TPS3828: WDI pin

Hi Ruixin,

I am looking into this and will provide feedback early next week.

Thanks,
Joshua

Hi Joshua,

this issue is important in customer side. Could you pls help to check ASAP? Thanks!

Rayna

Hi Ruixin,

1. The frequency of the Auto-Reset Oscillator is shorter the 0.9s watchdog timeout min, so as to always provide pulses in the floating/high impedance WDI configuration.

2. The internal transition detector of the TPS3828 detects the drive strength of the input at WDI. If the pulse is strong enough pulling either to VDD/GND the transition detector does not detect the "transition" (VIH to VIL or vice-versa from the Auto-Reset Oscillator) meaning there is no pulse and timeout occurs.

Thanks,
Joshua

Hi Joshua,

Customer need the exact value of the frequency of the Auto-Reset Oscillator, could you pls help to check it ? Thanks!

Rayna

Hi Ruixin,

I am requesting this information, but am not certain whether it will be available. I will update as soon as I receive feedback. 

Thanks,
Joshua

Hi  Joshua,

Thanks! looking forward to it.

Rayna

Hi Ruixin,

I cannot give you an exact value that all devices will oscillate at this exact frequency, however the internal Auto-Reset Oscillator operates at some unknown frequency (>5Hz) which is then digitally divided down to the needed watchdog timeout frequency. Please see below scope shot example of device oscillating at ~0.625Hz.
 

Thanks,
Joshua

Hi Joshua,

If the internal oscillator feed the dog every 1.6s. then it will have a problem here. If the internal oscillator fed the dog at 0s, then customer pull low the WDI at 1.5s, and the pulse on WDI is given in, for example, 0.5s. then it means the dog is fed after 2s. it exceed 0.9s so it will reset. That's the problem customer met here.

could you pls give some advice here? how to solve it?

thx!
Rayna