| View previous topic :: View next topic |
| Author |
Message |
gpsmikey
Joined: 16 Nov 2010
Posts: 588
Location: Kirkland, WA
|
|
 Posted: Fri Sep 28, 2012 3:32 pm |
 |
|
No - no extra holes for the missing cylinders, they built the block at the factory basically from the small V8 with the casting modified, but it still had the "V" angle of a V8. Engine had a few quirks, but in general good engine. I don't know how many of the V-6's out there were built the same way which was the basis of my original comment (I have the ability to run into these things it seems). Check out this link if you are curious about it:
http://en.wikipedia.org/wiki/Buick_V6_engine
Where they discuss the "198" being a derivative of the "215 Aluminum V8".
mikey
_________________
mikey
-- you can't have too many gadgets or too much disk space !
old engineering saying: 1+1 = 3 for sufficiently large values of 1 or small values of 3 |
|
 |
Mike Walne
Joined: 19 Feb 2004
Posts: 1785
Location: Boston Spa UK
|
|
| Posted: Sat Sep 29, 2012 1:56 am |
|
|
| Quote: | | You guys must be teachers. |
Not necessarily, but we've been around a long time and seen it all before (in different forms).
I don't think your reciprocal method will work as it stands.
1) If B0 is high you will go into the routine, and start timing. But, B0 may have been high for a while, and you'll have missed some of the period. You need to wait for an edge, before you start counting.
2) How long do you think it takes to go round the timing loop? (Hint it's NOT 25 cycles).
3) To get best resolution you need to be counting instruction cycles, not slower software loop cycles.
Mike
EDIT Something else to watch out for. If the signal suddenly stops, there is a risk that any reciprocal method will freeze at its last reading, your original counting method will show zero after one gating period. |
|
|
Ttelmah
Joined: 11 Mar 2010
Posts: 19539
|
|
| Posted: Sat Sep 29, 2012 7:51 am |
|
|
Partial only, but some basic ideas:
| Code: |
//several things will depend on your clock rate and chip
//Need your setup stuff here
int32 timer=0;
#word timer_msw=timer+2
#word timer_lsw=timer
//define a variable to access the 16bit parts of 'timer'
#INT_TIMER1
void timer_overflow(void) {
timer_msw++;
}
int1 wait_for_ccp() {
int16 overflow=0;
//wait for an edge to be detected on the CCP1 input
//returns true if an edge is found within about 0.5seconds
//false otherwise (stopped....).
do {
delay_us(10);
overflow++;
}while (!interrupt_active(INT_CCP1) && overflow<5000);
if (overflow<5000) return (TRUE);
return (FALSE);
}
void main(void) {
int32 first_timer;
int32 delta;
float revs=0.0;
//You will need your other configuration stuff here
setup_timer_1(T1_INTERNAL, T1_DIV_BY_1);
//set timer1 to count instructions
enable_interrupts(INT_TIMER1);
enable_interrupts(GLOBAL);
//We now have a 32bit timer potentially available
setup_ccp1(CCP_CAPTURE_FE); //start the CCP
//Use whichever edge is the best (rising or falling).
do {
//Now loop to take a time measurement from the CCP1 input
clear_interrupt(INT_CCP1);
revs=0.0;
if (wait_for_ccp()==TRUE) {
//Have seen an edge
timer_lsw=CCP_1;
first_timer=timer;
//Now have a 32bit timer value stored for the first edge
clear_interrupt(INT_CCP1);
//Now wait for the second edge
if(wait_for_ccp()==FALSE) {
revs=0.0;
}
else {
//If we get here, two successive edges have been seen
timer_lsw=CCP_1;
delta=timer-first_timer;
revs=INSTRUCT_CLOCK/delta;
//INSTRUCT_CLOCK needs to be #defined as the CPU instruction
//clocks/minute
//(for RPM output)
}
}
//here have a value 'revs', giving the measured RPM
printf("%6.1f",revs);
}
}
|
Now, several things missing (clock rate etc.), and it could be made significantly faster updating, if once _one_ measurement is made, you only look for the next edge (currently times between a pair of edges).
However has multiple things 'done':
1) If a clock edge is not seen in just over half a second, gives '0'0' as the output.
2) Will also not therefore hang when the clock stops.
3) Counts in processor cycles, giving as much resolution as possible.
4) Counts in 32bits, so will support a huge range of speeds and CPU clocks.
Have almost certainly typed something wrong. Just typed 'as is' to show how this can be done. Pulse input needs to be coming in on the CCP1 input.
Best Wishes |
|
|
nuno12345
Joined: 16 Jul 2008
Posts: 50
|
|
| Posted: Sat Sep 29, 2012 4:58 pm |
|
|
any way to do it without ccp1? I really want to test this on 12F675 before the 16f88 arrives  |
|
|
PCM programmer
Joined: 06 Sep 2003
Posts: 21708
|
|
|
nuno12345
Joined: 16 Jul 2008
Posts: 50
|
|
| Posted: Mon Oct 01, 2012 1:35 pm |
|
|
thank you pcm
I had already found that post here on the forums but I it works the same way get_timer does, at least I tried changing time and the frequency gets affected. |
|
|
PCM programmer
Joined: 06 Sep 2003
Posts: 21708
|
|
| Posted: Mon Oct 01, 2012 1:50 pm |
|
|
Then do some simple math on the result and adjust it for the different oscillator
frequency. |
|
|
|
FAQ 
Search
Register
Profile 
Log in to check your private messages 
Log in
