Instant Interrupts - Revisited

[Darrel Taylor]

Well, without the code, we're reduced to playing 20 questions.

Please post at least the Interrupt Definitions, handlers (other than Elapsed), and ANY and ALL lines that modify the following registers.

T1CON
INTCON
INTCON2
INTCON3

PIE1
PIE2
IPR1
IPR2

<br>

[sougata]

Hi,

DT's Elapsed Timer takes care of setting T1CON right in the "StartTimer" . If by any means T1CON oscillator is turned on then PORTC.1 becomes an input regardless of the TRIS settings. Still it does not explain why and how the PWM starts when Timer1 is started.

What happens when your PWM stops the pin in set low/high ?
I hope you are quite aware that if duty cycle is higher than period then the pin never gets cleared.

Mr.E where is your crystal ball ??!!??

[Darrel Taylor]

If by any means T1CON oscillator is turned on then PORTC.1 becomes an input regardless of the TRIS settings.

That's a good possibility.

Still it does not explain why and how the PWM starts when Timer1 is started.

So True!

> Mr.E where is your crystal ball ??!!??
Ummm, being that he hasn't logged in since 15th June 2008
I'd say he's in Jail.

Can't imagine mister-e not checking in, if he could.
<br>

[alanmcf]

Hmmm... Sougata's point about starting timer1 sounds like it might be the culprit, at least as far as RC1 going to input. PWM2 drives a fan via a MOSFET, so if RC1 goes to input, the pullup I have on the MOSFET would drive the fan to full speed, which is exactly what's happening. I haven't yet found in the data sheet why turning timer1 on does that to RC1.

I have attached the code relevant to Darrel's question. The INCLUDE files that aren't attached don't modify the registers Darrel was talking about. I have very little experience with using the timers or using interrupts-- is it possible to just modify the Elapsed Timer routine to use timer 0 or timer3?

Thanks--

--Alan

[Darrel Taylor]

Hmmm... Sougata's point about starting timer1 sounds like it might be the culprit,

It is!

The line T1CON = %00001101 '<-- settings for 16 MHz turns on the Timer1 Oscillator, which disables RC1's digital I/O functions.

Since you are using the Elapsed Timer, there's no need to set T1CON. The Include file does that for you.
<br>

[Ioannis]

Can't imagine mister-e not checking in, if he could.
<br>

In fact, I had some tel. discussion 1 and 1/2 month ago with Steve, but when I tried to find him again, he disappeared. No telephones (only tel. answering machine) no e-mails. If anyone knows about Steve please inform. He had some misfortunes in the past as we can recall, just hope all is OK now, lying in some beach.

Ioannis

[alanmcf]

Hi Darrel--

Are you saying that your elapsed timer routine uses Timer1 without actually turning it on, thereby avoiding the portc.1 problem? I will try this immediately!

Thanks--

--Alan

[Darrel Taylor]

Yes, It uses Timer1. But it doesn't turn on the External Oscillator.

Timer1 has a feature that lets you hook up a 32768hz crystal on the T1OSO(RC0) and T1OSI(RC1) pins. This oscillator will continue running in sleep mode, so it can be used to keep time like a clock.

But the way you are using the Timer, it's using the signal from the main System Oscillator, so there's no need to turn the ext. osc on.

For the Elapsed timer, you can use the StartTimer and StopTimer subroutines.

hth,

[sougata]

Skimask pointed this out

http://www.picbasic.co.uk/forum/show...0058#post60058


BTW I would encourage everybody to have a look at DT's codes and try to figure out the amount of hard work and pure AI (I beleive he is an android built by microchip using surplus production ) he has put in.

Meant to make our life easy and enlighten us, the mere mortals on PIC and PBP.