| ??? 03/13/07 22:16 Modified: 03/13/07 22:20 Read: times |
#134925 - Kai, you can\'t compare the two environments Responding to: ???'s previous message |
Kai Klaas said:
Richard said:
One other thing that has piqued my interest is that Maxim freely published information indicating that their DS89C4x0's, in which I'm quite interested, have a flash corruption problem despite the fact they took the trouble to build in their 1232 supervisor. Which one, the DS1232 or MAX1232? They aren't identical.. No, Richard, please, have you any proof that they built-in the 1232?? I think this is entirely untenable. No, in fact, I have recent indication from the manufacturer that they, in fact, DID NOT incorporate that specific circuitry. Their datasheet simply indicates that they've incorporated some of the functionality, and, for reasons you've already outlined, they're not identical. The MCU guy tells me he never uses "resets" which presumably means he doesn't use a supervisor since there's a complex reset/brownout-detection/watchdog/oscillator-detector, etc. already built into the MCU. The mixed-signal guy says it's definitely not the DS- or MAX-1232 that's been built in. Normally, these built-in brown-out detectors are a universe inferior compared with the MAX1232 or a similar reset chip. Remember, you have much more noise directly on the die than anywhere else on the multilayer board. So, you MUST reduce the sensitivity of internal brown-out detection by lowering the threshold level and slowing down the reaction time. Additionally, you suffer from huge tolerances of threshold voltage when the brown-out detector sits on the same die as the micro.
You cannot at all compare a built-in brown-out detector with the sophisticated functionality and performance of a MAX1232. This would mean to compare the Concorde with a Cessna... Kai You're absolutely right, in that the issues with which the mfg has to deal is quite different on the die than what we have to do on a board. I'm not so sure that the difference in overall effect is much greater in either case, though. Those bond-wires to the external environment are really small and likely to amplify, by virtue of their inherent resistance, any power-gnd noise, as well as the current fluctuations demanded by the internal circuitry. There's not much opportunity for us to bypass what's inside, so the mfg has done whatever he can. What WE try to do, you and I, with our boards, is to prevent the problems internal to one IC from impacting the behavior of another. In order to have an external supervisor IC to be effective, it MUST be much more sensitive than a brownout detector internl to the IC, as the noise margins inside are much wider than those outside. After all, we not only want to avoid noise flowing out, but we want to avoid noise flowing in as well. I'm not sure what the flash-corruption issue really is. I'm convinced the IC-makers would like, simply, for it to go away. They certainly don't want to publish any reliability reports or statistical information about the behavior of their products with respect to this "bugaboo." RE |
