Andrew Que Sites list Photos
Projects Contact
Main
   I was suppose to order 5 prints, but only 4 arrived.  After I looked again at my order I found I forgot to get the last print.  Two of these prints were framed quickly, but the other two required me to pick up glass.  I did that this afternoon.  Once home I worked on getting the two prints I did have into their frames.  I found it was cheaper to order 20x30 prints for both these frames than a smaller print size from an other site.  So I did some calculations on border size before ordering the prints.  The idea was to place the image in the middle, then a black border around the image to denote the hidden frame channel.  After the frame channel border is just the white print area to be cut off.  This worked perfectly. 
   The calculation isn't complicated, but it has several parts.  First, the image is in pixels.  It is assumed the by the printer the image will fill the print size requested.  Two borders are added: one to denote the frame channel, one to cover the remainder of the print area.  In addition I wanted a 1/4" margin over overlap factored in.  All of this was placed in a spreadsheet for the calculation, and then checked in Sketchup to see if it translated correctly, which it did.  There are times I like the fact I'm pretty good at math. 
   Monday is wash day, and a good day to work on projects in the basement.  Continued my wood project and made a good deal of progress, but I do not have any pictures to share.
   Pictured are some cattle I found on my bike trip last week along Schneider Road.  It is interesting how such large animals can be so skittish.  They were quite interested in having a look at me, but would get startled and runaway regularly.  I am pretty sure the cattle I encountered were all young males.
   These frames have been built for about 18 months, but I finally got around to getting prints.  Both are hickory and the purpose of making the frames was an excuse to work with hickory wood.  There isn't much to the design—hickory is too hard for my blunt tool collection. 
   Work continued on the Odroid's analog inputs.  I setup a voltage divider so I could measure both voltage and current.  With this I can calculate total power consumption.  The values are going to require a calibration factor as the resisters tolerances and wire lengths are throwing the numbers off quite a bit.

October 03, 2014

Getting the I/O Shield to work

I/O shield switching relay

I/O shield switching relay

About 2 weeks ago I ordered the I/O shield board for the Odroid.  It contains digital and analog I/O, and I needed these for the control circuit.  Getting the thing to work was no easy task.  The documentation is very sparse, the system keeps changing, and seems to assume the user already has a lot of knowledge.  So figuring out what I needed to do to get all the pieces to work wasn't easy.

The first part I was able to get function was the digital I/O.  This is provided by a SPI I/O expander chip.  According to the forums, “Starting with kernel 3.8.19 (of today 27.03.14) i2c-4 is not enabled automatically anymore.” So that required this line “modprobe i2c_gpio_custom bus0=4,200,199”. Once that was done, I was able to follow the example in the thread about the I/O Shield.

# modprobe gpio-pca953x

# echo tca6416 0x20 > /sys/devices/platform/i2c-gpio.4/i2c-4/new_device

# echo 289 > /sys/class/gpio/export

# cd /sys/class/gpio/gpio289

/sys/class/gpio/gpio289# echo "in" > /sys/class/gpio/gpio289/direction

/sys/class/gpio/gpio289# cat direction

in

/sys/class/gpio/gpio289# cat value

1

My first test was to change the I/O pin to an output.

echo "out" > /sys/class/gpio/gpio289/direction

After this was done, I could toggle the pin.

echo "1" > /sys/class/gpio/gpio289/value

echo "0" > /sys/class/gpio/gpio289/value

I confirmed with my oscilloscope the pin was changing. I need digital I/O in order to drive the power source select relay, so I decided to build a circuit to switch the relay. This required my relay and a FET, which I quickly wired up on my breadboard. It worked perfectly, and the in signaled the FET which in turn drove the relay.

Getting the analog side of things to work are significantly more tricky, mainly because a lot of assumptions were made in the documentation about how an Arduino works. The biggest problem was I needed to have the Standard Firmata protocol installed on the Arduino chip before any of the examples would work. Once I did this, the digital I/O worked just like it said it was suppose to.

I now have all the I/O I need for the solar project. There is a question of design. Since the Arduino is an external processor, it could do all the averaging for me. In fact, it could handle all of the logic for power management. I want to think about this more before I make a decision. In the end, the Odroid still has to have some kind of driver to talk to the Arduino so it can ask about the power state. I plan on logging solar energy production and system power usage, so there will be some kind of software required for the Odroid to make this happen. What I like about the Arduino is the fact it uses an Atmel processor. I have been using a 32-bit Atmel CPU at work, and rather like it. They are very low power. I did a power check with both the plan Odroid, and the Odroid with the I/O shield. I found my meter doesn't have enough precision to show the additional current draw from the I/O shield. This means the less than 0.1 watts of additional power draw, and I can easily live with that.

The next step is to build the circuit. I am not going to have enough room on the I/O shield board to do this, mainly because the connectors won't fit. So I am going to order an external breadboard I can mount above the I/O shield and build the circuit there. The time is getting close to when I will need to order the solar panel and battery.

Xiphos doing repairs

Xiphos doing repairs

   Parts have arrived for the Solar Power Webserver, but today I have my woodworking project to work on.  I started by getting out my planner and making a large amount of sawdust.  I now have both of my pieces of walnut planned down to 1/4" and am pleased with the results.  I decided to sand the board I will be working with before work begins.  I figured it would be much easier to do this before I start cutting the board into small pieces.  The results will still require sanding, but not as much.
   When I got home Xiphos had had our drill all opened up.  The power cable feeding the main motor had come loss and overtime had built up a lot of carbonation.  This caused the drill to stop working.  So he opened it up, cleaned the terminals, and had everything reassembled and functioning in no time.

October 01, 2014

Happy October!

   October is here and the colors are bringing in the season.  There is not a lot of color yet, but I went out for a bike ride today to capture what there was.  The weather is suppose to take a turn for the worse the next few days so this might be the last time I get in a ride for a bit.  I did my standard loop, which is about 11 miles and takes me past Pheasant Branch Creek Conservancy and the town of Ashton.  I had a fairly strong wind from the south, but concentrating on photography I wasn't in any hurry.  I had hoped the sun would have played a little better with coloring the clouds, but no luck.  All of my pictures were fairly bland.
   Today's picture is Pheasant Branch Creek Conservancy just outside Middleton.  Once more of the trees turn this should be a pretty scene.

September 29, 2014

RAID drive dies

   This morning my RAID array shut itself down after losing a drive.  The Blue-Dragon has a 3 TB RAID-5 array consisting of 7x500 GB drives.  Drive number 7 was failing the SMART tests and needs to be replaced.  After some investigation I found this drive was the oldest in the array.  Since it is a RAID array all I have to do is replace the drive.  I quickly ordered a replacement drive, but would be unable to use the system until the replacement arrived.  I was pretty sure I had a 500 GB drive somewhere in my collection of unused drives, and after some searching I found it.  This drive is also dying and use to be the boot drive for the Blue-Dragon before I switched to a solid-state drive.  However, the RAID card refused to switch to using this new drive.  After a lot of fighting I was able to get the array back up, but it wasn't the simple "replace the bad drive" expense I was hoping to have.
   Pictured is the complete door.  The re-drilling is complete.  After the glue dried that held the plug in place, it was time to drill a new hole for the new door handle.  That went very quickly and the new handle fit just fine.  Now I will have to get use to having to use the handle rather than just pulling on the door to open it!
   When we left off, our group had just defeated a pack of Sahuagin that were attempting to sink the river boat taking us to Newport.  Marv was in the water and decided to investigate the Sahuagin's transport devices.  They were tentacles of some sort. 
   An other battle was emanate, and we needed Marv back on the boat.  Annalis fumbled a search for some tools that might assist in hauling Marv back in.  Scar messed up an attempt to shoot a ballista into the tentacle holding onto Marv, almost hitting Marv in the process.  Andal fell of the mast and into the water trying to get down.  With everything against us, Ellenoria tries a bluff to improve moral and succeeds.  Marv fails a rope skill, but with an action point was able to succeed on pulling himself closer to the boat--or the boat closer to him.  Annalis balances on the rope and makes it to Marv, and Scar tries to pull Marv in.  Andel tries to swim back to the boat but it isn't working.  Ellenoria learns the monsters are Water Weirds from the other sailors and it is known that Sahuagin often ride them into battle.  Andal fails to find the boat, thus failing our skills challenge.  We were now in a worse spot than when we started.  Marv, Andal, and Annalis were all off the boat.
   The Water Weirds climbed onto the boat and the attack  begins.  Ellenoria takes a lot of damage the first round and runs for the mast.  Sparky bloodies a Water Weird in the first round using his special lighting attack, and it is clear they have a weakness against electrical attacks.  Marv also does some heavy damage in his first round, with Jam making a decent attack as well.  Scar gets an attack in that doesn't do much damage.  However, Andal calls lighting using Sparky as a source and really puts on a show.  Soon there were three dead Water Weirds killed by Jam, Scar and Annalis.  One of the enmey fumbles and takes electrical damage from Sparky for this mistake.  After this failed attack, Sparky blows the enemy up with an electrical attack.  Annalis ends up with a natural 20 and destroys the enemy she is fighting.  Ellenoria then sings this song to boost attacks:
Yo ho, yo ho, the fighter's life,
The a flag o' skull and bones,
A merry hour, a hempen rope,
And Marv kills Davy Jones.
   Our setup today made use of my project the other day where I installed LED strip lighting to our game board.  This illuminates the table and makes it easy to see one's character sheet.  Everyone was pleased with the results.
   Worked on several projects today while partly watching some historical documentaries.  I have 5 picture frames that require prints, 4 of which were completed but never had a print and one that broke the glass during the move and scratched the print.  While I thought the process of selecting prints and ordering them was going to be quick I was mistaken.  One of the frames in particular ended up with very strange dimensions.  I had to figure out how to offset a print such that I could get it to fit in the strange size.  In addition, I had designed a couple of frames for print sizes I can only get from one print house.  But the prices are much higher than my usual place.  It is cheaper to order a larger print and trim the print down.  Wanting to get it right I spent much longer than I probably should have needed verifying the math.  However, prints have been ordered.  More projects are now in the queue, but none should take too long.  I think I'm going to have to find more work to do.  It is fall though, and the daylight hours should be more occupied with photography.