18 Dec 2012 » Inductive charging on a Droid Bionic
Update: I tried another touchstone, and my phone is charging without any apparent issues! Next, I’m going to litter my house with touchstones, and maybe embed one in my keyboard at work.
My Droid Bionic is pretty battery hungry, so I keep it plugged in most of the day while I’m at work. I’m tired of dealing with wires on my desk and forgetting to plug them in, so I decided to try inductive charging. Motorola makes an inductive charging system for the Bionic, but that would run me around $100, and I hack things.
Inspired by [Ryan Rix] and Will Stone’s [article], I thought I’d try to modify my phone to work with a Palm Touchstone charger. [Ryan Rix]: https://twitter.com/#!/rrrrrrrix [article]: http://uselesspuzzles.blogspot.com/2011/09/droid-bionic-optimizations-extended-bat.html
I basically needed to connect the receiving coil’s contacts with the inductive charging port built into my phone. To do this, I routed some copper tape to a partially tape-covered piece of foam.
One problem I ran into while doing this is that there is no reliable pinout information available for the four-pin connector on the Bionic. Looking at the last picture in Will’s post, I wrongly assumed that the top pin was +V. One of the comments mentions that the picture might be wrong, and I’m confirming that: The top pin is GND, and the bottom pin is +V. I still have no idea what the center pins do.
For a few minutes after first connecting it, I was afraid I had fried my phone because the mobile data silently turned off.After fixing the polarity and re-enabling data, I placed my phone on the charger and… success! … kind of.
Now, when on the base, my phone toggles rapidly between ‘charging’ and ‘disconnected’. I’ve done some research, and this appears to be a common issue with these mods, as well as stock Pre/Pixi devices. My multimeter and oscilloscope show a fairly consistent 5-5.6V coming from the receiving coils, so I’m not sure what the problem is or how to fix it.
My next step is to try another base. I’ll update this post when/if I find a solution and/or more information. In the meantime, here are a few pics of my mod:
11 Nov 2012 » Parametric Nosecone
About ten years ago, my sister and I were launching model rockets that we built together. She built a Quest Courier. After some initial trouble with getting the nosecone’s parachite to deploy, I got it to work.
The last I saw, it was floating over the freeway about three miles away
She’s still annoyed with me.
could go buy a new kit, or write Quest asking for a replacement, but that’s not how makers work. Instead, I decided to use my hackerspace’s 3D printer. HeatSync Labs recently acquired an Ultimaker, and I recently acquired some modeling skill. Using OpenSCAD, I designed and printed a replacement nosecone for her rocket.
#### measure To make the nosecone fit and behave like the original, the nosecone had to meet a few specifications: - connector section must fit tightly in body tube - payload capsule must snugly hold a chicken egg. - outside diameter must not exceed launch rod standoff height
#### design My current preferred 3D design system is OpenSCAD, so that’s what I used to design this nosecone. I ended up making it modularly by generating several segments of hollow cylinders. Rather than create each segment individually, I built a parametric tube function. It takes a length and two radii, and generates a tube with those dimensions, tapering as necessary.
- The tip kept melting. Playing with the extrusion head travel rate, I was able to reduce this effect. I still had blobs, which were easy to sand down, and some pits, which I filled with cyanoacrylate.
- The connection at the shoulder was weak. The rocket tipped over, and the bulk of the nosecone just snapped off. Reprinting the lower piece with a thicker shoulder seems much stronger.
#### paint I applied two coats of spraypaint to protect the outside and make it look more like the original. When painting the new base, I didn’t take notice of which black paint I grabbed, so the top is glossy, and the base is shiny. Oops.
### lessons learned #### Measure carefully and plan ahead. After designing and printing the cone, it occurred to me that I might have made it wide enough to obstruct the launch rod. Thankfully, it ended up being just small enough to work. #### Make sure equipment is working The printer had a broken Bowden tube mount, so the filament easily lost tension, fouling the print. I discovered the broken piece after the third attempt failed, wasting several hours. After replacing the connector and adjusting the feed mechanism, the prints were much more consistent #### Pay attention Once the prints were completing, I ran into heat trouble. When reaching the tip of the cone, the plastic wasn’t getting enough time to cool, causing blobs and other unevenness. Increasing head speed only made the problem worse. After noticing that, I slowed it about an inch from the tip, allowing a clean, sharp finish.
This project really helped improve my understanding of the 3D printing process, and my ability to use OpenSCAD. I’m really pleased with how it finally came out, and my sister is excited to be able to fly her rocket again.
26 Apr 2012 » Switching EL Wire
My motorcycle has an been inspiration for many project ideas. The most recent of these to come to life is electroluminescent wire under-lighting tied into my turn signals.
EL wire needs (relatively) high voltage AC to operate – typically around 100V at 2kHz. This provided me with a few design challenges: 1. Convert 12VDC to 100VAC 2. Switch voltage to multiple EL strands 3. Avoid leaking AC into the bike’s DC systems 4. Do the above using the least space and weight necessary
Challenge 1 was fairly easy. Originally, I had intended to design and build an inverter. After size and effort considerations, I settled, instead, on purchasing a purpose-built inverter from adafruit.
The remaining challenges proved to be more interesting. I have been switching DC for years, mostly with jellybean transistors and microcontrollers. I’ve done some AC switching with small mechanical and solid-state relays, but don’t want the noise or space typically involved when using relays. Enter the opto-isolator and triac.
Two quick electronics lessons: >A triac, or triode for alternating current, is essentially a transistor that is capable of passing current in both directions.
An opto-isolator uses an LED to turn on a phototransistor, allowing two circuits to be connected, but electrically separate. This allows a low-voltage/current device, like a microcontroller, to switch higher voltages without exposing it to potentially hazardous levels. I used an opto-isolated triac, as the larger triac requires an AC signal to switch on.
Using KiCAD, I designed a simple board consisting of an opto-isolator and triac for each of two channels.
I used PCBNew, which is packaged with KiCAD, to convert the schematic into a board layout, which I could then use to make a usable circuit board.
Lesson learned: Verify pin identifiers during and after layout, before printing/etching. KiCAD had the triac pins numbered incorrectly. As I had turned off the silkscreen layer during layout, this went unnoticed, leading to frustrating hours of hardware debugging.
Once I had the board(s) laid out, it was time to actually make the thing. At HeatSync Labs, we have a few methods available for creating PCBs. My favorite so far is using the laser cutter. Unfortunately, the laser is unable to cut or etch copper, so we use black spray paint as a mask.
First, the copper is covered with a layer of black spray paint. Once dry, it’s placed on the laser cutter’s bed, and the design file is imported to the control software. Since the software isn’t very well-written, importing requires several file conversions. I have found the best path to be: 1. Export from KiCAD as SVG 2. Import SVG into CorelDraw, export as bitmap (b/w, 600dpi, ensure scaling is 1:1) 3. Import BMP into LaserCut 4. Invert image
This creates a negative mask, i.e., white portions will still have copper when finished
This is critical for back-side/bottom-layer copper!
- Verify all dimensions are what you expect
- Follow laser procedures to “engrave” board
Etching a masked PCB has been covered many other places, so I won’t go into that here.
Once etched and cleaned, I drilled holes for the components. I did this with a Dremel drill press and a tiny high-speed drill bit. Be careful when you’re doing this; Most of the drill press accessories I’ve seen are very wobbly, and that doesn’t work well with these bits. They’ll shatter and fly everywhere at the slightest lateral pressure. I tightened all screws/bolts, then added extra tension on the drill mount by tying it to the frame with some Velcro strips.
After drilling, components were soldered into place.
The boards were tested using a typical bench power supply. Third try was the charm!
As excited as I am to have this set up on my bike, I haven’t installed it yet for two reasons: End-of-semester crunch, and I don’t like the color or intensity of the wire. I purchased high-intensity yellow wire, but it turned out to be more of a sorta-bright greenish. I’m looking for a better source now, but might install this anyway in the meantime.
When it is installed, the board will be housed in a weather-resistant enclosure mounted to the front of the bike’s frame, since that’s close to where I need to pull power and signal status from. The enclosure I have should be large enough to also house my future headlight modulator.
I am currently searching for EL wire that is better suited to this purpose and matches my bike better. At some point in the near future, I’ll be expanding the system to accommodate a third channel for a brake indicator, and adding an extension to light my jacket. Stay tuned.