Long delay affected by low temperature?

[ronsimpson]

I was just reading up on EC.

We still have possibility of bad solder joints.

 

[AGCB]

But, since you are running the motor and PIC off the same battery, a voltage sag or spike could be an issue. When the motor is colder, it will draw more current, causing a larger voltage drop when it starts.

This is the most feasible reason I see. (It also answers my OP ? [Should temperature affect oscillator timing or anything else?]) Since it all works so flawlessly on the bench, most of the other possibilities are less likely. That being said I have provided a separate power supply to the MCU side of the circuit via a 3)AA battery box so if there is a voltage drop in the 6 volt battery it won't effect the operation of the PIC.
I will try again outside and let you know what happens.

Can you post a complete schematic that's bigger than a postage stamp - including ALL components. What you posted above is unreadable.

Mike
I don't use any of the schematic programs. Usually just hand draw something but in this case it was so simple that I just built it on a breadboard and tested before transferring to a solder board. The output amp circuit I got online and drew on the graph paper that was in the photo above (post #9). Sorry it didn't come out more readable.

Thanks to all who have responded! I have read and considered your ideas and have not discounted them.
Aaron

 

[Pommie]

Mike
I don't use any of the schematic programs. Usually just hand draw something but in this case it was so simple that I just built it on a breadboard and tested before transferring to a solder board. The output amp circuit I got online and drew on the graph paper that was in the photo above (post #9). Sorry it didn't come out more readable.

What you don't seem to understand is that what you assume to be valid may be totally wrong. Post a schematic, hand drawn or otherwise and we may be able to help. There are many questions asked above that you have still not answered - answering them may get you help.

Mike.

 

[AGCB]

Here's a hand drawn complete schematic and a picture of the board before the separate power was added to the MCU side.

 

[AGCB]

On the schematic there is a common ground for all.
If there are still ?s please list them simply 1., 2., 3.
Thanks
Aaron

 

[Pommie]

So, have you tried switching to EC oscillator?

Mike

 

[sagor1]

Agree if that is an external oscillator, then use the EC_ option for clock source. Is that clock source always stable? (is there any other way to check it?)
Your grounds have to be tied together, else the GPIO output has no reference to the NPN ground.
And a reminder about my final point, the P channel MOSFET may not be turning on fully with lower gate voltages. While it may start to conduct at -4V Vgs, your 6V supply may not be enough to turn it on fully. AT colder temperatures, this may be a small factor, though I tend to agree it is more likely something around the PIC circuit itself...

Another test is to turn off the MCLR (reset) in the config and see is that clears it up. That will determine if there is some fault (power or noise) that is triggering the reset. ie:
config _EC_OSC & _WDTE_OFF & _PWRTE_ON & _MCLRE_OFF & _BODEN_OFF & _CP_OFF & _CPD_OFF

EC = external clock source on pin 2
WDTE = watchdog timer OFF
PWRTE = power up timer ON (a few ms delay at power up - allows other things to stabalize like external clock source)
MCLRE = reset OFF as a test, put it back ON if there is no difference after testing so you can reset the PIC
BODEN = Brown out detection OFF
CP, etc are code protection bits = OFF

 

[alec_t]

Unless I've missed something the TMR0 overflow interrupt is never enabled. I'm no programmer, but won't that be problematic for the delay routines?

 

[Nigel Goodwin]

Unless I've missed something the TMR0 overflow interrupt is never enabled. I'm no programmer, but won't that be problematic for the delay routines?

He's not using TMR0, he's using simple software loops.

 

[Diver300]

I would have thought that the fault would be more likely to be caused by the motor trying to take more current in the cold. That could also cause a bigger voltage drop in the MOSFET. The battery will have a higher impedance at lower temperature.

 

[AGCB]

I'm going to test it with the change made in the 1st part of post #22. I'll let you know when that test is complete. I'll be out of town till 2 December.
Thanks
Aaron

 

[AGCB]

The only change being made was to supply the PIC and oscillator separately with power from 3) AA batteries. It's now 10 days and it goes off at exactly 15:24:10 everyday. So the code and circuit were right on.

Thanks for all the ideas.
Aaron