TPS3813: Are WDI rising and falling edge valid as a WDI input?

Kazuya,

A1. Yes, either rising or falling edge on WDI is detected to prevent device from resetting. If no edge is detected, device resets.

A2. If falling edge occurs before t_boundary,min then device resets (/RESET goes low). To prevent device from resetting, falling or rising edge must occur between t_boundary and t_window. Note: t_boundary has min/max tolerance and t_window has min/max tolerance. To be certain device does not reset, trigger between t_boundary,max and t_window,min. In Fig. 1, the shaded region represents the tolerance of t_boundary which means the t_boundary can be between t_boundary,min and t_boundary,max and you must pulse WDI after t_boundary elapses to prevent reset.

A3. If WDI falling edge is input after t_boundary,max, this means that no falling edge is detected within the window frame so device resets (/RESET goes low) and WD timer is reset after the delay.

A4. Rising or falling edge must occur after t_boundary and before t_window to prevent reset.

A5. If rising or falling edge occur before t_boundary or after t_window, the device will reset.

Please let me know if you have any more questions of if anything is unclear! Please click “This resolved my issue” button if the answer provided resolved your issue. Thanks!

-Michael

Kazuya,

A1. The pulse width shouldn't matter since the device looks for edges instead of pulse duration. Table 6.6 requires a min pulse width of 3us at VDD so I would try to apply at least pulse widths of 3us if possible. "t_boundary" doesn't represent a region but rather a specific time. That time varies slightly from device to device which creates the "t_boundary, min, typ, max" region. You want your pulses to occur after t_boundary for your device. The t_boundary can be as far out as t_boundary, max. If your pulse occurs before t_boundary, this is outside of the window frame thus the device will reset.

A2. No need to apologize! We are here to help. The diagram is actually very confusing. There is no label explaining the shaded region in Figure 7 so it is a little confusing. In figure 7, there should be a "lower region", "window frame", and "upper region" with the good region that you want to aim for is "window frame". This is described in much better detail in figure 6. In figure 7, I believe the shaded region is suppose to be the middle good region for window frame. The reason the device resets in "trigger pulse lower window boundary" is because after the first good pulse, the second pulse occurs right after which would be in the lower region which is before the middle window frame region so this would cause reset.

Please let me know if this is unclear or if you have any more questions. Thanks!

-Michael

My apologies for the delayed response.

A1. Yes, If WDI falling edge (negative edge) is input before t_boundary_min, /RESET=L occurs. Fig1 diagram doesn't show the lower boundary very good. It shows a shaded region but this should represent the window frame (the middle region). Please see Figure 7 for a more detailed timing explanation. In Fig7, you can see the lower boundary, window frame, and upper boundary. To prevent reset, WDI must pulse inside the lower and upper boundaries (in the middle window frame).

A2. You are correct. The WDI terminal is connected to a rising edge detector. The text is wrong. I have ordered some EVMs to double check this but I believe the WDI can only detect the rising edge. This is how it is explained in the automotive version (TPS3813-Q1) as well.

If my testing finds that both edges are detected, I will let you know. Thanks!

-Michael