Andrew Que Sites list Photos
Projects Contact
Main

January 24, 2015

Mad Scientists on Vacation

   Tonight was an other Elmwood Park themed event: Mad Scientists on Vacation.  Everyone who wanted was to dress as a quirky scientist ready to hit the beech at the Bikini Atoll.  I decided I was a young Dr. Merkwürdigliebe (that's Dr. Strangelove for those who are wondering).  Xiphos did a great job of pre-party planning and we had a great set for the evening's festivities.  A decent number of people showed up and several in costume.
   My polynomial regression site had been more popular than the King's Quest game walkthrough, but this month they are about tied.  I decided to see who was linking to the polynomial regression site and found it had been cited in a master's thesis paper for someone at the Budapest University of Technology and Economics.  Unfortunately the author says he did not use my library but had examined it for his work.  I am pretty happy to know my library was considered as a tool for someone doing research—enough so it was cited in a paper.  Comments made on the online regression calculator page also show that people have been making use of the site.  One person says it helped them with an engineering project as they needed to calculate heptic (an 8 coefficient polynomial) regression.  I wasn't even sure what applications would need such high-order polynomials when I wrote the library, so I am glad to see there are uses.
   My bread boards have arrived at last, and I can start on my latest project.  I want to build a controller for some high-wattage LEDs.  After reading about how the circuit works I decided to give it a short.  I ordered an LED controller, some FETs, coils, diodes, capacitors and load resisters.  The controller will run directly from 120 VAC, and here is the first state.  A thermistor feeds AC into a bridge rectifier.  This creates full-wave rectified DC power that then feeds into a filter capacitor.  I put a 100 k resister across the cap for a small load, and sure enough I now have 170 volts of clean DC power.  The thermistor is to keep the inrush current for when the circuit first turns on to a tolerable level
   There is a lot of power available here, which is why I am doing this test.  The initial test will run a 10 watt LED, but I have been looking at 30 and 100 watt LED chips.  Running these devices directly from 170 VDC is the most efficient method to power them.  And if I have success Xiphos and I have some projects in mind we can do.
   An other view of my basement work area looking mostly at my work bench.  My entire setup in the basement uses LED lighting as the other day I retired the last of my compact florescent bulbs, and I have not used incandescent bulbs in several years.  The mess on my bench is typical.  Not too messy to be unusable, but not teribly organized either.
   This view of my main work area shows my primary work bench illuminated.  There are two rows of lights.  The first is under the bench and consists of some bright LED strips.  The benches are typically illuminated by LED reels, but these are mostly for basic navigation.  When one needs to work far more light is required.  So these LED strips take care of that.  The second row of lighting comes from track lights which are focused on the edge of the bench.  This help cover areas not as well lite from the under bench lights.  The track lights are mostly 3 watt LED bulbs, with one 15 watt bulb right above my main work area. 
   This is a view covering a little over 90° of my work area stitched together.  I put it together while watching a bad action movie from the 1980s everyone says is pretty good.  There are about 7 photos in the one finial.  Getting the seams hidden was quite the task.  It is fairly easy to do on large low-detail areas like the floor.  But in the areas with more detail matching becomes tricky.  It gets especially tricky when objects are separated by distance as well because they move at different relative positions as one turns.  However, I've been stitching pictures together for some time so while it takes awhile this stitch was nothing new.
   My colored LED light bulbs arrived today and I retired the compact florescent bulbs that had been in all my fixtures.  The power draw was reduced from 13 watts/bulb to 5 watts/bulb.  In addition the LEDs have no warm up time.  This was most noticeable on my yellow bulbs although they all have a warm up period.  In the winter it is much more noticeable because the basement stays around 65° F (18° C).  The bulbs came direct from China and took a couple weeks to arrive, but the cost is about the same as I pay for the color CPF bulbs I can buy locally.  They do not dim, but that is not required for my main lights. 
 

This is an BeMicro MAX 10, an evaluation board equip with a Altera Max 10 FPGA. I have wanted to experiment with a Field-Programmable Gate Array (FPGA) for some time. These devices are like liquid logic in a chip—you can make them into a huge number of devices including CPUs. About a month ago I came across a project idea that was a perfect application for a Programmable Logic Device (PLD—and a FPGA is a type of PLD). The project is simple enough that I don't need a complicated device, but I figured I could get a good introduction into FPGAs if I were to use one on this project. After consulting with an electrical engineering coworker I started searching for devices. I found the BeMicro MAX 10 evaluation board had an FPGA far more powerful then I required but was a great price. So I could try my simple project on this board and expand when the project was over.

Unlike Arduinos and Raspberry Pi's, Arrow Electronics doesn't seem tailored to hobbyist. They wanted me to have an account on their web site and tell them about my company. Clearly not designed for people who want to tinker. The development tools form Altera were the same way.