Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
05/31/08 14:14
Modified:
  05/31/08 14:35

Read: times


 
#155328 - Have you ever been involved in CE testing?
Responding to: ???'s previous message
Jan said:
...that EMI sufficiently strong to throw the micro out of way, never occurs.

Jan, CE standards are so demanding, that your product is immune against 99.9% of all strange effects.

Jan said:
And how did you come to the conclusion that Alex's case is not due to unforeseeable situations?

Because he talked about "lightning". And CE complying products are immune against lightning, as they are "surge" tested, which is a very harsh test by the way. If you design your power supply unit properly, lightning on mains is totally filtered out of your supply voltage. (I do not talk here about a direct hit, of course, but which is very very seldom.)
He also talked about "unstable" power. When being designed properly, then only power downs and power up can be critical and these aren't cured by the help of a watchdog, but by a proper supervisory/reset controller.

So, this all tells me, that Alex has an improper design which he wants to cure by the help of a watchdog.
Jan, our products are designed in that way, that the watchdog never starves. We made long term tests and found that the watchdog not even once had to reset the micro.

Jan said:
Ask the utility (electricity) provider. You cannot simply say "power dropouts shall not occur" if they DO occur in certain regions. You have a slightly different perspective: in Germany you have a substantially more stable and controlled utility grid than say here in Slovakia; and also it's less likely to have strong EMI injected into the grid by faulty devices.

Jan, but that has nothing to do with the watchdog! What are you telling here? If you have mains dips or other fluctuations of mains voltage, then you still have the big electrolytic storage cap which filters much of the dips out. And the rest of fluctuations is handled by the help of your reset controller, just as simple power up and power down situations. The watchdog has never to bite in these cases.

Even fast spikes on mains are no problem if you use proper filtering. Remember, your CE testing includes very demanding "burst" and "surge" testing. I have seen improper parts exploding during the surge testing...
We use three pole mains filtering, PTCs from Epcos and transzorbs on secondary side of transformer, caps across the rectifier diodes, switchers as preregulators and 7805 for fine regulating. Each step with two pole lowpass filtering including ferrite beads from "Würth" and ceramic high caps from "Taiyo Yuden". Even behind the 7805 there's two pole low pass filtering on the boards, when using mixed analog/digital circuitry. Again, the watchdog did never bite.

Kai

List of 48 messages in thread
TopicAuthorDate
Looking for a 80XX chip with build in Watch Dog            01/01/70 00:00      
   Complex?            01/01/70 00:00      
   they are legio            01/01/70 00:00      
      Rx2, 66x, LPC, 80F ...            01/01/70 00:00      
         you will easily recognize            01/01/70 00:00      
            Ateml            01/01/70 00:00      
               probably            01/01/70 00:00      
   Why is the power unstable?            01/01/70 00:00      
      ANY certificate is NO absolute guarantee...            01/01/70 00:00      
         Have you ever been involved in CE testing?            01/01/70 00:00      
            oh, how many times have that been the case            01/01/70 00:00      
            why do you use WD then?            01/01/70 00:00      
               Recovery            01/01/70 00:00      
                  A quote from "The Firmware Handbook":            01/01/70 00:00      
               the fallacy of 'testing'            01/01/70 00:00      
                  why redesign            01/01/70 00:00      
                     'normal' is where it is to be located            01/01/70 00:00      
                        depends on the circumstances            01/01/70 00:00      
                           reduced            01/01/70 00:00      
               To calm my boss...            01/01/70 00:00      
                  why would I not know how it works?            01/01/70 00:00      
                     wait a minute ...            01/01/70 00:00      
                        I couldn't agree more...            01/01/70 00:00      
                     I wanted to focus on something different            01/01/70 00:00      
   Good Design vs. Uncontrollable.            01/01/70 00:00      
      I am glad I am not your customer            01/01/70 00:00      
      What you CAN do            01/01/70 00:00      
         and            01/01/70 00:00      
            but it might save the engine...            01/01/70 00:00      
               neither will a watchdog qualify ....            01/01/70 00:00      
                  It was Neil Kurzman...            01/01/70 00:00      
                  watch the dog            01/01/70 00:00      
                     Yes, I agree with you            01/01/70 00:00      
            I know that it is not a good resolution.            01/01/70 00:00      
         I wish I can do that            01/01/70 00:00      
            I think you need another job            01/01/70 00:00      
            Try googling "SEPIC"            01/01/70 00:00      
               one of many            01/01/70 00:00      
                  It's a type, not a brand ...            01/01/70 00:00      
            Try this            01/01/70 00:00      
               I would recommend against it            01/01/70 00:00      
                  If you have had a look at the datasheet...            01/01/70 00:00      
                     you missed a word            01/01/70 00:00      
               I've had mixed results ...            01/01/70 00:00      
                  Did you test the LM2674 actually?            01/01/70 00:00      
                     I wouldn't call it testing , but ...            01/01/70 00:00      
                        Are you sure it was the LM2674?            01/01/70 00:00      
                           Well ... as it turns out ...            01/01/70 00:00      

Back to Subject List