Post by thread:[PIC]: Too hot to wake up

Alice Campbell wrote: {Quote hidden}> >Esteemed Piclist: > >Once i discovered that the ADC in my datalogger wouldnt run >at 33 kHz, i replaced the watch crystal with a 4 MHz crystal. > There are two caps, 47 pF. The dataloggers live outside,and >sleep 7 minutes to one hour. Its pretty hot out, about 100 >deg or more in the daytime. I have a temperature >compensation routine to correct the sleep time for >temperature, that isn't the problem. While the circuits run >pretty well under about 80 deg, once they get hot, they only >sluggishly wake up, one hardly wakes up at all. I tried a >.1uF cap across the pic's power pin to ground, to boost it a >little, helped a little but not fixed. I changed the caps to >33 pf but thats worse. The capacitance of my finger will >generally wake up the more sluggish unit, but i really cant >stand around all day slapping the thing to make it run. Im >starting to reach for the big rubber mallet, does anyone have >any suggestions? > >Thanks, >alice > Hello Alice, I have never tried sleep mode with a PIC, myself, nor tried operating a PIC at 100 degF [although all I would have to do is open the outside door around here right now], but the first thing that occurred to me was - Do the loggers start up reliably from a "hard" reset at these temperatures? You may be looking at the wrong problem. - Also, you might want to look at using an industrial temp range PIC - it may be way over 100 degF inside the box, esp in the sun. [might set up one of your channels to measure internal box temp]. - Also, maybe you battery is pooping out at those temps. - Also, you might check the datasheets re operating temperatures for the crystals you are using, and possibly try other xtals. Someone else may know if resonators work better at hi temps. Might also check http://www.murata.co.jp - Also, when you went from 32 Khz, you did change over from the LP to the XT oscillator, didn't you? If all else fails, you might even try the HS oscillator - [really make the xtal's day]. - Also, interestingly, the PIC datasheet says using lower value caps should "decrease" the startup time - but yours appears to work the other way. Hmmmm????? Still, you might try the low end value shown in the datasheet, 15 pF. - [the key to fixing this sort of thing is to just keep asking questions, and testing things - and keep your sun hat on]. Maybe someone else has had hi-temperature experience with PICs ???????? best regards, - Dan Michaels Oricom Technologies =================== -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]: PIC only [EE]: engineering [OT]: off topic [AD]: advertisements

Dan Michaels wrote: > Maybe someone else has had hi-temperature experience with PICs ???????? We've successfully long-term tested 16C74s at 150 degC and 16F877s at 175 degC (hi temp versions), and have had no start-up problems either from a cold start or from sleep mode. Both designs used standard crystal configurations: 3.6864MHz (XT) for the '74 and 11.0592MHz (HS) for the '877. As Dan said - check that oscillator fuse. A commercial temperature rated PIC should operate much higher than 100 deg F. If the problem is temperature related, the battery and/or the crystal would be likely suspects (again, as Dan suggested). How much start-up current does the logger require? Or, are you talking about the crystal on an external ADC? If you are, I've seen start-up problems with Analog Device's 7713 ADC, both from sleep and power-up. --Matt -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]: PIC only [EE]: engineering [OT]: off topic [AD]: advertisements

Following forwarded to piclist, as proxy for Alice: ================================================== > I have never tried sleep mode with a PIC, myself, nor tried > operating a PIC at 100 degF [although all I would have to do is > open the outside door around here right now], but the first thing > that occurred to me was > > - Do the loggers start up reliably from a "hard" reset at these > temperatures? You may be looking at the wrong problem. Yes. for reasons lost in time, the mclr button is used as an interrupt for this circuit, and triggers a dump to the rs232 line. it incidentally will fix any error conditions and leave the machine in a known, initialized state, but NOT change the main count. see, the calibration of the sleep timer relies on generating a wdt timeout, and parsing out the wdt timeout was enough of a nuisance that it was easy to parse out the mclr reset and use it instead of an interrupt. as long as all the subroutines have flags, i can find where it left off, which is inevitably in the sleep routine. > > - Also, you might want to look at using an industrial temp range > PIC - it may be way over 100 degF inside the box, esp in the sun. > [might set up one of your channels to measure internal box temp]. > yes, i am considering this--actually, all i have to do is save the sleep timer calibration variable. But i am annoyed that the same pic + board worked fine with the 33kHz crystal and 47 uf caps. the change to the circuit is just the crystal and caps, nothing else. i promise, i was there soldering. > - Also, maybe you battery is pooping out at those temps. > battery voltage is great. the stupid thing just sleeps all the time. I can tell from the battery voltage, which is maybe 8.7 volts and 3 weeks doesnt even dent it. i think i licked the earlier power leak from not trissing all the pins and setting them low. > - Also, you might check the datasheets re operating temperatures > for the crystals you are using, and possibly try other xtals. > Someone else may know if resonators work better at hi temps. > Might also check hmm, have some 4.1 colorburst crystals in bottom of bin,... <mumbles about all those timing loops for 2400 baud> > http://www.murata.co.jp > > - Also, when you went from 32 Khz, you did change over from the LP > to the XT oscillator, didn't you? If all else fails, you might > even try the HS oscillator - [really make the xtal's day]. > xt is enabled. i was about to try HS next, but it seemed a little paltry. Another thing i am thinking about is locating and trying 100khz crystals, i dont need speed, i need reliability. Junk box just doesnt have any 100kHzs. > - Also, interestingly, the PIC datasheet says using lower value caps > should "decrease" the startup time - but yours appears to work the > other way. Hmmmm????? Still, you might try the low end value shown > in the datasheet, 15 pF. This i will try. maybe i just didnt push it far enough, or there is a peak and im not across it yet. Another thing i will try is unbalanced caps, say a 15 pf and a 20 pf, to rock the boat. also a cheezy carbon resistor on the inner pin, to generate some noise. > > - [the key to fixing this sort of thing is to just keep asking > questions, and testing things - and keep your sun hat on]. > i was afraid of that. thanks, i will keep slogging and report progress. alice -- http://www.piclist.com hint: PICList Posts must start with ONE topic: [PIC]: PIC only [EE]: engineering [OT]: off topic [AD]: advertisements

> We've successfully long-term tested 16C74s at 150 degC and > 16F877s at 175 degC (hi temp versions), and have had no > start-up problems either from a cold start or from sleep > mode. Both designs used standard crystal configurations: > 3.6864MHz (XT) for the '74 and 11.0592MHz (HS) for the > '877. As Dan said - check that oscillator fuse. Are you sure 150 C ??? We have performed several projects run up to 125C but SUCCESSFULLY LONG-TERM TEST on 150 C sounds like real magic. I'm interested what PCB laminate you have used for your PCB as FR-4 would melt and deformate at temperatures over 130C. What type of capacitors (especially electrolyte caps) you have been used? All quartz crystal specs shows up to 70C operation. What happend with them at 150C after 1 week operation? Even don't dare to ask for the connectors, switches and etc. plastics as for high temperature the recommended parameters derate down to 10% of the normal (for instance the carrying current). Tsvetan --- PCB prototype for $26 at http://run.to/pcb -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

Tsvetan Usunov wrote: > Are you sure 150 C ??? Yes. > I'm interested what PCB laminate you have used for your PCB as FR-4 would > melt and deformate at temperatures over 130C. It also deforms at temperatures below that, too, after enough time. We don't use FR4 - I'm not sure which material it is, but it's rated to about 185 degC. Solder selection is important, too. (I do the electronics and firmware - the production department takes over from there.) > What type of capacitors (especially electrolyte caps) you have been used? Tantulum. > All quartz crystal specs shows up to 70C operation. What happend with them > at 150C after 1 week operation? Nope, you can get high temp crystals and oscillators. > Even don't dare to ask for the connectors, switches and etc. plastics as > for high temperature the recommended parameters derate down to 10% of the > normal (for instance the carrying current). One connector, again Mil rated, no switches or sockets. All this stuff is *very* expensive, though. --Matt -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics