Transition to WordPress

As of today, this site is running on WordPress. Everything went fairly smoothly, and modifying the theme was straight forward. I didn’t want to go through the hassle of updating my old custom content management system to meet new web standards in order to look pretty on new mobile devices – using WordPress made this simple, helps to simplify updates on the go even further, and means I have more time to spend on projects that I enjoy.

I am currently digging through notes and photos from the many projects I have neglected posting about – look for some updates here soon.

Curtain Clock

I’ve never been a morning person. But, I’ve found the best way to wake up is with natural light. These indoor solar simulators hold no candle to the bright sunlight on a clear morning. My solution was an alarm clock to draw back the curtains.

This device mounts to the wall beside the window, neatly hidden behind the curtain edge when open or closed. The alarm from a clock assembly (pulled from an old programmable coffee machine) triggers a 555 timer acting as a monostable oscillator which drivers a relay-controlled synchronous AC gear motor that draws the curtains closed via a cord. The 555 pulse length must be set as to generate the correct fixed opening time. It is also a good idea to have a torque-limiting clutch (ie, find a bobin that fits tight to the motor shaft that will slip if there is a jam).

I built this at the end of 2008, but never really documented it. It works well, but I would like to redesign it to automatically find the full open position.

Increase Laser Pickup Gain on Chrysler Infinity CD deck

The CD player in my Eagle Vision was slow to change tracks, didn’t seek to outer tracks, or handle my heavily scratched cd collection well… And, this finally led to some minor embarrassment – enough to drive me to fix it.

This CD player design was featured in basically all Chrysler vehicles through the 90’s (plus a earlier and later). The dash surround, at least in my Eagle, is attached with only speed clips – it pops right out when you gentley pry at the edges. Be careful detaching the wires since the harness doesn’t let you pull the deck out very far. This deck is grounded at one point only – through the strap bolted onto the back. Keep this in mind if you test with the cover off.

Now, if you’re smart, put a CD in, and seek to the last track before ejecting the CD and removing the deck – this will avoid most of the tedious manual seeking by finger tip. Remove the screws holding all the metal covers, including the heatsink side and ground bolt plate. Pull the top logic/RF board from the deck, exposing screws through the bottom board – remove these so that the bottom cover can be partially removed, giving access to the laser assembly. You will have to rotate the gear driving the slide the laser is mounted on in order to bring the laser close to the outside edge, exposing the gain potentiometer (see image below). Turn this pot clockwise to increase the laser gain – too much gain will burn out the laser, ruining your deck! You have been warned! I only had to turn the pot on mine a hair (about half a degree) to get the gain I needed – Unfortunately I couldn’t find solder points easily accessible with an ohm meter to measure the change the adjustment made.

Reciprocating Saw Repair

While cutting out sheet metal to repair a portion of the rusted floor on our old Dodge Caravan, I ran my cordless reciprocating saw at a bit too high of a duty cycle. The plastic in the drive motor melted. Luckly, everything else in the saw (gear box and switch?) were well built, high quality parts. The motor was standard, and a trip to Princess Auto found an easy replacement.

CA1010 Backlight Repair

When I got my Yamaha CA-1010 Amplifier, it worked quite well with the exception that most of the backlight did not work. I promptly tore it down, checked the power consumption of a working bulb, and replaced the bulbs with LEDs with current limiting resistors. Also, while inside, I sprayed the controls with contact cleaner (isopropanol based) and followed with silicon lube to stave off oxidization.

From 2008 to 2010, I worked in a lab where my primary project was the upgrade of a helium dilution refrigerator so that RF measurements could be taken on a experimental package. This system, previously used for helium experimentation, was massive. Huge pumping speed with 6 or 8″ lines, double sealed mechanical pump, roots blower. This is no small dilution fridge!

Key Analysis Progress

fteeth Having worked on digital filtering using MATLAB, I’m fairly comfortable with it. I had not used the Image Processing toolbox – but, it is easily learned.

The images shown here steps along the process of extracting the teeth from the image for manipulation. I will fill in the theory here soon.

81 Yamaha XS400 – Stripping the Tank Paint

The weather warmed up a bit today, was about -3°C, and I was able to dig out the garage door, and spend some time getting my gas tank ready for painting. I’m not an autobody guy, but, this still shouldn’t be too hard. I first removed the fuel petcock and the gas filler cap from the tank – I blew air through the tank with a vacuum on reverse to remove vapours.

Started off with an abrasive wheel mated with my drill – worked really well to remove the soft top layer of non-factory paint (and worked well on the factory paint). I decided to go down to primer level, and start with a couple new coats of primer and go up. Unlike a wire wheel, this abrasive wheel didn’t seem to heat the metal – good, because that heat could lead to deformations.

Started sanding with 200 grit after that, and will work up from there. I have to do the finer stuff wet, so, that’s left for another day (and the bondo will come first)

Key Analysis

sectioned After looking at the number of keys I had on my keyring for the same building, and making some quick comparisons, I started thinking about the possibility of generating a master key from the quite freely distributed zone keys, for academic reasons alone. This would be an interesting lesson in image processing.

The General Idea:

Scan each key in colour/grayscale
Automatically convert image into a 1-bit profile of the key
Automatically Rotate image so that all keys are in the same orientation
Automatically section the key to isolate the teeth
Find the central location on each tooth from key sets (all local max/min)
Define Integer values for all tooth-height possibilities
Define each key as a set of these integer values
Compare similarities and differences
Option to precategorize with key zone

CD/DVD Storage Jukebox

The Loading Lever

I designed a replacement loading lever using my academic copy of Solid Edge v11. It almost mirrors the original, except, because I no longer use the upper lever also, there can be a bit more material. This will be made out of aluminum to improve rigidity also.

The data CD/DVD jukebox
I was always fascinated with those large cd changers when they were new; but, they were always expensive. Plus, with no more than 50 to 100 audio CDs, they aren’t that useful to me – it’s much easier to carry around compressed copies. I had seen SCSI 4×4 disc changes in standard 5.25″ form-factor, and had seen tape drive autoloaders as well. I ended up getting a line on a non-working 101 audio disc changer. I could now build a prototype.

When I first envisioned this project, hard disk storage was still relatively expensive; I had 20Gb and 30Gb drives, the later had just been purchase at a cost of over $300CAD. Compare this to buying good quality CD-R discs in bulk at $0.15/ea. Hard drive storage was running ~$10/GB, while CD-R storage was ~$0.22/GB – not accounting for the cost of a CDR drive, or considering the relative frailty of the media.

The obvious main difference between the storage schemes boiled down to access speed, latency, write speed, media handling, and the amount of data you can have online at a time. I didn’t want to have to handle CD-R discs, risking damage, and, I didn’t want to have to swap them in or out often. While filesystems such as mediadatabase or hierachical storage file system exist, and could be used (or modified) to run an active filesystem on such a device, the extreme latency, low access speeds, and small capacity of single discs make this a less than perfect option. However, a device such as this is well suited as an automated backup system – the 101 disc change would hold 70.7GB on CDR before compression. The option to simply upgrade the unit to a DVD-R (or dual layer) drive when cost effective would further expand the backup capacity to 474GB (or 949GB)! Needless to say, years after the initial idea, I was still thrilled to redesign the disc changer robotics to accommodate a highspeed data drive and computer control.

For the prototype, I used the CDRom drive from an old-style slot loading iMac – slimline drives were much easier to mount to the jukebox. The only change to the drive was an external connection to the “disc inserted” sensor, allowing loading to be triggered at will — using an external sensor, this allows loading discs from wider aproach angles. This can be triggered through the jukebox’s own loading mechanism when it goes to load a disc.

At this point, I have temporarily stopped work on the control system, after dropping the new drive mechanism, breaking a critical part. I will machine (or scrounge) a replacement eventually.