27
« on: April 11, 2013, 09:35:54 PM »
I think what is happening with that transistor is that when VIN is grounded or left floating (in which case R3 pulls it low), Q1 is closed circuit (FETs are normally closed, the opposite of BJTs). Since USBVCC is connected to the USB connector, then the +5V net will effectively be connected to +5 power supplied by USB (if plugged in to a USB power source). When VIN goes high, the FET is open circuit and the +5V rail is no longer connected to USBVCC source. This would agree with the notation in the ISP Shield spec that says "Will supply +5V to the system if VIN is not present". It looks like nothing is normally connected to VIN (pin 10 on the shield connectors), so unless you go out of your way to apply a voltage to that pin, +5V will always be live (if connected to a USB host or hub).
I think maybe the purpose of the VIN signal is if you wanted to provide your own +5v power source, and keep it isolated from the USB link. Like maybe if you needed to drive more amperage than you want to pull from the USB source (i.e. anything over 500mV). In that case, you could just wire your power source to the +5V net (pin 8 on the shield connector) and pull VIN high (pin 10 on the connector). Or if you needed a non-standard voltage such as 4v.
Then there is a power module on the Processor unit that appears to select either VBATT or +5V as the VCC source to power the microcontroller (also used as analog and I/O reference voltage). What it looks like is happening in that module is that VBATT is used for VCC unless +5V is there. If +5V is present, then the top half of Q1 (not the same as "Q1" in the ICP board!) is suppressed, which closes off VBATT from VCC, leaving +5V attached to VCC.
VBATT is of course 3v since it is a coin cell, so it seems like the system can end up running on either 3v or 5v, depending upon which power source(s) are hooked up. For the MCU that's OK, but it also means that peripherals would also need to be able to handle both +3 and +5 volts I think? That seems wrong - wouldn't the voltage jumps potentially cause unexpected behaviors... such as overdriving LEDs (or worse)?
Another thing that looks odd is that we have a +3v3 pin on the connector (pin 6), but I don't see where this is used or supplied anywhere? Pin 16 of the FT232R chip has 3v3 output via LDO regulator, but that can only drive up to 50mA, and it doesn't show as being connected to that net anyways. So that seems a little bit weird.