To LUGNET HomepageTo LUGNET News HomepageTo LUGNET Guide Homepage
 Help on Searching
 
Post new message to lugnet.robotics.handyboardOpen lugnet.robotics.handyboard in your NNTP NewsreaderTo LUGNET News Traffic PageSign In (Members)
 Robotics / Handy Board / 7017
7016  |  7018
Subject: 
Re: Time available during 1 kHz system interrupt?
Newsgroups: 
lugnet.robotics.handyboard
Date: 
Tue, 2 Nov 1999 17:15:37 GMT
Original-From: 
Christopher Prosser <cprosser@SPAMLESSacm.org>
Viewed: 
1002 times
  
The controls are spread across a couple of files, mainly in libhb.c we get

/*
  pulse width modulation control:
  if off, all motors run at full speed
  if on,  speed bytes are used to determine motor speed
*/
void system_pwm_on() {bit_set(0x39, 0b00000100);}
void system_pwm_off() {bit_clear(0x39, 0b00000100);}

/*
  printing to the LCD
  WARNING:  printf's will wedge once the print buffer becomes full
    if system printing is disabled  */
void system_print_on() {bit_set(0x39, 0b00000001);}
void system_print_off() {bit_clear(0x39, 0b00000001);}


in r22_ir.c
void ir_transmit_on();
void ir_transmit_off();
void ir_receive_off();
void ir_receive_on();

encoders.c enables and disables encodes.

Note: I'm not sure if these completely disable the code at the asm level.

In regards to what happens if you exceed 1ms, the pcode.asm has the
following comment:
System interrupt performs the following tasks:
*
*     0.  sets up for next interrupt
*     1.  increment system time
*     2.  decrement "process_ticks".  If zero, pokes
*  BRN (branch never) into pcode_run loop, so that
*  current process exits.
*     3.  deals with LCD print.
*     4.  does PWM stuff.
*     5.  does shaft encoder stuff.

A quick read of pcode.asm dictates that memory location 0x39 is a set of bit
flags for all the functions you can control. bit_clear(0x39, 0x00) will turn
everything off. Here are the bit definitions

PRINTBUFFER EQU %00000001
IRDECODE EQU %00000010
PWM  EQU %00000100
SHAFTENCODER EQU %00001000


-chris

----- Original Message -----
From: Bruce Moore <bamoore@bloomington.in.us>
To: <handyboard@media.mit.edu>
Sent: Tuesday, November 02, 1999 4:47 AM
Subject: Re: Time available during 1 kHz system interrupt?


Interesting discussion but it raises a few more questions.

Are there any (bad) side effects when user code that is inserted in the • ISR
takes longer to execute than the 1 msec interrupt cycle.  From experience
it appears that the ISR runs on the next interrupt following the • completion
of the current cycle.  The only thing I noticed is that the PWM routines
don't always calculate their next transition correctly but i figure that
there are plenty of other things that are going on that I'm not watching
(and that may eventually bite me).

Chris:  How do you turn off the things like IR and quad shaft decoding to
get rid of the little overhead.


----------
From: Christopher Prosser <cprosser@acm.org>
To: Greg Starr <starr@unm.edu>; handyboard@media.mit.edu
Subject: Re: Time available during 1 kHz system interrupt?
Date: Monday, November 01, 1999 8:48 PM

Hi Greg,
  It all depends. Various features of the HB can be turned on and off to
change the amount of work done in the main ISR. By default the code uses
300ms to execute. The file lib_hb.c provides the following:

/********************************/
/*** System Interrupt Control ***/
/********************************/

/*

    These functions allow you to turn on and off various features
    controlled by the system interrupt routines.  The more features
    you turn off, the faster your code will run.

    On reset, the features have the following state:
       pulse width modulation ON
       infrared decoding      ON
       LCD printing           ON
       quad shaft decoding    OFF
       IR tranmission       OFF

    This uses approx. 30% of total CPU time.

    Approximate benchmarks:

x    Feature  % of CPU
    -------  --------
    PWM       3
    IR decode     11
    LCD printing (active)   8
    LCD printing (inactive) 1
    quad shaft decode     5
    IR transmission     1

*/



-chris prosser


----- Original Message -----
From: Greg Starr <starr@unm.edu>
To: <handyboard@media.mit.edu>
Sent: Monday, November 01, 1999 7:19 AM
Subject: Time available during 1 kHz system interrupt?


Does anyone know the approximate amount of time available for user
assembler programs which "add on" to the existing 1 kHz HB interrupt?

Another way of stating that is...how much time is used by the 1kHz HB • ISR
code?

--greg

*****************************************************************
* Greg Starr, Professor                                         *
* Department of Mechanical Engineering        ph (505) 277-6298 *
* University of New Mexico, Albuquerque NM 87131   FAX 277-1571 *
* email- starr@unm.edu  WWW- http://www.me.unm.edu/~starr/      *
*****************************************************************






Message is in Reply To:
  Re: Time available during 1 kHz system interrupt?
 
Interesting discussion but it raises a few more questions. Are there any (bad) side effects when user code that is inserted in the ISR takes longer to execute than the 1 msec interrupt cycle. From experience it appears that the ISR runs on the next (...) (25 years ago, 2-Nov-99, to lugnet.robotics.handyboard)

3 Messages in This Thread:


Entire Thread on One Page:
Nested:  All | Brief | Compact | Dots
Linear:  All | Brief | Compact
    

Custom Search

©2005 LUGNET. All rights reserved. - hosted by steinbruch.info GbR