Post by thread:How long PIC's life?

Hi all, Does anyone know how long a PIC could works a certain code such as a real time clock which is never be interrupted or reset until it stops or hangs by itself? Let's assume that there is no bug in code at all and power supply is in a very good condition. Also watchdog is disabled. You may think this is a strange question but I think it's interesting because I'm developing a product that has to be powered all the time until user throw it away because he buys a new model, not because the old one is dead. Thanks in advance, Wiriya Sasisakulporn

> From: Thanai Anantawatanawitaya <spam_OUTthananTakeThisOuTMOZART.INET.CO.TH> > Hi all, > > Does anyone know how long a PIC could works a certain code such as a real > time clock which is never be interrupted or reset until it stops or hangs > by itself? > > Let's assume that there is no bug in code at all and power supply is in a > very good condition. Also watchdog is disabled. You may think this is a > strange question but I think it's interesting because I'm developing a > product that has to be powered all the time until user throw it away > because he buys a new model, not because the old one is dead. It will go forever. Well, at least until the chip wears out. Yes, there is a phenomenon in which the metallisation on the die surface migrates slowly in the direction of sustained current. Also, the dopants in the semiconductor junctions will slowly diffuse from thermal effects. Hard radiation also degrades the die. Atomic bombs going off in the ionosphere or nearby will probably snuff it. However, except for The Bomb, for your purposes these processes will not affect the chip until _way_ past your next product release. If The Bomb goes off then you will have other things to worry about. Regards, SJH Canberra, Australia

Well, I would say a description of what is critical is important, not asking how long a brick will stay square. Under normal operation, I am sure it will work fine for you, it is not MS Windows95 that likes to lockup for no real reason. But anyway, what is critical?, Timing? Data? Outputs? Inputs? Once you have figured that out, can implement fail-safe methods of handling the problems. A real-time clock/calendar with a super-cap to back it up is good for keeping time (I use this, it is good for over a week) , A EEprom for storing critical data, and if you are worried about data just randomly getting corrupted, you can incorporate a self test to constantly check memory and fire a reset if something does change. Well, without a description of the application I would say I have run out of guesses. (grin) Giles L. Honeycutt Thanai Anantawatanawitaya wrote: {Quote hidden}> > Hi all, > > Does anyone know how long a PIC could works a certain code such as a real > time clock which is never be interrupted or reset until it stops or hangs > by itself? > > Let's assume that there is no bug in code at all and power supply is in a > very good condition. Also watchdog is disabled. You may think this is a > strange question but I think it's interesting because I'm developing a > product that has to be powered all the time until user throw it away > because he buys a new model, not because the old one is dead. > > Thanks in advance, > Wiriya Sasisakulporn

> > > From: Thanai Anantawatanawitaya <.....thananKILLspam@spam@MOZART.INET.CO.TH> > > Hi all, > > > > Does anyone know how long a PIC could works a certain code such as a real > > time clock which is never be interrupted or reset until it stops or hangs > > by itself? > > > > Let's assume that there is no bug in code at all and power supply is in a > > very good condition. Also watchdog is disabled. You may think this is a > > strange question but I think it's interesting because I'm developing a > > product that has to be powered all the time until user throw it away > > because he buys a new model, not because the old one is dead. > > It will go forever. Well, at least until the chip wears out. Yes, > there is a phenomenon in which the metallisation on the die surface While it may do, it may at any time be hit by a decay product of some radiactive isotope in the packaging, or a cosmic ray, this may cause anything from a change in a register (internal or user) to altering eeprom/eprom/rom The first ones on the list are more probable How often will it happen? I doubt microchip charecterises chips for this, the only way would be to impact the chip with cosmic ray like events, and see what average interehaction rate there is. {Quote hidden}> migrates slowly in the direction of sustained current. Also, the > dopants in the semiconductor junctions will slowly diffuse from thermal > effects. Hard radiation also degrades the die. Atomic bombs going off > in the ionosphere or nearby will probably snuff it. However, except > for The Bomb, for your purposes these processes will not affect the > chip until _way_ past your next product release. If The Bomb goes off > then you will have other things to worry about. > > Regards, > SJH > Canberra, Australia > -- Ian Stirling. | http://www.mauve.demon.co.uk/ AKA Caeser, Bolonewbie. | With information on the PDA I'm making.

It's probably safe to assume the PIC will keep running until the rest of the circuit (particularly switches, relays, connectors) have worn out, assuming that the part is not subjected to overstress. At reasonable temperatures, humidity, and vibration levels, it should last darn near forever. ______________________________ Reply Separator _________________________________ Subject: How long PIC's life? Author: thananKILLspamMOZART.INET.CO.TH at internet Date: 11/15/96 10:37 AM Hi all, Does anyone know how long a PIC could works a certain code such as a real time clock which is never be interrupted or reset until it stops or hangs by itself? Let's assume that there is no bug in code at all and power supply is in a very good condition. Also watchdog is disabled. You may think this is a strange question but I think it's interesting because I'm developing a product that has to be powered all the time until user throw it away because he buys a new model, not because the old one is dead. Thanks in advance, Wiriya Sasisakulporn