Two big anouncements:
- The roadtest challenge has been extended one week, so the deadline is now friday the 19th. I'm still planning to finish my project this week, otherwise my girlfriend and son will go crazy because I'm spending too much time in my 'lab'.
- Hooray! This morning, using a flashlight shining close over the floor, I found the LTC2935 back that I lost yesterday evening! Now my project can continue!
I finally got my board working, at least harvesting and going to a stable 3.3V! Yesterday I wrote about the cunning comparator circuit used on the demonstration board. Today I tried to replicate it from the schematic of the demonstration board, but the output would not turn on my FET... In despair (had I ruined the comparator?) I verified the schematic with the actual demo board and the error was clear:
The documentation did not match the populated components on the board!
In the schematic that was documented, the voltage level to trip was set on 3.15V. You'd expect the output to go high at 3.15V, then. But reading more carefully, this threshold is only for a falling input on the comparator! The rising voltage should exceed this voltage by 5%, and thus the output trips at 3.15 * 1.05 = 3.3075V. Which is not how high the regulator will get, so the output never trips.
Luckily for the engineers of the demonstrator board, they could populate some other positions to shift this level to 3.0*1.05 = 3.15V!
Please, Linear or Wuerth, update the docs! [edit: Lars Fahrbach forwarded this internally, so it'll be taken care of!]
After this fix, my board gives a nice and shiny 3.3V to the micro! YEAAAHHH!!! I can continue my project!
- Read the documentation: AN0062 of EnergyMicro already warned for draining the storage capacitors
- Do not always trust the documentation, verify with the reality!
- Debugging makes me grumpy; finding the solution makes me happy!