MC34978 Read switch status register doesn't change

[Pascal06]

Hello dear community,

I ported to a Microchip PIC16F the MC34978 SDK

To test my program on a real pcb, I'm using the SDK example: basically by first initializing the MC34978 with POR default parameters and then by reading the switch status register each 500ms (except that I configured all inputs at "switch to ground".

To verify if it works, I'm using a digital analyser with SPI decoder and in the same time I'm pulling one or other inputs to ground. The problem is that nothing change in the second read switch status 32 bits transfer.

Read switch status master command:

First 32 bits transfer (MOSI 0x3E "Read switch status" command)

Read switch status slave reply (note that for an unknown reason, all switches except SG0 are seen as high state):

Second 32 bits transfer (MISO 0x3E reply)

A few time later, I'm reading the Fault Status Register,
First 32 bits transfer (MOSI 0x42 "Fault status register" command)

But there isn't any error:
Second 32 bits transfer (MISO 0x42 "Fault status register" reply)

I can't understand that nothing happen.

Here the partial schematic:

MC33878

Is anybody can help me ?

Thanks a lot in advance.

[guoweisun]

Hi

What are the value of VBAT and VDD?

What's the value of R32?

[Pascal06]

Hello,
thanks a lot for your reply.
VBAT = 6V, VDD = 4V
R32 to R53 = 100 ohms (like in the application schematic)
I added a wake_b pull-up, and it is the same thing.
I verified the INT_B pin while I pull down one of the pin, nothing happen too...

It looks like the IC detect nothing... IC's build is CTEP2140E (so 40th week of 2021)

[Pascal06]

Hello,
about exposed pad, datasheet says that "it is recommended that the exposed pad is terminated to GND (pin 1) and system ground".

But can a none connected exposed pad induce this kind of problem ?

[guoweisun]

Hi 

You can try to connect this exposed pad to ground for getting better heat dissipation。

I assume you use the verified software so you need pay more attention on the hardware.

[Pascal06]

Hello,

thanks a lot for your reply... I admit that could be a hardware issue, but wow, on such a simple design it's crazy !

I wanted to test my hardware by removing the "faulty" MC34978 on one of my boards, and link the uC to a KIT34978EKEVB board.

What a shame, it seems that theses kits aren't available any more.

[guoweisun]

Hi

How many boards do you test?

can you use the 12V battery voltage to try again?

[Pascal06]

Hi,

thanks a lot for your reply.

I'm using two boards (these are custom design, not NXP evalution board).

I already tested with VBAT=12V but it is the same thing.