Upgrade in reliable controller use

CHIPKEN

The concept of having many processors on one chip could lead toward a system design that yields a high reliability by utilizing a redundancy setup on the cog ports. As an example four cog ports could be connected to seperate sensors in the same measurement location. The logic here would be to accept for control purposes only, at least, two sensor outputs that have nearly the same output. Cog ports operating outside acceptable sensor outputs may be shutdown if sensor failture is suspected with a reduction in processor power. All of this is being done on one chip instead of several single type processor chips which has the advantage of being more reliable for control. Am I right here?

Chuck

Dave Scanlan

Your logic sounds good to me.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔




··

CHIPKEN

Dave,
The chip logic programming ability vs reliabilty of the controller is often not discussed. Thanks for answering.

Chuck

Tracy Allen

The notion of using multiple sensors and a majority vote scheme could work with any processor, say, the BASIC Stamp, so long as the timing allows data to be acquired in sequence. Where the propeller chip could excel I think would be in its interrupt free logic, and in a relatively simple (and therefore less bug-prone) program running to service each sensor. I wonder though, if much MTBF could be gained by using two cogs instead of one. Say three thermocouples. One cog or three cogs could get data from the three and then use rules to decide if one reading is unreasonable. However, I bet one cog could scan all 3 thermocouples and implement the rules with hardly any change in reliability from using 3 cogs for them separately. The failure rate of thermocouples will be many orders of magnitude greater than the failure rate of cogs. Unless somebody pulls the plug on the propeller, and that is always a step back to the next level issue in reliability.

I don't know, maybe I'm missing your point, Chuck?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tracy Allen
www.emesystems.com

CHIPKEN

Tracy,

The question is -·does the cog concept on the Propeller chip provide a better system reliability? It does only if less chips are needed to do the same job.

Chuck

SleepMaster

Gee, if only they had had this when Shuttle's tank sensors were being deisgned. *sigh*