![]() |
quick question: TIMERS
The code below works but i need just a few more loops in the code in order to get it to go exactly where i need it to any one know how i can make it start a new count after it gets past 254 so it doesnt loop, and i can add a few more steps to it, or does some one have a better way of doing this ? I got this way to work it just needs to run for a little bit longer.
Code:
void User_Autonomous_Code(void) |
Re: quick question: TIMERS
you could just do an if of when it goes above the time.
if(timer>40) {} else if(timer>80) {} |
Re: quick question: TIMERS
Your code is begging to become a state machine. It is a codeing techinque that uses a switch statement and where each case is an independant state. This works well because of the 26.3 ms looping process. The benefit of state machines is that it can be written so that some states are re-entrant where they look for some event like time or distance to trigger the program into the next state.
switch (state) { case 1 : // start motors t = time + SomeDiration break; case 2 : if (time < t) // This is re-entrant. state = 2; else state = 3; break; case 3 : // stop motors and start next manuever... ... Get the picture? It's not hard. It does require some thought. But this is how most very complex programs start out. There are other ways of doing this with pointers and trillions of function - but this is just so simple and readable. If you are at the Great Lakes Competition look me up and I'll show you more. |
Re: quick question: TIMERS
huh ? dun quite get it, too tired maybe ...
|
Re: quick question: TIMERS
A few things:
At the most basic level, you don't need to worry about the int looping over. Ints in C are two byte variables; an unsigned one will store up to 32000+. If you need more than that, you can use long, which is a four byte variable. Second, you don't need two seperate if statements when you are setting pwm01 and pwm02--they do the same thing, so put all the code you need in one! You'll save processing power and increase readability. State machines can be a little confusing at first, but they will make your code much more readable in the end, and much faster. A state machine uses the switch command, which will look at one variable, and jump to the "case" that the variable number points to. So you would have an int variable to count another int variable that would be the state. int count; int state = 0; void InitAuto(void){ count = 0; //any other variables that you need to reset } void User_Auto_Code(void){ switch(state){ //look at the state variable for code to jump to case 0: //do some stuff if(count > 100){ state++; //move to state 1 InitAuto(); } break; case 1: //do some other stuff if(count > 50){ state++; //move to state 2 InitAuto(); } break; etc etc etc.... }//close switch count++; As you can see, its really easy to make readable code this way, and you know exactly how many program counts are in each state without having to do any math =). This is also a much lighter load on your processor; your program will execute a lot faster and be cleaner in memory. A couple symantic things: the break; statement is needed at the end of every case, to tell the compiler that you are done with the case. Just set up your cases in order, and tell your motors what to do =). If you can implement real timers, it will make the program that much better (254 uses a millisecond timer), but you should be fine with the program counts. Good luck! |
Re: quick question: TIMERS
Quote:
Code:
Static state variable starting at 0Hope that clarifies it some. |
Re: quick question: TIMERS
Quote:
case 1 is to happen for 2 seconds and case 2 is supposed to happen for 3 seconds? pseudo of what im talking about : case 1 for 2 seconds pwm01= 200 pwm02= 200 pwm03= 150 pwm04= 200 then go to case 2 case 2 for 3 seconds pwm01= 200 pwm02= 100 then go to case 3 case 3 pwm01= 200 pwm02= 200 etc . what keeps it with in time constraints and what cuases it to go to the next case statement. |
Re: quick question: TIMERS
That's all up in my post...
Make a "StateInit" function that runs everytime you move to a new state, that at the very least resets your count to 0. Put a count++; statement after the switch. Then, inside each case, say if(count > someNumber){ StateInit(); state++; } That will move it to the next case when the count is above the number you want. Since you don't have a real timer you'll have to do the math yourself for how long to stay in each one but it shouldn't be that bad. |
Re: quick question: TIMERS
Quote:
|
Re: quick question: TIMERS
Again...read my post...
If you use an int for your counter variable, you will have more than enough space. An int in C is a two byte variable, meaning the signed version can store from -32000 to 32000. What you're thinking about is a char, a one byte variable. That will turn over at 254 (or 128 if it is signed). Have you tested your code yet? |
Re: quick question: TIMERS
Quote:
so what if i use my orrigianl code and just change char to int then i wont have to eorry about the overflow problem right ? |
Re: quick question: TIMERS
Quote:
You could have a variable that gets increment every 26.2ms. If this functions is getting called that often, it could be static to it. If not, it could be a global variable that is incremented in something that is. Then, for the tests, you would check if the variable was greater than, say, 78. This would make it run for approzimately 2 second ((1000ms/26.2ms)*x seconds). Or, if you need more accuracy than that, and that's not bad, you can have an external clock attched to pin 1/2 and have an interrunpt that increments it. |
Re: quick question: TIMERS
Quote:
|
Re: quick question: TIMERS
Quote:
|
Re: quick question: TIMERS
5: 0-4. 0 is used by FRC_library.lib and is slowest of all, 1&3 are faster than 2&4.
Range: (Prescaler Range|Postscaler Range) 0: 1:2-1:256|1:1 1: 1:1-1:8|1:1 2: 1:1-1:16|1:1-1:16 3: 1:1-1:8|1:1 4: 1:1-1:16|1:1-1:16 The prescalers are powers of 2 (*2), and postscalers are integers (+1). The base frequincies of all the timers is the same: w/o scaling they are 10MHz clocks and a 100ns tick. This all out of Kevin's interupt.c file. WELL commented |
Re: quick question: TIMERS
Yes, it does have timers (4), but I couldn't figure out how to use them, and didn't feel the need to try much. If someone knows how, those would also work.
|
Re: quick question: TIMERS
Quote:
-Kevin |
Re: quick question: TIMERS
I have had problems with interupts.c and timers. I use Timer 2 (and/or 1) to increment a variable and if that variable overflows, it increments another one. The problem is that I can't ever seem to get either to change. Help?
|
Re: quick question: TIMERS
Quote:
Here's a sample initialization of Timer 4 to cause a 4ms interrupt. Code:
OpenTimer4(TIMER_INT_ON &Code:
unsigned long Clockms; // 1 millisecond clock |
Re: quick question: TIMERS
Quote:
|
Re: quick question: TIMERS
now if i was to use the timers that kevin has on his site inorder to do a dedreckoning autonomus what would it look like ? I begin to have a little more trouble when the interrupts are thrown in there i looked through the actuall clock.c file and the user_routines_fast.c file how would i specify time lengths for an action ? Im not too sure how the interupts work i read the first white paper on them but my understanding of them are still kinda wavy.
|
Re: quick question: TIMERS
Quote:
e.g., Code:
if (Clock < 10) // Clock is in tenths of seconds[edit] Corrected the ">" , typed too freely. Thanks Jamie! |
Re: quick question: TIMERS
The timer (1-4, actually) is initialized, and I have checked the enabled bit. The overflow variable (a char) is to slow it down. I send the long both over the terminal and through the dashboard. Both always say 0. The code looks like it should work, but doesn't. This is my ISR:
Code:
extern char TICK; |
Re: quick question: TIMERS
make sure its not only enabled, but i think you have to start it too
|
Re: quick question: TIMERS
i got kevins interrupt code and i looked through it, how do i call one of the timers to start counting ? then how would i implement it into autonomus ?
|
Re: quick question: TIMERS
Call the start method in the User_Initialization function (Timer2Start() or Timer4Start() or whatever it is). Then, use the state machines that we told you about in the first page of this thread. In the StateInit() function, set a new variable like StartTime or something to keep track of where the timer was when you moved into the next state. Then just subtract the start from the actual, and when that value is larger than the one you want (how long you want to stay in the state), move to the next state and call the Init function.
|
Re: quick question: TIMERS
Initialize_Timer_X, And you have to modify the line that reads: TxCONbits.TMRxON = 0; to TxCONbits.TMRxON = 1;, else the timer won't go.
|
Re: quick question: TIMERS
Quote:
Your code looks file of course. Have you tried displaying TICKS on the dashboard or terminal to see if it is also not getting incremented? I have to run now, but I'll be back later and try to duplicate what you did in a test program. |
Re: quick question: TIMERS
Quote:
Code:
extern long TIME;The User_Routines_Fast.c Interuptor (Actually executed by proc) is this, just to check: [code]#pragma code InterruptVectorLow = LOW_INT_VECTOR void InterruptVectorLow (void) { _asm goto InterruptHandlerLow // jump to InterruptHandlerLow() _endasm } #pragma code //... #pragma interruptlow InterruptHandlerLow save=PROD,section("MATH_DATA"),section(".tmpdata") void InterruptHandlerLow() { //... }[/check] I didn't edit this (to my knowledge :yikes:), so I can't imagine that would be it. The code is compiling fine, no errors or warnings. |
Re: quick question: TIMERS
Quote:
|
Re: quick question: TIMERS
I just relized something. In the initialize routines for the timers, the first 2 lines are address setters (I think). Most of mine are 0x00! Does this make a difference?
|
Re: quick question: TIMERS
Quote:
Believe I found the problem. TICKS is declared as char (-127 to 127) You are checking for 255 as an overflow. TIME will never increment. I grabbed Kevins interrupt code and added your code to the interrupt handler. Only difference is I just declared your globals as static just within interrupts.c for my ease of duplicating your results. TICKS and AUTOCLOCK increment as you wanted. In Initialize_Timer_1 I changed: Code:
PIE1bits.TMR1IE = 1; |
Re: quick question: TIMERS
before i enable one of those timers, is there anythign i should know about which one would work best are there significant differences for them, i noticed you can change the ammount of time each tick increses the clock by.
once i enable one im still confused as far as the state switches and the case statements, idon't understand how the whole thing works is there any documentation or sample code i can look through, im really stumped and i would like to get a timer working. |
Re: quick question: TIMERS
OK...I really want to help you dude but it seems like you don't understand what we're trying to say each time, or I can't explain it well enough...why don't you IM me (sn: TenintheCrunch)? I think I'll be able to help you more talking directly :)
Cheers, |
Re: quick question: TIMERS
I changed a compiler option to assume Char is unsigned. I only use a signed char one place, -1 - 1, and it's declared as signed. I also sent TICK to dashboard, and it didn't work. I didn't have IE set to 1, but I didn't think it would make a difference (it would just be slower, right?)
|
Re: quick question: TIMERS
Where's the option for making chars unsigned by default? Does that affect ints and longs as well?
|
Re: quick question: TIMERS
Quote:
Project->Build Options...->Project under tab "MPLAB C18" There is a quick radio button for treating "char" as unsigned and one for forcing all calculations to be performed at least as int rather than the smallest type included in the calculation. You can also enter your own flags by clicking on "Use Alternate Settings" and typing into the box that becomes enabled. You'll find the available compiler flags documented in the MPLAB C18 C Compiler User's Guide that came on the MPLAB installation CD, Appendix C. e.g., -k is the flag for treating char as unsigned. When you click on the radio button you'll see "-k" appear in the "Inherit global settings" window. |
Re: quick question: TIMERS
Quote:
INTCON3bits.INT2IE = 1; Yes, this is required. It's your on/off switch for individual interrupts. It enables the interrupt which is disabled by default. Otherwise, nothing will happen. |
Re: quick question: TIMERS
Banner interupts work, but not timers.
|
Re: quick question: TIMERS
Quote:
Code:
PIE1bits.TMR1IE = 1; |
Re: quick question: TIMERS
i got my ded reckoning worked out and i have a switch wired up to switch between the two modes how would i do that using case statements?
it should be something like case1 if (rc_dig_in_10=1) {auton 1} case2 else(rc_dig_in10=0) {auton 2} i read about them but how would they be used to work with a digital input from the rc |
Re: quick question: TIMERS
Quote:
Code:
if (rc_dig_in_10 == 1) //or just "if (rc_dig_in_10)" |
Re: quick question: TIMERS
Quote:
|
Re: quick question: TIMERS
Quote:
If you try it and it doesn't behave as you expect post the error here and we'll help you debug any problems. |
Re: quick question: TIMERS
Quote:
When you enter autonomous mode, all inputs are reset to neutral, so you have to capture the value from rc_dig_in_10 before autonomous turns on. Create a prevRcDigIn10 variable, and set that equal to the current one at the end of every loop (right before putdata()), then when you move into autonomous, look at that previous varialbe, not the real one (the "real" one will always be zero). |
Re: quick question: TIMERS
Huh? You are talking about the IO on the RC, right? In auto mode, the OI transmits inputs as neutural.
|
Re: quick question: TIMERS
so it would be like:
#define digin10old then later on .... digin10old = (rc_dig_in10) putdata() |
Re: quick question: TIMERS
Quote:
Contrary to an earlier post, the RC inputs such as rc_dig_in10 do work during autonomous mode, so there's no reason to do it this way. The advice to save the values for later use applies only to OI inputs (knobs and switches, mostly) that you want to use to configure autonomous operation. Also, variables in c aren't declared using #define statements. The proper syntax is not trivial, but it is not important here, since you don't need to use a variable for this purpose. |
Re: quick question: TIMERS
so it would be done how then ?
|
Re: quick question: TIMERS
Quote:
|
Re: quick question: TIMERS
Code:
if (rc_dig_in18=1) { |
Re: quick question: TIMERS
Quote:
Your code seems ok at first glance...why don't you put this statement in your Process_data function: printf("%d\n", (int)rc_dig_in18); Check the console printout while your program cable is plugged in. Make sure that it changes properly with the switch on your robot. Too improve efficiency: use else if statements. For example: if(ticks < 15) blah; else if (ticks < 35) blah; That will handle your decisions faster... Sorry again if I confused you eariler. :) |
Re: quick question: TIMERS
Quote:
|
Re: quick question: TIMERS
you guys are all really smart...i need to learn this stuff, or else i don't get to be on my team next year
|
Re: quick question: TIMERS
Quote:
What your first line should say instead is this: Code:
if (rc_dig_in18==1) {Code:
if (rc_dig_in18) { |
Re: quick question: TIMERS
Quote:
|
Re: quick question: TIMERS
Quote:
|
Re: quick question: TIMERS
Quote:
Code:
// somehow get a number in the "auto" variable |
Re: quick question: TIMERS
Quote:
Code:
On Off |
| All times are GMT -5. The time now is 12:47. |
Powered by vBulletin® Version 3.6.4
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.
Copyright © Chief Delphi