BQ25798: Seamless transitions between ACFET-RBFET pairs

Hi Jeff,
Thank you very much for confirming the VAC1/2 switchover time.

For using the TPS2121 as a solution, and the initial power requirements, I'm sorry I omitted explaining our small low power SoC (embedded ARM Cortex-M4 and 192KiB Flash) that handles our product's low-level system & power housekeeping.  Upon attaching a power source, the low power SoC measures each input source condition & voltage levels for validity and controls our main FPGA/SoC high-current buck regulators, which will be fed by BQ SYS.

We plan for our low power SoC to be the EC host for the TPS25751D target and issue the patch file,  presently exporting at 13,184 Bytes for the low region binary, which we have adequate space in its Flash.

Our SoC requires a couple mA peak.  The 3 power sources are simple schottky ORed into its low current 3.3V regulator (which can tolerate 40V input over-voltage), from the TPS25751D PPHV (5V), Ext DCin, or battery.

We plan to ground the BQ TS pin, preventing default charge.  Our low-power SoC checks the battery pack comm data for non-pirate attachment and its temperature data.  We plan for it to monitor the Bat temp data, write 4CC commands to enable/disable BQ charging, and pet the 160-second watchdog.

Therefore, do you think the TPS2121 will be ok, since our initial power requirement is a low power SoC, which issues the patch file, and measures if voltage levels are adequate to allow the high current FPGA/SoC to power-up and boot?

Thanks again for your advice!

Norbert

Hi Jeff,
Thanks.  Sorry I am confused, or maybe I was also confused in your first reply?

I think your first reply covered 2 issues:
---
Issue 1: "switching over has a brief time where both are off... likely impractical capacitor could handle the load"

I think you were confirming that the VAC1 to VAC2 switch delay could risk a reset (when no battery present), during our example high-power switchover scenario (17W), when the the Ext DCin Jack is pulled (VAC1 priority source), attempting to transfer over to a good USB-C 15V,3A VAC2 source that had been attached for a while.
---
Issue 2:  "the USB-C source starts at 5V*2A=10W so PD controller must increase the input power to meet the 17W-20W demand."

I thought you may have been worried about our initial power-on with USB-C, if our power was 17W.  But we have a low-power 3.3V SoC to handle the EC host for the TPS25751D before the high current FPGA/SoC boot.

Do you think the TPS2121 in line feeding BQ VBUS will provide us an adequate solution?

I just received the TPS2121 EVM board. I plan to verify our application scenarios with the CHG-EVM  by inserting (fly wiring) across a cut between the TPS25751D PPHV and BQ VBUS input.  I plan to configure the TPS2121 in "Fast switchover" (CSP = 3.3V), with Ext DC jack INT1 level to undervoltage below 8V to hopefully observe a fast transfer to the valid USB-C source (on IN2, TPS25751D PPHV) or vice versa upon attaching the Ext DC Jack input again.

Thanks again.
Norbert

HI Norbert,

I think there is a good possibility TPS2121 will work but you might need to add additional capacitance on SYS.

Regards,

Jeff

Hi Jeff,
Using the TPS2121 switch between the TPS25751D and BQ25798, I am observing the BQ SYS shutting down. See first 2 waveforms attached.   Is it maybe due to the BQ VBUS input level change triggering the DPM settings?   I can try to adjust the VINDPM and other settings next.

How does the TPS25751D and BQ25798 work together when the switch DC source is not from the TPS25751D PPHV output?  Do you have recommendations or app note for using the TPS25751D I2Cc → BQ25798 configuration that also includes a second BQ VBUS input, either by using two ACFET-RBFET pairs, or using an external DC switch (TPS2121)?  Do you have app notes overview of the algorithm, which registers affected, etc. between TPS25751D I2Cc and BQ25798?

For our application with the TP2121 switch, and our primary use case without a battery attached requiring seamless switchover, would we be better to control the BQ ourselves, connecting our EC host directly to the BQ25798 instead of using the TPS25751D I2Cc controller?
--------
3rd waveform is of the TPS2121 EVM which I believe is stable now after adding sufficient hysteresis and decoupling caps to stop it from oscillating and latching up in the wrong switch selection mode when presented with unclean, real-world bouncing DC connections.
--------
Present CHG-EVM flash program settings from USBPD App Customization Tool:
1. TPS25751D with BQ Power source and sink
2. Source: 15W(5V)
3. Sink: 45W (15V)
4. Preferred: Power Sink
5. USB highest speed: USB2
6. Preferred data role: Host & Device
7. Support BC 1.2 CDP\
8. No liquid detect
9. Vendor ID: No
10. Product ID: No
11. BQ25798
12. INDPM: 5%
13. VINDPM: 10%
14. Charge voltage: 8.200V
15. Charge current: 1.0A
16. Termination: 0.320A
17. Precharge: 0.400A
--------
Thanks.
Norbert

Hi Norbert,

There is no additional documentation for the TPS25751 firmware.  You might try posting another forum question with only that part number for that team to answer.

If its VBUS input drops below 3.4V UVLO voltage, the charger goes through its startup routine for 250ms before the converter starts switching again.  SYS is supplemented by battery during this time.  

Regards,

Jeff

Hi Jeff,
Thanks for the reminder about the 3.4V UVLO.

I have the BQ SYS properly staying on now (see attached waveform) when the TPS2121 switches BQ VBUS from the TPS25751 PPHV 15V output to a lower priority level.  Attached waveform shows switching BQ VBUS to a DC jack input of 11.0V.  Lowering the VINDPM register setting solved SYS from turning off forever.

I struggled a while with the 4CC I2Cw to change the BQ registers.  I think the TPS25751 TRM may have omissions for I2Cw and lacked definitions of the 4CC Task Return codes. I finally found better 4CC info in the old TPS25750 TRM (Table 3.1) which helped me continue.

I'll post under the TPS25751 to understand the high level system operation of the TPS25751 I2Cc → BQ25798, affecting register parameters, and advantages of the I2Cc controlling the BQ versus our EC host.

Thanks.
Norbert