O.K., so the menu system in WordPress just started crashing after trying to add a page, giving me a “The given object id is not of that menu item” error many times over.
Argh, so simple things such as reloading the old menu (using the back button on the browser), or disabling all plugins has not fixed this. One source claims that it could be a problem with the maximum php memory limit being set to low (something the server owners have to change).
Worst case scenario is I have to recreate the entire menu structure again. My website was basically entirely linked from within this menu. Groan…
The waterproof LED lights that I had ordered from Deal Extreme months ago finally arrived. I already had one of these brought a long time back for playing around with, but needed another two for the skateboard. The plan is to mount two facing the front, and cover the remaining one in a red filter of some sort and make it rear-facing. They run directly of 12V, and are easily mounted.
The I2C comms page is now pretty comprehensive with added information about long-distance I2C communication, I2C variants, USB-to-I2C drivers, screenshots of I2C signals, and much more. You can visit it by clicking the link below:
The Electric Skateboards arrived today from PCB Cart. The first thing I noticed is that the yellow soldermask I asked for was not yellow at all, but rather what can only be described as a dark spew orange. I was kind of disappointed about this, since I was hoping for a bright yellow that would match the colour of the pacman on the PCB. Also was the problem that the text on the remote PCB did not stand-out well against the dark orange, and it would be hard to read the user interface. I sent PCB Cart an email with a photo showing the colour, and they replied saying that it was the colour that they always used for yellow, and no one had complained about it before.
We reached an agreement that I could get them redone in another colour, paying have the manufacturing costs (about NZ$40), without paying for set-up costs (presumably because they still had the stencils), and with free shipping. I thought this was a fair deal and resubmitted the design to be made with a red soldermask and white silkscreen.
The photo below shows the gross orange colour compared to the kind of yellow I was expecting.
Finally, the electric skateboard PCB’s have been sent away for manufacturing. I eventually chose PCB Cart after discovering that I couldn’t run the PCB’s as a ‘Prototype’ at Silver Circuits (their maximum Prototype PCB size was 160mm in any direction, my panel was 200mm). See the Electrical Suppliers page for more info on PCB manufacturers.
The boards ended up costing NZ$330 (ouch!). This was for 3 panels of 3 boards (adding two panels only added NZ$14 to the price, so it’s worth it for back-up boards).
There is the the motherboard PCB, the half-bridge PCB (both to be mounted on the skateboard), and the remote control PCB (mounted on the hand-held controller). I’m crossing my fingers and hoping there are no mistakes (well, at least no major ones that would require a redesign). One thing that I will do differently next time I design a PCB is to make sure the minimum annular ring is 0.3mm or larger. Since I had 0.2mm vias with 0.4mm holes, I had to raise the annular ring spec when submitting it to PCB cart (from 0.3mm to 0.1mm), which added about NZ$40 to the price.
I chose the longest lead-time (12 days) to keep costs down, hopefully they will ready by the weekend after next!
The motion sensor and alarm is now working! This would of had to been the easiest part of the circuits to get working. Everything just worked, first time (so did the software, and big part due to the PSoC’s easy to use drag’n'drop hardware blocks and configurations tools.
I found this cool little motion sensor, the MS24 on Element14 for less than NZ$10. It’s heaps better than most tilt sensors since it is sensitive to motion in all orientations, unlike tilt sensors which tend to only change state at a specific angle from the horizontal.
The sensor is normally closed, but when a small amount of motion disturbs it, it momentarily goes into the open state. All you need is a resistor and capacitor to interface this into a microcontrollers interrupt pin. The combination of the resistor and capacitor control the sensitivity.
I made a simple BJT full-bridge to control a speaker (or peizo) at 12V from the microcontrollers PWM outputs (the PSoC allows you to configure one PWM module to have two outputs, each the inverse of the other, all in the GUI!).
For the prototype, the motion sensor triggers an interrupt on the micro, which then enables the PWM, changing the frequency every 500ms to create an alarm sound.
Here’s a little video showing how it works. Note that I had the motion sensor on maximum sensitivity (aka no capacitor), hence how a little table vibration sets it off, which is quite impressive!
The subscription for this sites default domain name, www.undergroundmadlabs.com, was coming to an end. I thought this would be a good time to change the name to reflect the site title. Choosing undergroundmadlabs.com was a bad decision on my part, made at a time when I wasn’t too sure on a good name. Soon after I brought it for a year, I realised Clandestine Laboratories was better, and rebranded the site. However, I was stuck with the domain name (well, unless I wanted to pay double), until the subscription ran out.
So now, after a week of hassle to get cladlab.com transferred to justhost.com, and then moving the WordPress installation from one account to another, it’s finally working! The site was out for a few days while I struggled to get the WordPress installation working happily on the new server. I couldn’t of done this without interconnectIT’s tutorial “Migrating A WordPress/WPMU/Buddypress Website” and the php script “WordPress (and others) Safe Search And Replace Tool“, again by interconnectIT.
I finished building and started testing half-bridge mach 4 this weekend. To my surprise, it’s working much better than the previous 3! For one, no MOSFET’s blew up within the first few seconds of connecting it up to the motor. I have a feeling it’s to do with the higher Vds (100V) of these MOSFET’s over the previous ones (75V). This gives the drain-source TVS diodes plenty of play-room to kick in and absorb transients before the MOSFET gets damaged.
The schematic the prototype board was made from.
Tentatively testing the half-bridge with a lightbulb before hooking it up to a motor.
Testing it with a motor. I was able to stand on it for short times! (last time the MOSFET’s blew).
However, the MOSFET’s and the bulk capacitor were getting unusually hot. I hooked a scope up to the gates of the MOSFET’s and noticed ringing on the low-side FET when the high-side FET switched on. I think this is causing momentary shoot-through. This would explain hot FET’s, and, if the voltage rail was sagging, hot bulk capacitors.
