proto
is enough
It can also regulate the fan speed if one day we need to attach a fan to the heatsink.
OK looks good
R16 was added because we add a new method of power cycling which might or might not work. Once it is validated/fixed (if needed) R16 should be removed for the series production of v3.
OK, looks good
I don't have strong opinion, ports/net labels are most important.
yes, we can
yes, please change the capacitor to 100nF as per datasheet recommendation
No, it's OK, make it mounted. We will test this once the prototypes arrive, worst case we can remove D4.
OK, so we'll need to add a correction factor in the software ( @milettri ). But on top of that, we need to check the PCB routing and modify the capacitor value for v3 ( @Filip.Switakowski )
Well, to start, I would expect both of them to report a very similar value, since MAX6639 inputs temperature sensing DXP/N from the SoC. Unless there is some temperature effect on the MAX6639 itself as it is placed close to IRPS circuits?
The difference is about 10degrees (MAX6639 measurement is higher)
IC35 is powered from USB port, so if P3V3 is not there, that I2C coming from the FTDI will be connected through IC35 to I2C-eeprom, which then goes to IC8 (powered from P3V3_USB or P3V3). EN
of IC35 cannot be connected to PS MIO in this case, because SoC is not powered if System Board is not plugged in the crate.
Then if P3V3 is there (i.e. board plugged to the crate) IC35 is always disabled, even if powered.
This came from @abyszuk as EPC requirement to be able to read the MAC address of a System Board before it's deployed in the crate. Programmable DC/DCs are configured at the production stage.
Ah indeed, in v2 this reset signal was coming from a 1.8V PL pin. Now it's coming from a 3.3V PS pin so indeed the inverter can be removed. However, we should leave the R26 pull-up as it is.
Already corrected in EDMS
Asked CERN design office to fix this in EDMS production files.