Yahoo Groups archive
Thread
2008-06-17 by David Appleton
Don't AVRs have a WatchDogTimer? This is used to ensure that a processor hasn't gone off into a bad place. If your software doesn't "service" the Watchdog Timer, the processor will go into a reset condition. [Non-text portions of this message have been removed]
2008-06-17 by Dennis Clark
> Don't AVRs have a WatchDogTimer? > > This is used to ensure that a processor hasn't gone off into a bad > place. If your software doesn't "service" the Watchdog Timer, the > processor will go into a reset condition. > Yes, they do. I prefer external hardware watchdogs myself. DLC > > > > [Non-text portions of this message have been removed] > > > ------------------------------------ > > Yahoo! Groups Links > > > > -- Dennis Clark TTT Enterprises
2008-06-17 by Zack Widup
Me too. I usually turn the internal WDT off. Zack
On Tue, 17 Jun 2008, Dennis Clark wrote: > >> Don't AVRs have a WatchDogTimer? >> >> This is used to ensure that a processor hasn't gone off into a bad >> place. If your software doesn't "service" the Watchdog Timer, the >> processor will go into a reset condition. >> > Yes, they do. I prefer external hardware watchdogs myself. > > DLC >>
2008-06-17 by Philippe Habib
A lot of the stuff that I work on, the internal WDT is not allowed for regulatory reasons. I must use an external part.
On Jun 17, 2008, at 9:57 AM, Zack Widup wrote: > > Me too. I usually turn the internal WDT off. > > Zack > > On Tue, 17 Jun 2008, Dennis Clark wrote: > >> >>> Don't AVRs have a WatchDogTimer? >>> >>> This is used to ensure that a processor hasn't gone off into a bad >>> place. If your software doesn't "service" the Watchdog >>> Timer, the >>> processor will go into a reset condition. >>> >> Yes, they do. I prefer external hardware watchdogs myself. >> >> DLC >>> > > ------------------------------------ > > Yahoo! Groups Links > > >
2008-06-17 by Enki
What is the advantage of using an external WDT instead of the internal one? Mark Jordan
On 17 Jun 2008 at 11:57, Zack Widup wrote: > > Me too. I usually turn the internal WDT off. > > Zack >
2008-06-17 by Philippe Habib
In my case, it is required. The fear is that if the micro dies in such a way that the watchdog doesn't work either you'd be in trouble, so you have an external watchdog which does the reset. You also have some other way of detecting a dead micro and then arrange to fail in a "good" way.
On Jun 17, 2008, at 2:33 PM, Enki wrote: > > What is the advantage of using an external WDT instead of the > internal one? > > Mark Jordan > > > On 17 Jun 2008 at 11:57, Zack Widup wrote: > >> >> Me too. I usually turn the internal WDT off. >> >> Zack >> > > > ------------------------------------ > > Yahoo! Groups Links > > >
2008-06-18 by David Kelly
On Jun 17, 2008, at 5:45 PM, Philippe Habib wrote: > In my case, it is required. The fear is that if the micro dies in > such a way that the watchdog doesn't work either you'd be in trouble, > so you have an external watchdog which does the reset. You also have > some other way of detecting a dead micro and then arrange to fail in > a "good" way. So its based on fear rather than fact? Fear that the watchdog is implemented as a software function that will fail the same way when/if the CPU fails? But if that was the case wouldn't the CPU be smoked and beyond recovery via reset? -- David Kelly N4HHE, dkelly@HiWAAY.net ======================================================================== Whom computers would destroy, they must first drive mad.
2008-06-18 by Roy E. Burrage
The external WDT can easily be used to set off an external alarm if the controller gets too pooped to participate. It can also be used to turn on another controller whose job it would be to set off an alarm and shut down the system in a reasonable manner to a safe state. David Kelly wrote: > On Jun 17, 2008, at 5:45 PM, Philippe Habib wrote: > > >> In my case, it is required. The fear is that if the micro dies in >> such a way that the watchdog doesn't work either you'd be in trouble, >> so you have an external watchdog which does the reset. You also have >> some other way of detecting a dead micro and then arrange to fail in >> a "good" way. >> > > > So its based on fear rather than fact? > > Fear that the watchdog is implemented as a software function that will > fail the same way when/if the CPU fails? But if that was the case > wouldn't the CPU be smoked and beyond recovery via reset? > > -- > David Kelly N4HHE, dkelly@HiWAAY.net > ======================================================================== > Whom computers would destroy, they must first drive mad. > > [Non-text portions of this message have been removed]