The new boards have arrived and it looks like we have a winner. Now time to write some firmware….
Small changes to initial conditions.
The first microcontroller project I ever built was a Nixie clock using two bargraph tubes. One for hours and another for minutes. This used a Adafruit trinket, RTC and high voltage supply. All mounted on a nice antique clock movement. Is was pretty nice and featured on the front page of Instructables. Aces.
Well recently it was dropped and while still turns on, it no longer keeps the correct time. Now a couple of years have past and I know a little more I thought I’d rebuild my first project.
So here’s the first update: board designed with Design Spark PCB (reduced version of Easy PC) and ordered.
It uses a Kinetis KL03 and the same HV supply as I did before, I just put it on the same PCB. So boards on the way from China and a box of bits from Farnell sitting on my desk… tbc
As alluded to in my previous post, I have been working on a coffee table. The idea was to use the idea of carving waves into wood but make something also functional as well as nice to look at. I bought a beech kitchen work top and decided to use that as the table. I also bought four hand made hair pin legs to support it. Beech carves amazingly and was a dream to machine. So it was a simple job of routing out the ripple waves. The surface was finished with sand paper all the way up to 600. This basically left it mirror smooth. I also ditched my usual Tung oil in favour of something more hardwaring and decided to try out Osmo polyx oil. This stuff is amazing, a little more pricey but goes a very long way so actually cheaper in the end. Plus the finish is fantastic. Pretty pleased with the results.
I have recently become interested in carving wood, not by hand but on the CNC mills at hackspace. I got the idea after carving the saddle potentials for the ball traps. The lime wood gave such a lovely smooth curve that I thought it would be nice to carve more functions into wood.
So I bought a large piece of rough sawn oak, levelled it and carved a wave patten across it’s surface. The pattern was generated in Mathematica. I wrote a fairly random 2D function and tweaked it until I came up with something I liked. Exported it as stl and used the large CNC router to do the wood munching. Turned out really well (the contours can be a little hard to see in the pic):
I want to hang it on my wall but being quite heavy and living in a rented house that’ll have to wait. So I tried something a little smaller. I had a lime carving block left over from the saddle so I carved a bowl shape. Again generated in Mathematica. This had much finer detail than the large piece but also worked really well.
Excited by all this I now have a garage full or wood waiting to be carved! The next job, an oak beam being turned into a garden bench, with wave carved seating…
I wanted to make a Christmas present for my girlfriend this year so I decided to make a RGB bedside light. The LEDs themselves are cheap enough of eBay as are constant current drivers.
The LED I chose was a 10W RGB, so 3W red, 3W green and 3W blue. To drive this rather over spec’d lamp I decided on the PT4115 LED driver costing a massive £0.99 for 10. These just require a resistors, inductor, capacitor and diode. With a simple circuit diagram for 3W LEDs on page 11 of the spec sheet. The brightness can be controlled by PWM from a microcontroller directly to the IC. Whipped up a board in Eagle and bought 10 PCBs from China.
The PT4115 is an SOT89 so a little fiddle to solder with the massive tips at the Hackspace but manageable. You can see I have put three on one board, one for each colour.
Now it was simply a case of soldering the board to the LED and plugging in my arduino.
And yes it was. However, what I was surprised at (but shouldn’t have been) is how hot these get. Being used to little LEDs I don’t think of them as giving off any heat. However, whilst they may be extremely efficient, say 80%, then that’s still 2W of energy being turned into heat. It was about this time the first LED stopped working. Time to find a heatsink.
First heat sink found in a box of bits…
This was pretty effective and prevented the LED overheating. However, after about 10-15 min it became too hot to touch. Next heatsink.
This one even had a fan on it! This beast did nicely.
For permanently mounting the light decided to laser cut a ply wood box, the top would hold the heatsink and beneath would be the drivers and a controller. To control the colour of the lamp I opted to use two pots. One pot would control the overall brightness of the lamp and the second would control the colour. I used an Adafruit Trinket I had spare. In fact the first microcontroller I ever bought. I never liked it much though as I could never get it work on Windows only on Linux, which I don’t often use. So a good project to get rid of it in. Fortunately it has just the right number of pins, two analogue inputs and three PWMs. It also has a power regulator so I can take power directly from the 12V supply.
Here’s a pic of the laser cut box.
There are two lines on the right face, the top is a slot to hold a support across the box to hold the heat sink. The second lower slot is a vent to allow air to circulate from the bottom, up through the heat sink and out the top slots.
To act as a diffuser/lamp shade I simply used a sheet of nice paper. Folded it into a rectangle and placed it on top. It works really effectively. Here’s a pic.
It has been a while since I posted anything, mostly because I have done nothing other than work on some demonstration experiments for my old phd supervisor. These are now finally finished so here we are.
First a little background. I was asked before Christmas to build 2 demonstration experiments to show undergrads, prospective students and interested members of the public the fundemental principles of trapping single charged atoms, or ions. The first of these experiments is a mechanical analogy, a saddle shaped peice of wood/plastic/something representing the electric field. In one direction it will trap the ball, since it can’t roll up a hill. However, in the other direct it will happily roll down the side. To prevent the ball rolling off the saddle it is rotated, and when the rotation speed is just right the ball will oscillate back and forth inside the saddle, effectively trapping the ball. This is exactly the same principle as how atomic ions are trapped inside an ion trap, except swap the wooden saddle for an electric field. The second experiment is just to build an ion trap but to trap dust particles. An ion trap is a very simple device, it consists of 4 electrodes, which are four parallel rodes each on the corner of a square. Then two more acting as end caps. This arrangment of wires can be used to trap charged particles ranging in size from specks of flour to signal atomic ions. To trap big heavy particles of flour we need a lot of voltage, about 6000v should do it and oscillating at tens of Hz… 50Hz is convenient.
So as far as building these devices they offered two different problems each, the saddle trap was a) how to make a saddle b) how to rotate the saddle. For the dust trap a) how to generate such a large voltage and b) how not to kill anyone.
To address the first issue of how to build a saddle I had two options additive or subtractive fabrication. I quickly ruled out the 3D printer due to the poor surface finish and opted on a piece of kit I have never used before, a model mill. I managed to get trained up and after test with foam I managed to mill out a beautiful wooden saddle. It took about 2.5 hours but the results were lovely and required only minimal sanding. See this video for a time lapse of the milling. For protection from all those grotty little fingers I just gave it a couple of coats of tung oil, giving it a nice wet look finish. The saddle completed I dug through the trove, a room full of “junk” at the hackspace, this stuff has all been kindly donated and if it hasn’t got a stick you can use it. In amungst all the stuff is a couple of large boxes full of fans. Mostly PC fans but all sorts. Out of here I found a 12v PC fan and a 240v mains fan. The mains fan being designed to run from AC has a lovely smooth bearing and will rotate freely, none of the magnetic steps in DC fans. This would act as my bearing to hold the saddle. To drive the saddle the 12v fan would drive a spindle and a belt would connect the two. To control the speed of the saddle the current to the fan would be controller by the PWM output from an Arduino, this would also show the target speed on a little Nokia LCD. Nice. Now just put the whole thing in a laser cut box and we’re in business. (There were a few more trial and error stages that the author may or may not have omitted.) In the end I was asked to build two, this time with a larger saddle. Fine except for a 4.5 hour mill time! However the results have been awesome and I’m gonna have to build myself one! Here’s a video of the large saddle in operation.
A rotating wooden saddle is capable of trapping a ball in centre, preventing it from falling off.
For the dust trap my main concern was safety. If this is to go near the general public it can’t kill anyone! After playing with lots of different transformers and audio amps I borrowed a microwave amp and I was in business. These have the benefit of being built for purpose. They are designed to generate the voltage I need. However, they also will kill you just if you look at them wrong. My solution, current limiting resistors. Lots of them. Since all I am interested in to trap dust is voltage then the current can be arbitrarily low. The trap itself acts as a very small capacitor, the impedance of this capacitor at 50Hz is huge so placing 10Mohm in series before the trap makes virtually no difference to the trap voltage, it does mean though, should a stupid finger get between the electrodes so little current can be drawn no one will die. Also encasing the electrodes in plastic helps. This works great for trapping dust, flour, cumin and best of all baking powder. Though strangely mustard power doesn’t trap. Here’s some pics and a video.