| ??? 08/04/06 21:15 Read: times |
#121674 - does not matter when they will happen eventhough ' Responding to: ???'s previous message |
Explain, please, how you'd cope with an event that ALWAYS occurs at precisely the instant at which your MCU will go metastabile? Such a moment always exists. How do you deal with it?
This discussion is NOT about unavoidable metastability (one case in 10egadzillion), but about flawed software/hardware that 'burp' if certain asynchronous events (e.g. interrupts) happen at certain conditions. -- see below I mentioned - for the benefit of Jez - the requirement of a synchronizer (double D) at asynchronous inputs to a CPLD/FPGA, not including such would be allowing avoidable metastability. While you're at it, explain, please, how you can design a system wherein you can't predict when things are going to happen. How will you "predict when things (e.g. external interrupts) are going to happen" If the behavior of your code is not absolutely predictable, then something is wrong in the code. Surely you won't admit to that! I do not know how to answer this construed statement but I will state that everything should work, whatever the circumstances. The behaviour of the code should be predictable even if the occurance of incidents is not. -- see below 'below' Designing faultless interrupts require concern about things like atomicity, interaction with other ints, interaction with the main .... That does not mean that you have to predict when they will happen, that means that you design you stuff so that it does not matter when they will happen eventhough 'when' is not predictable. Erik |
| Topic | Author | Date |
| reduced pin count multiplexing | 01/01/70 00:00 | |
| with e.g. ST2221A you need 3 pins | 01/01/70 00:00 | |
| No you are right Erik | 01/01/70 00:00 | |
| Well | 01/01/70 00:00 | |
| the wiring is simpler with ST2221A (equivalent) | 01/01/70 00:00 | |
| was it somone named Charlie ... | 01/01/70 00:00 | |
| ??? | 01/01/70 00:00 | |
| check out the appnote.. | 01/01/70 00:00 | |
| Reading the appnote always helps | 01/01/70 00:00 | |
| anyway | 01/01/70 00:00 | |
| who is he? | 01/01/70 00:00 | |
| walking through the description give you a charlie | 01/01/70 00:00 | |
| Small but real benefit | 01/01/70 00:00 | |
| who does your soldering? | 01/01/70 00:00 | |
| A billion dollar assembly house | 01/01/70 00:00 | |
| very small compared to | 01/01/70 00:00 | |
| Statistically or not | 01/01/70 00:00 | |
| I strongly disagree | 01/01/70 00:00 | |
| blah | 01/01/70 00:00 | |
| deterministic vs predictable | 01/01/70 00:00 | |
| and i know | 01/01/70 00:00 | |
| yes, but I am not talking about kettles | 01/01/70 00:00 | |
| just a minute, there, pilgrim ... | 01/01/70 00:00 | |
does not matter when they will happen eventhough ' | 01/01/70 00:00 |