| View previous topic :: View next topic |
| Author |
Message |
benoitstjean
Joined: 30 Oct 2007
Posts: 566
Location: Ottawa, Ontario, Canada
|
| PIC18F4620 INT_RB maximum speed |
 Posted: Wed Aug 15, 2012 6:54 am |
 |
|
Hello, I guess I will try this by trial and error when I get a chance but hopefully in the mean time someone might be able to answer.... I realize that many things can impact the speed of interrupts BUT considering the following:
In my main, I have a simple infinite while loop:
I also have the #INT_RB interrupt routine simply do the following:
| Code: |
#int_RB
void RB_isr(void)
{
// Read PORT B
uiPortB = input_b();
(sPinB4 below is a one bit SHORT)
// Read the entire port B, isolate PIN_B4 with the bit mask,
// shift all bits over by 4 to have sPinB4 be the actual PIN_B4's
// state
sPinB4 = ( uiPortB & 0x10 ) >> 4;
// Ouput sPinB4's value on PIN_D4
output_bit( PIN_D4, sPinB4 );
} |
Currently, on PIN B4, I give it a 50% duty cycle square wave of 8kHz. This results in having the same 8kHz onto PIN D4, but with a delay.
So my question is: what's the maximum frequency I could put with this simple program onto PIN B4?
I will do some more tests since I have a partial program running. I am currently testing with 8kHz and it's working. The delay between B4's low-to-high transition and D4's L-t-H transition is exactly 15 us.
Has anyone tested this with higher frequencies? I will share my findings here once I manage to figure it out but in the mean time, if someone has answers, please let me know.
I'm using a PIC18F4620 and my crystal is 16.384Mhz (for specific reasons not explained here but I need that crystal).
CCS Compiler version 4.056.
Thanks! |
|
 |
drolleman
Joined: 03 Feb 2011
Posts: 116
|
|
| Posted: Wed Aug 15, 2012 9:54 am |
|
|
| Test it. hook a scope to an unused pin turn in on at the start of the isr amd then off and the end of the isr and you get your answer. |
|
|
benoitstjean
Joined: 30 Oct 2007
Posts: 566
Location: Ottawa, Ontario, Canada
|
|
| Posted: Wed Aug 15, 2012 10:05 am |
|
|
Thanks but if you read closely, that's what I'm doing currently with 8kHz. I read the port and make PIN D4 the same state as my interrupt-on-change PIN B4.
On my scope, I have probe A to PIN B4 and probe B to pin D4. If I measure the delay between the rise time of both signals (it's an 8kHz 50% duty), there's 15us.
What I want to know is how fast is the fastest I will be able to do this.... I am trying with 64kHz and I don't think my PIC can handle interrupts that fast as PIN D4 is totally NOT the same as PIN B4 in this case (looking at my scope).
The problem is I don't have a frequency generator so it's a bit long to test with the PWM at different frequencies.
Again, I will most likely get my answer soon by trying out different PWM frequencies but as mentioned, I was just curious to see if someone already had the answer before I did all the work.
So, so far, 64kHz seems too fast. |
|
|
benoitstjean
Joined: 30 Oct 2007
Posts: 566
Location: Ottawa, Ontario, Canada
|
|
| Posted: Wed Aug 15, 2012 10:28 am |
|
|
So I think I've figured-it out. Using a 16.384MHz crystal, if I set my PWM to these settings:
| Code: |
setup_timer_2( T2_DIV_BY_4, 49, 1 );
set_pwm1_duty( 25 );
|
It gives me a 20.48 kHz 50% duty wave on my PWM which I feed in interrupt PIN B4. The result is output PIN D4 (since I read pin B4 and apply its state to output PIN D4).
I say 20.48kHz because that's a perfect round number using the given crystal frequency. It's probably a bit more but afterwards, the delay between the rise-times of the interrupt versus the output pin will start overlapping.
Can someone confirm that this would make sense? |
|
|
PCM programmer
Joined: 06 Sep 2003
Posts: 21708
|
|
|
|
FAQ 
Search
Register
Profile 
Log in to check your private messages 
Log in
