FS8530 Turn Off and strange Control Signals

[AU_]

Hello,

we are initializing our first hardware based on the PMIC MC33FS8530A0ES. We soldered the PMIC to the board without programming and planned to program it later on the pcb via a NXP programming board from the KITFS85SKTEVM via FlexGUI and I2C. Therefore we connected the I2C bus from the NXP programming board to our hardware and removed any PMIC from the socket on NXP board.

We performed the following procedure:

• connecting VDDIO and VDDI2C to an external voltage supply (3V3)
• connecting debug pin to an external 5V supply
• disconnecting all other parts of hardware from the PMIC
• providing the PMIC with an input voltage of 12V
• Wake1 == High

The PMIC is held and awaits connection in debug mode, current consumption from 12 V supply approx. 6 mA:

• opening flexGui and initiating connection
• connection established successfully
• load OTP config file (see attachment) in script editor and program into mirror registers

The PMIC should now be set and is ready to be evaluated in emulation mode, is that correct?

After that step we remove the 5V from the debug pin which enables a pull down and starts the PMIC.

Instantly the connection to the PC seems to be lost but the PMIC seems to run for some time.

The attached oscilloscope record shows the PRE_GHS (blue), PRE_GLS (light blue/black) and Vbos (purple) signals. 

PRE_GHS seems to be okay and is in high state approx. 5 V above the input voltage of 12V. PRE_GLS seems to be also okay most of the time (e.g. 1st period, 3 periods in the middle of the record). But in some periods PRE_GLS does not turn on when PRE_GHS turns off. In This case both Transistors should be off which I guess is a undesired state, isn't it?

One can even see some kind of oscillation on the PRE_GHS signal which is really strange.

Vpre seems to have an adequate output level of 5V as long as the device keeps running. 

As the connection to the programming board is instantly lost when leaving debug mode the device turns off after some seconds. I cannot decide if this is due to a failure connected to the strange behavior of the pre regulator or just due to the safety turn off features.

What can be wrong here? The schematics seem to be very much like the reference design. I've attached a schematic sheet to this message. Only thing we already noticed we forgot was the supporting capacitors for Vbos which we added to the hardware prototype. I also attached the OTP config file for use in flexGui. We therein turned off all monitors and all regulators. Only Vpre should start. The config worked fine on the programming board.

Thank you very much for your support!

Kind regards

Andreas

[AU_]

Hello Jozef,

thank you for the quick reply. Unfortunately we were not successful. I couldn't identify some of the components you mentioned, but we tested the extra capacitor on Vpre FB pin but it didn't change the behavior.

We already mounted a 5k1 pull-up between Vbos and PSYNC, adding a capacitor to decouple didn't help either. From flex Gui we get the feedback that programming was successful, so I don't think that there should be any problem. The bootstrpping capacitor is 47 nF already. Vbos is decoupled by 4.7 uF.

I'd really like to know what's the cause for the Pre GHS signal to look like in the osci record. Furthermore it would be helpful for us to better understand why the PMIC is turned off. Is there any helpful error register we can read? 

Our Vpre is now loaded with no load, can this be a problem? Is there a minimum load for stable operation?

Thank you very much in advance 

Kind regards

Andreas

[JozefKozon]

Hello Andreas,

I'd really like to know what's the cause for the Pre GHS signal to look like in the osci record. Furthermore it would be helpful for us to better understand why the PMIC is turned off. Is there any helpful error register we can read? 

[A] I suspect the R77 and R79 resistors are too low. According to the KITFS85AEEVM evaluation board, the R77 and R79 should be 1MOhm. In your case they are 100kOhm, which might pull the signal too GND. R77=R27 and R79=R26. The values of the R77 and R79 should be different for programming and for the use case. The R77 and R79 should be 100kOhm for programming and 1MOhm for use case. And the R77 should be connected to GND for programming, and for the use case the R77 should be connected to the PRE_SW pin. Please see below screenshots for differences.

• KITFS85AEEVM evaluation board

     

• KITFS85SKTEVM Programming Board

At least for the programming of the FS8530 I would recommend you to purchase the KITFS85SKTEVM, which was tested for programming of our SBCs and is guaranteed, that the SBC will be programmed correctly.

Our Vpre is now loaded with no load, can this be a problem? Is there a minimum load for stable operation?

[A] The SBC/VPRE should also work correctly in zero load condition. You can check the VPRE_G bit in the M_FLAG register, if the VPRE event has happened. 

With Best Regards,

Jozef

[JozefKozon]

Hello Andreas,

thank you for the detailed description and for the files. I have checked the KITFS85SKTEVM programming board and the KITFS85AEEVM evaluation board schematics and it seems, that for programming and for using the programmed FS85 the configurations are slightly different.

In both cases the VBOS should be also connected through a 10kOhm resistor to PSYNC pin with 1nF decoupling capacitor. Perhaps this is the reason why the FS85 turns off after few seconds.

If this doesn't help, please try the following modifications on top of connecting the PSYNC through the pull-up to VBOS.

For programming please follow the SCH-SOCKETED-FS84_FS85_Programming.pdf attached. Please connect the R77 to the GND. Please add 22pF capacitor to the VPRE_FB pin. Please increase the value of the C84 to 47nF. Please connect the 4.7uF capacitor between the VBOS pin and GND. 

For evaluation of the programmed FS85 please follow the FS84_FS85_EVB.pdf schematic attached. Please connect the R77 back to the PRE_SW pin and please increase the R77 and R79 values to 1MOhm. Please add 22pF capacitor to the VPRE_FB pin. Please increase the value of the C84 to 47nF. 

With Best Regards,

Jozef