This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

MSP430 Absolute maximum and recommended operating voltages

We have a device using a Lithium battery with an open circuit voltage of ~3.7V (Li-SOClbatteries).  We're planning to use an LDO as a regulator, but I was wondering, what are the consequences of not using a regulator?  The maximum recommended voltage is 3.6V, but the absolute maximum voltage is 4.1V.  

  •  Hi, till voltage never exceed 4.1V can be none,  just more power drawn from battery. I left an application running from a 3 AAA cell with a diode in series from 2007 and is still running now, battery voltage dropped from initial cell voltage of 4.7 to under 4 volt battery expiry date has passed few mount ago.

     In the past and recent due to a fault at LDP more than one time MSP got powered from 6V rail and visibly nothing appeared, this has not to be used as a rule.

     None of the two sentence can be used as a rule

     Regards

     Roberto

    march 10 editing to post

  • Roberto Romano said:
    till voltage never exceed 4.1V none

    You cannot assume that!

    Outside of the specified operating range, the performance is undefined:  that means that absolutely anything might happen - it doesn't have to be reliable, or reasonable, or repeatable. The only guarantee is that the device will not be damaged.

     

  • Buffalo,

    If you're just playing around with the part and not planning to go to production with the design, 3.7V is probably OK.  However, officially we won't guarantee anything over 3.6V, as it can degrade the part's performance.  So for production, you want to use a regulator and keep Vcc under 3.6V.

    Hope this helps,

    John

  • Andy Neil said:

    till voltage never exceed 4.1V none

    You cannot assume that!

    Outside of the specified operating range, the performance is undefined:  that means that absolutely anything might happen - it doesn't have to be reliable, or reasonable, or repeatable. The only guarantee is that the device will not be damaged.

    [/quote]

     Hi Andy, on my post some correction need to be applied, in first:

     this can't be a rule -> .Dot.  These can't be rules -> SO read as : none of both sentence can be a rule.

     On your writing from experience I have to say more:

     Absolute and operating range: the 200 unit MSP that where posed in operation never failed and still are working near absolute maximum rating, useful life ended early due to the assigned task and approach to a stop due to battery life.

     The failure of power supply unit teach more: Absolute maximum rating 11V operating range to 9V, the example design powered from a 7V source, one of the first sample unit failed @ 8V, so I lowered the supply voltage, a new part where announced by TI but not available for production. Over a 3000 unit lot on a different board type (1 main controlling 2 peripheral of an application) we learn PSU was some way defective, we lowered again the main power to 6.8V (the minimum acceptable), some near 25% part failed again on a random basis, in the main unit power supply a double footprint where prepared to carry the new psu as it become available in the future, so on half the lot all PSU unit where removed and exchanged.. After recall all 1000 unit where reworked to change the defective chip.

     SO learned some chip survive to more than absolute maximum rating, other cannot withstand the operating range..

     Again THIS CANNOT BE A RULE!

     A chip or an assembly that don't respect the specification can be marked as defective.

     Regards

     R

  • Yes, there most certainly are rules - but you are misinterpreting the rules!

    The Absolute Maximum ratings tell you limits below which it is guaranteed that the device will not be damaged - they do not say that the device will fail if you exceed them.

    Similarly, the Operating Conditions tell you limits within which it is guaranteed that the device will operate as stated - they do not say that the device will not operate outside those limits; just that you have no guarantee.

     

  • Andy Neil said:

    Similarly, the Operating Conditions tell you limits within which it is guaranteed that the device will operate as stated - they do not say that the device will not operate outside those limits; just that you have no guarantee.

     Andy, the MSP operate well over the limit, none failed also the part exposed to 6.8V, on over 10000 MSP unit I used just two failed due to PSU failure from 12V rail.

     THe PSU I reported was INSIDE the declared operating limit!!

      Regards

     Roberto

  • Roberto Romano said:
    the MSP operate well over the limit,

    You're missing the point again!

    Nothing says that they won't operate - just that you have no guarantees.

    Roberto Romano said:
     none failed also the part exposed to 6.8V, on over 10000 MSP

    Proving only that you've been lucky!

    I was run over by a car once and, obviously, I survived - but that is clearly no reason to suggest that anyone else should try it!

  • Roberto Romano said:
     Andy, the MSP operate well over the limit, none failed

    It depends on what you see as a fail. Above 3.6V I can imagine several things: writing to flash with >3.6V results in flash retention time not met or even improper flash programming. As the flash charge pump produces a voltage that is too high. But once VCC is reduced, all is well and no permanent harm is done. Having more than 4.1V VCC might result in a programming voltage that permanently damages the flash, perhaps significantly reduces the number of write cycles.
    Or with VCC above 3.6V, leakage currents significantly increase, so entering LPM won't reduce power consumption to anything near the values in the datasheet.

    Maybe the ADC precision or the voltage reference are outside specs now. Or the crystal oscillator draws more than specified. Maybe the port pin output drivers cannot stand a shortcut anymore on this output voltage.

    Maybe none of them has been observed by your applciations yet, and perhaps you never will (as you're not programming the MSPs more than a couple of times, don't draw maximum current form the port pins, don't care for a few µA standby current). Still the device would fail a complete check against the datassheet specs.

    Failure is not limited to immediate and total destruction.

    To make it simple: outside the recommended operation conditions, you cannot trust any min/max value in the datasheet anymore. And once you've been outside the maximum ratings, all other information in the datasheet becomes quesitonable too.

    It may still be acceptable and reliable for your specific application.

  •  Andy, I am so sorry you get run over by a car, can this have damaged you????

     OK Again please don't broke phrases, on production where the PSU failed all parameters where in range, to answer also to Jeans, this threat was on 2003 so all MSP are still in place and still reprogrammed on board and still reacting to erasing and restoring data on information area program and data transferred are subject to security check so CRC and checksum apply'd, no more exposed to out of range power supply and none failed at this date. Some can be got removed by service but due to financial failure of customer organization.

     All PSU unit (3000 unit!!!) where dropped to trash and exchanged so on first run about 200 unit where exposed to this failure and some 300 other on field. This exposed about 900 pieces of MSP to more than absolute from a failed pieces that WAS ON SPECS TOO!!!!!

     

      Andy say about I am lucky.. I don't believe that.. PSU chips where absolutely defective.. over 50% defective is a "not to use this rule".

     

     MSP chips got a so low failure rate to be classified as extremely reliable.

     Statistic driven result say that : THIS IS THE ONLY good RULE!!!!

     From statistic data if someone get runned by a car truck or train we experience a very low survival percentage.. in this scenario where p factor is near zero and q near ONE we just can say Andy IS VERY lucky and none other we know can survive this experience so populate just dead men set.

      So as a statistic driven data rule: Andy take care about car when you walk around.

     Regards

    Andy Neil said:

    the MSP operate well over the limit,

    You're missing the point again!

    Nothing says that they won't operate - just that you have no guarantees.

    Roberto Romano said:
     none failed also the part exposed to 6.8V, on over 10000 MSP (Missing part of my phrase here!!)

    Proving only that you've been lucky!

    I was run over by a car once and, obviously, I survived - but that is clearly no reason to suggest that anyone else should try it!

    [/quote]

  • Roberto Romano said:
    MSP chips got a so low failure rate to be classified as extremely reliable.

    I totally agree. It' sastounding how much an MSP can stand when operating it outside the specs. Even more if inside the specs. All failures I have encountered so far on our devices could have come from improper soldering (applying 5V/2A instead of regulated 3.6V due to an improperly soldered regulator, improperly soldered GND pins etc.) and other external reasons. And even of those, many survived after all was fixed.

    What I was talking about was that even if an MSP works properly after (or even while) operating outside the specified limits, it does not mean that it didn't take any damage at all. Maybe the damage was done to parts you don't use in your application, so you'll never notice. Or maybe you use them, but the 'damage' doesn't change the behavior. E.g a shift in the low/high threshold of the port pins, which is unimportant as logn as your input levels are far above/below teh threshold anyway. Or an increase of offset or gain error in the ADC, which is unimportant because you calibrate it by your own anyway, so you never notice. So the device would fail a thorough test against the specs listed int the datasheet, while it doesn't fail for your specific application.
    It's dangerous to generalize here.