06-16-2018, 04:34 AM
(06-16-2018, 03:35 AM)lucho Wrote: Thanks for the information! Unfortunately, I can't find the datasheet of the TCS4484B voltage converter, but from the schematic, one can see that this is the resistor was a part of a voltage divider to reduce the level of the PCIE_PWR signal on the "enable" line (EN) of the TCS4484B in half. (Probably this EN line is connected internally to a base of an NPN transistor whose emitter is grounded.) I don't know the voltage level of this signal when it's in HIGH State, but on sheet 15 there is a note: "RK3399 part E is 1.8V/3.0V mode". Perhaps at first it was assumed (or determined) that the HIGH State level of this signal would be 3.3 V, but then it turned out to be just 1.5 (or 1.8) V, which, divided in half by the aforementioned voltage divider, turned out to be insufficient to enable the TCS4484B, so it become necessary to get rid of the voltage divider by removing R895381 and connecting the PCIE_PWR to the EN pin of the TCS4484B only via R223. Is this explanation of mine close to the truth?
Yup, that makes sense.
If the 'enable' signal is a positive signal, it has to be a voltage divider. It can only be that or a pull-down resistor to ensure a known state before the SoC GPIOs are configured, which is pulling down too hard. That fact that it looks like it may be a NC part in the future make it look more like a voltage divider that's not necessary. Hopefully the sata problem is just driver related now, as the SATA card does appear on the PCI bus so it is being powered and is talking to the rockpro64... or at least the controller is to some extent.