Designing the PCB, step #1: schematics and footprints

Hello All,

After my previous post, about having worked to get something finished before 12th of June, where it should have been July, I've started designing my schematic and PCB. Somehow, this process always takes more effort than I see upfront. I had to make parts (footprints and schematic symbols) in Eagle for the LTC3108, Wuerth transformer and the EFM32GG940, and of course determine which parts should surround them. I've considered taking a smaller variant of the EnergyMicro controller (I guess even the TinyGecko would be able to handle my application), but to play a bit safe I chose to use a part from the GiantGecko family, one that has a QFP housing. I know it's more difficult to solder than the TQFP, but I like the looks of QFN housings...

Designing your own board brings the necessity and pleasure of diving deeper into the features of the components you're using. I'll share some insights below.

LTC3108 design

I was very happy with the schematics and BOM of the Energy Harvesting Solution To Go kit on Wuerth's website. I could basically copy the schematic, and remove some optional features:

• The output voltage can be set to 4 fixed values: 2.35, 3.3, 4.1 and 5V. Because I'd like to use as little energy as possible, but don't want to go as low as 2.35V (makes component selection for other components quite hard) I'm using the 3.3V output
• The LTC3108 has a second output, which I am not going to use (at least not as far as I know now). It's quite nice to have though; it's current limited, and will help when a 'high power' (in energy harvesting terms) load has to be switched on and off. I'm still doubting whether I'll use it to switch a buzzer on or off.
• Another cool feature is the possibility to add a 'storage' capacitor or cell.This is something I had in mind to design myself, but apparently Linear already put it in this chip. When you're energy harvesting, and you're using very little power, but your power source is still supplying it would be ultracool to have some place to store this excess energy for later usage (when you're driving a solenoid, or use a lot of high-power peripherals). When you're using a battery or cap at the 'Vstore' pin, excess energy is stored. Here you'll have to make a tough choice: having a lot of storage is cool, but charging a large capacitor takes a lot of time. Choosing a smaller capacitor means you'll be able to use the extra energy available quite soon after you've started harvesting, but you can only use a little capacity because.. you chose a smaller capacitor. Again, designing for energy harvesting means that you'll have to know a lot about both the energy source, and about how the energy is used....
• The switching design can be made with several options of transformers and capacitors. I'll first draw the design, then simulate to see what values are needed while the board is being produced.
• The design on the demo board uses a Sanyo Polymer capacitor of 220uF. I can't exactly figure out why these capacitors were chosen, unfortunately Farnell doesn't have them on stock. I guess it's the combination of high capacity (not available in one device for ceramic capacitors) and low ESR (not available in one device for aluminium electrolyte capacitors). I chose to place several ceramics in parallel, but maybe someone can point me out what the specific design characteristics are that got the Sanyo's in the design...

EFM32GG940

Haven't been able to do much here, but this is on my todo list:

• Find out how to connect the debug interface
• I'll take a bit of risk and connect an LCD to the LCD interface. The datasheet on this topic is quite beefy, so I hope I'll do this right.....
• Have to find out what pins to connect, and which pins not to connect

I'll post more details later on, combined with the Eagle files...

Greetings,

Victor