And I don’t know how, maybe ESD. The failure mode is especially unpleasant. It starts switching normally for a second (the LED turns on) and then stops. But it just doesn’t shut down, no, it keeps drawing current through the inductor. I’ve looked at it with my old scope. First second or so: heavy switching, after that: several amps of DC current. That creates a lot of heat. I burnt my temperature probe while monitoring the chip for abnormal values. My pinky still hurts. Next time I’ll use the other hand. So now I’ll have to get me a bunch of these chips to keep going. The good thing is that the adapter board has survived ;-)

Related posts:

1. Seeedstudio Super White LED lighting bar 3W — first test I finally got my trusty digital power supply back. It...
2. RGB LED Toy — It works Well… sort of. I finally got the PCBs from sparkfun!...
3. Trying to fix the sound of my Sega Game Gear — part 1 Last time I turned it on, I didn’t get the...

Related posts brought to you by Yet Another Related Posts Plugin.

And I’ve already managed to solder the LT1618 boost converter chip to one of these. Compared to soldering TQFP-32 packages this screams for solder paste and a hot air station. You need patience and flux… LOTS!

This needs a very steady hand. Even with a quite pointy tip pin-by-pin soldering just doesn’t work anymore. Flood all of them, apply plenty of flux and use a clean tip to wipe the excess off. If I had to do many of them, I’d get solder paste and a hot air station.

The other side is for TSSOP8 chips.

If you need some of them, drop me a message (No headers, no chips, just the bare adapter PCB). But I don’t want to find any of these on ebay! This is intended for hobbyists who actually need these for prototyping and don’t want to pay the hilarious prices for adapters. If you need thousands, just download the gerber files and have them made.

Related posts:

1. Who needs cheap MSOP10/TSSOP8 to DIP adapters ? Hi, I’ve been planning to build a prototype driver for...
2. Note to self: 26.0V >> 2.60V — A short ChipQuik review Nothing good can come from a freak idea after midnight...
3. V3.04 PCBs from MakePCB have arrived I’ve waited for 5 weeks for the 2nd batch of...

Related posts brought to you by Yet Another Related Posts Plugin.