Fun video with demos:
Train at 1:05 (I got a bit carried away), Qihuahua at 2:23
Hello! The last months I haven’t posted anything on this blog, which is a bloody shame. However, I’ll compensate with this blog post. A few months ago a fellow Tkkrlab member told me about his project to let a remote controlled LEGO train run via a web interface. He had this up and running, but got tired of constantly changing the batteries and started looking at wireless charging systems. Around the same time, element14 started a RoadTest with a combined Wuerth / TI demo kit for their new industry standard Qi wireless power transfer solutions. I applied for the RoadTest, and was happy to be selected as a candidate to try out this new hardware. But…. then element14 changed the rules, and instead of a regular ‘RoadTest’ (use the hardware & write a review) this became a Roadtest Challenge; a design competition between 6 candidates. You can find the element14 blog posts here, but I’d like to make a personal summary on this page. Below the progress of two months designing and hacking!
Short technology roundup
Qi (pronounced “chee”, like in Tai Chi) is the industry standard set and maintained by the Wireless Power Consortium. Both manufacturers of end user stuff (LG, Nokia, Blackberry) as component manufacturers (Wuerth, TI, Panasonic) have taken place. You’ll be able to charge your Qi-compliant Blackberry on a Qi compliant Nokia charging pad and vice versa. Very good, in my opinion. The technology used is inductive charging, and the evaluation kits in this roadtest were bq51013aevm-765 and bq500211evm-045, both using Wuerth coils. This setup is able to transfer 5W (5V/1A) without physical contact, at an efficiency up to 90%.
Of course I wanted to get the train running, and charging on Qi. But the guidelines of the challenge asked for something that would integrate with your daily life, so I took a bit wider vision and took to charging toys with AA(A) batteries in general, as that is a common problem in my household. For the Open Power Functions I already had a battery box of LEGO (6AA), and I thought it would be a nice challenge to convert that to Qi as well. This all applies to the ‘receiver side’. I thought it would be nice to have the transmitting side be ‘fun’ for children. My girlfriend took on her the job of creating the ‘Charging Qihuahua’, a charger in pet form. Wuhahahaha!
All dates below link to the original blog post at element14, with larger pictures!
I unboxed the evaluation modules, played around with them and decided I’d take a go at at least making a Qi compliant LEGO Power Functions battery box. Making something for the train looked hopeful.
OCTOBER 17th I did the first testing on the LEGO train, to see if the LEGO train idea would be feasible at all; the distance between transmitting and receiving coil had to be small enough to start transmission. Placing the transmitting coil under the rails made placement of the train very finnicky. After the connection was made, it was reliable, and I got good efficiency over the power path.
NOVEMBER 2nd With one month to go, I had been working very hard on the schematic and layout of my own receiver board. The biggest challenge was to find a working solution to charge 6 AA’s from 5V/1A output. In the end I used a boost converter as current source (TPS61165, actually meant as LED driver) and a NiMH charger management IC that terminates charging on change of temperature (bq2002T).
I made a layout and sent it to OSH park.
Also, I adapted a piece of LEGO rails to accomodate the transmit coil (dremeled away some support sections, glued in coil):
NOVEMBER 10th Movement test: the Qi system uses some communication to enable the transmitter at full power; this takes some time, and therefore the output of the receiver isn’t instantly on when a train passes the transmit coil. This needs overthinking. As can be seen below: NOVEMBER 18th I received the evaluation kit for the boost converter. It was actually meant to drive white LEDs instead of charging batteries, and driving white LEDs it does:
This enabled me to try out some protection features I had in mind (preventing the output from blowing up the charge manager). To make room in the LEGO battery box I changed to AAAs instead of AA cells. This gives enough room for a PCB on one side, and a coil on the other. To let the AAA cells make contact (they’re shorter) I used some cardboard and clipped pushpins:
The other thing I started doing was modifying one of my Open Power Functions Receivers to accomodate cell voltage sensing; had to do some solder-fu here:
NOVEMBER 20TH I got my PCB’s from iTEAD! Yes, iTead instead of OSH Park. I hadn’t taken into account the time & cost it would take to ship from the States to Europe, so I ordered PCB’s in China and had them shipped by DHL. Due to a part number error I couldn’t test the Qi receiver yet, but I started working on the charger.
NOVEMBER 25th I got the QFN housed receiver IC’s and soldered them. I recently bought an Atten 858D+, and I’m very happy with it. This video shows soldering the IC: NOVEMBER 26TH First proto is working! YEEEEAAAHHHHHH!!!! 4 days to go!
DECEMBER 1ST To my dismay, the deadline has been postponed; I worked so hard to get it done in time, that adding another week seems to lengthen the project, not relax it. But: I’ve got results, lots of them: LEGO POWER BRICK
I’m really proud of this one: the outside of the brick is completely unchanged, yet the inside has a completely vamped functionality; the box has two lids: behind one is the coil, behind the other is the electronics. QI ROBOT TOY Charging 4 AA’s is also possible by changing one resistor in the receiver circuit!
Qi train The ‘piece the resistance’ is the LEGO train; this took a lot of time in programming and testing, we had to add a Hall sensor to the receiver, and a small magnet to the transmitter rails to detect the charging position; but now, it works! When the battery voltage gets too low (should be replace by some sort of fuel gauging in the future) the train ignores all infrared commands that are normally used to steer it, and starts looking for the charging station. When found, it stops and starts a charging cycle! WOW!
DECEMBER 4TH I added videos,both on the top of this post DECEMBER 10TH I still don’t know who won; but I wanted to investigate an effect I saw when charging the robot; I finally had the current that I wanted to have. The protection circuit I used actually cut off the available current to the AAs before reaching the setpoint I wanted. As side effect, it kept the connection to the Qi alive by not overloading the circuit. Interesting stuff. I did some simulations. Click on date above for more details.
DECEMBER 12TH Still no decision in the challenge. I made a comparison between OSH park and iTead:
DECEMBER 12TH How small can you go? I soldered one of the OSH park boards, and cut off half of the board. This makes a very small Qi circuit, in this case powering an LED:
Dang it; I didn’t win. Chad Johnson made a RasPI with Qi, and powerpath routing that convinced the jury more than my design.
I finished the train design AND a LEGO box AND the robot. The design files for Eagle can be found here(zip), when you’re interested in the LEGO train please drop me a line in the comments and I can give you more info on code and implementation.
Qi technology is really cool, it’s efficient and easy to implement. Whether I’m going to do a next RoadTest Challenge? Not sure yet, as it took a lot of time….