dspic30f4013 and timers

[on7nh]

hello group
i have a timer issue and i know not what is happening

i have a external 20ms intterrupt form zerocross detection (mains)
i want to syncronize a second timer for thyristor firering...
but the timer count only to 8....

any sugestions?

Carl

[Ttelmah]

Comments:
117920000.......
Max supported frequency is 40MHz.

If you are firing a thyristor off a pin, you presumably want it 'X' uSec after the external interrupt triggers. As such, leave the timer interrupt disabled, till the external interrupt triggers. Then in the external interrupt, setup the timer, and enable it's interrupt, then in it's interrupt, disable the interrupt. This way the trigger will only occur when the external event is seen, avoiding accidental triggers. I can't see any reason to keep triggering the pin after this point. Once a thyristor is fired, it remains on, till the next zero crossing (except for things like GTO types)..

Then, your 'print' statement, which is meant to reflect the count of interrupts, won't. You send something like 26 characters each time. At 115200bps, 2.2mSec. Your output will be falling behind the counter all the time.
Your print, also uses '%u'. Look up the correct format for an int32.

Best Wishes

[on7nh]

@ Ttelmah

max frequency is 120Mhz !
internal osc *16 = 117920000

If you not set this value into the osc, all communication is wrong
40 mhz by external osc....then internal PLL !

I see also the 8 output pulses on my scope.... so i am very sure about this value...

[Ttelmah]

Yes, may internal frequency, but not maximum crystal. This is 40MHz.
The 120MHz, is not the 'XTAL_FREQUENCY', but the internal clock rate. The maximum allowed external clock, is 40MHz. This is the 'XTAL_FREQUENCY'. Your internal clock rate can then be a multiple of this, but to my mind your nomenclature is totally wrong.
You have both the fuses selected to run an external clock without PLL, and the internal RC. Wrong. You talk about a 40MHz external clock, but don't appear to actually be using this (FRC=use internal RC oscillator). 7.37*16.

Not a crystal at all....

This sort of error, is a sign of 'careless' thinking, and is not the way to get things to actually work right.

The timer will wrap at 449.8 times per second. 2.22mSec. At 50Hz, you get an interrupt from the edge detector, every 20mSec. Since the first pulse will occur at this interval after the interrupt, you would expect to see eight pulses, before the next external interrupt occurs, which sets 'tel' to zero, so the count will never go past 8.

Best Wishes

[FvM]

I think, the timer behaviour explanation is correct. Regarding conflicting fuses, CCS C is using the last setting in this case and simply ignores the wrong EC fuse. I also didn't understand what's been said about a 40 MHz crystal.

[on7nh]

thank for the input,
problem was in timer settings,

the correct way was :

[Ttelmah]

No, CCS doesn't just use the last fuse. It effectively 'OR's the patterns together. If the first one happens to be a '0' pattern, then it has no effect, but otherwise using two conflicting fuses results in a third pattern that may well not work at all.
On the oscillator, change the name you give your frequency to something like 'CLOCK_RATE'. It is _not_ a XTAL at all, and while you think it doesn't matter, it is a bit like calling a screwdriver a hammer. It doesn't matter while _you_ are the only person involved in the code, but if someone else later has to get involved with the code, they will get mislead. It is vital in code development, to 'call a spade a spade'....

Best Wishes

[FvM]