Help on PIC 16f877a crystal problem

I'm going to suggest you have another problem going on here.

First, are you using bypass capacitors at the power pins of your PIC? If not, place a 0.01 - 0.1uF cap from the power pin to the ground pin, as close to the PIC as practical.

Next, I've had problems with PICs when the reset input is floating. Tie it up (or down, whichever is appropriate) through a resistor.

Third, make sure your power supply circuit can supply the necessary current to drive all of your LEDs. 5 LEDs will draw 100mA, or more, depending on what kind you use. If your power circuit cannot supply that much, the Vcc at the PIC will drop and the chip will reset itself.

An issue with your code, generally when you are testing both cases of an input variable (pin_a0) you want to write your code so that one case has priority over the other, so that if both cases happen simultaneously - the logic doesn't hang up. For example:

if (pin_a0) = 0 then ....

else if (pin_a0 = 1) then ....

end

This way the "0" term has priority over the "1" term and will be tested first.

When you close your switch, the input (pin_a0) doesn't transition smoothly from 0 to 1 - the contacts bounce, so the data into the PIC looks like:

0000100010101011110101111101111111111111111, or similar. Since one part of your code is looking for (pin_a0), and the other part is looking for !(pin_a0) - that is, there is an extra inverter gate delay on the second term, it's possible for both terms to be true at the same instant when (pin_a0) changes.

I suggest that you de-bounce your switch input. If I recall correctly, this capability may be built into the PIC. If not, program a de-bouncing filter. (Change your code to require that the switch is open or closed for about 100mS before you change the state of the LEDs.)

Good luck!

Circuit layout is critical with any crystal oscillator and load capacitor values may need adjusting. Ceramic resonators are pretty reliable, accurate and don't need load capacitors. The micro' may need configuring so that it 'knows' what sort of ascillator you are using.
The commercial products I design nearly all use 8Mhz ceramic resonators. It is only when I need an accurate Real Time Clock that I bother with a crystal.

You realy need an oscilloscope if you are going to investigate the problem with the crystal oscillator. Try Googling 'pic crystal oscillator' ...?

Hope this helps...

Del

yes i use the 100uf on the power pin as shown in picture

and i think the supply have enought power for my LED to light up

but i still donno why the led still keep blinking and give me very fast speed pulse output.

As well as 100UF you need a small ceramic capacitor say 0.1uF or 0.01uF, this is because the 100uF electrolytic will not be very good at supressing high frequency noise.

Del

I haven't used the PIC16 family, but I have used the PIC18F family which has many similarities.

Firstly as drawn, the PIC is supplied with 5V with 4.7K in series but the reset taken direct to the 5V. I hope this is a drawing error as it won't light any LEDS if built this way.

Secondly a PIC should be quite fast enough for static logic applications using its internal RC oscillator. Configuring the clocks in these PICs can be confusing at first, there are a great many options. You need to check the clock divider ratio as this needs to be configured explicitly.

this site might be able to help you as well, he also has links to pic forums.

yes is a drawing mistake . i am very sorry about that . and thanks for the link ^^it help me a lot thank ^^

Here's an application note that came into my Google reader today:

http://ww1.microchip.com/downloads/en/AppNotes/93009A.pdf?from=rss