| View previous topic :: View next topic |
| Author |
Message |
hemnath
Joined: 03 Oct 2012
Posts: 228
Location: chennai
|
| auto_off function |
 Posted: Fri Mar 06, 2020 5:00 am |
 |
|
Hi,
PIC 18F2620, 1MHz internal oscillator, CCS version: V4.114
I want to auto-off the circuit when there is no change in PV for about 10 minutes. Due to some factors, PV will change slightly. So that we shall provide a limit like PV+x and PV-X. If the PV is within this range, a digital pin will switch off the circuit. Could you please help on the logic to work it out. I have been trying it. But couldn't find the solution.
Let's say, on 5th minute, if the PV changes largely or outside the limits as mentioned above. Time should start from 0. |
|
 |
Ttelmah
Joined: 11 Mar 2010
Posts: 19244
|
|
| Posted: Fri Mar 06, 2020 8:17 am |
|
|
| What on earth do you mean by 'PV'. We don't know.... |
|
|
temtronic
Joined: 01 Jul 2010
Posts: 9124
Location: Greensville,Ontario
|
|
| Posted: Fri Mar 06, 2020 9:05 am |
|
|
PV = Potential Voltage or Photovoltaic Voltage ??
The algorithm to control power is easy enough though... |
|
|
Ttelmah
Joined: 11 Mar 2010
Posts: 19244
|
|
| Posted: Fri Mar 06, 2020 11:51 am |
|
|
Photovoltaic voltage brings different problems to potential voltage.
Problem you are going to have is any photo voltaic source is going to
be very high impedance. You will need a buffer amplifier.
Simplest way would be to program two comparator outputs one from a
voltage just below the source, the second just above (use the comparator
reference module to set the limits), then take the two outputs and feed
them into a logic gate, or have hove code to set a pin when the voltage
is between the two limits.
However resolution will be quite low. For accuracy, you would have to
actually have the PIC running, and read the ADC and test the value
being between the limits. You could use the watchdog and sleep for
(say) 28mSec then make one reading, update the output and go back to
sleep. If a pin is to control this then wake for just a handful of machine
cycles test the pin and if it is still set, go straight back to sleep. |
|
|
dluu13
Joined: 28 Sep 2018
Posts: 395
Location: Toronto, ON
|
|
| Posted: Fri Mar 06, 2020 4:16 pm |
|
|
What circuit are you talking about? Do you mean you want to switch off your PIC, or you want to disable the power of the amplifier that you are using to measure the voltage?
If you are talking about the amplifier then there are opamps with a disable pin that you can pull to turn off the amplifier. |
|
|
Jerson
Joined: 31 Jul 2009
Posts: 122
Location: Bombay, India
|
|
| Posted: Fri Mar 06, 2020 6:52 pm |
|
|
| I guess the original poster meant process value |
|
|
temtronic
Joined: 01 Jul 2010
Posts: 9124
Location: Greensville,Ontario
|
|
| Posted: Fri Mar 06, 2020 7:08 pm |
|
|
We need more information.
What is 'PV' ?
Is PV an analog signal or a digital signal ?
This is CRITICAL information.
Another very important piece is HOW the 'time keeping' is being done.
Do you use an RTC device like the DS1307 or a SW based RTC ? Both require good battery backup and really a HW RTC is preferable the DS3231 is a very accurate RTC. |
|
|
hemnath
Joined: 03 Oct 2012
Posts: 228
Location: chennai
|
|
| Posted: Sun Mar 08, 2020 10:09 pm |
|
|
PV - Process Value
In my case, it varies from 0.000 to 10.000.
Assume it is a variable as Y. Don't confuse with PV. |
|
|
Ttelmah
Joined: 11 Mar 2010
Posts: 19244
|
|
| Posted: Mon Mar 09, 2020 1:03 am |
|
|
Some comment:
Having a number like 0 to 10.0 floating point, is a sure sign of 'bad design'.
Floating point numbers are never accurate, involve a huge amount of
processor work, are bulky & slow.
Much better to have an integer value like 0 to 10000.
Now no matter how this is generated there will be 'errors'. You are never
going to get 'no change'. Simple noise implies that values will fluctuate.
You need to be using a 'margin', and allowing small changes to be treated
as insignificant.
Now, how to suspend a system, will be down to what hardware you actually
have. Simplest something like a RTC chip, and program an alarm for a time
interval. However don't be afraid of simply having the system wake for a
few machine cycles at a much shorter interval. A one per second 'tick',
and a system that wakes, does some tests, updates a counter and goes
back to sleep, costs almost nothing in power. |
|
|
hemnath
Joined: 03 Oct 2012
Posts: 228
Location: chennai
|
|
| Posted: Mon Mar 09, 2020 2:53 am |
|
|
Ok. I'll come in your way.
If the value is integer (lets say the value is 50) is constant for 10 minutes, I'll send a logic high signal using microcontroller pin. I don't have RTC chip in my hardware. Using internal timer, I have to use it.
I have used the internal timer section for 1 second delay.
| Code: | #INT_rtcc // This function is called every time
void clock_isr()
{
// the RTCC (timer0) overflows (255 - > 0) .
// FOR this program this is apx 76 times
if (--INT_count == 0)
{
// per second.
++seconds;
INT_count = INTS_PER_SECOND;
}
}
|
How to build the algorithm/logic to check that the value is not changing using this timer ? |
|
|
Ttelmah
Joined: 11 Mar 2010
Posts: 19244
|
|
| Posted: Mon Mar 09, 2020 4:11 am |
|
|
Even with an integer, you will see change. The key is to set limits for it:
| Code: |
int1 change_seen=FALSE;
#define LEVEL 4 //set this to determine the movement accepted
#INT_rtcc // This function is called every time
void clock_isr()
{
static unsigned int16 old_pv;
int16 diff;
if (old_pv>pv)
diff=old_pv-pv;
else
diff=pv-old_pv;
if (diff>LEVEL)
change_seen=TRUE;
else
change_seen=FALSE;
old_pv=PV;
// the RTCC (timer0) overflows (255 - > 0) .
// FOR this program this is apx 76 times
if (--INT_count == 0)
{
// per second.
++seconds;
INT_count = INTS_PER_SECOND;
}
}
|
|
|
|
temtronic
Joined: 01 Jul 2010
Posts: 9124
Location: Greensville,Ontario
|
|
| Posted: Mon Mar 09, 2020 5:18 am |
|
|
re: timer based RTC
the following link ...
http://www.ccsinfo.com/forum/viewtopic.php?t=26177
...shows how to implement an accurate RTC using an internal timer.
Your version has a comment ...'apx 76 times'....
While it might be OK for a few seconds, 5-6-10 minutes could become inaccurate.
Also if NOT using an external xtal/caps for the PIC clock, the internal oscillator will not be accurate. Just something to be aware of.
Jay |
|
|
Ttelmah
Joined: 11 Mar 2010
Posts: 19244
|
|
| Posted: Mon Mar 09, 2020 8:30 am |
|
|
and as a comment for looking for change over a time:
| Code: |
int1 change_seen=FALSE;
#define LEVEL 4 //set this to determine the movement accepted
#define LIMIT_TIME 4560 //counts - here using 60*76
#INT_rtcc // This function is called every time
void clock_isr()
{
static unsigned int16 old_pv;
static unsigned int16 max_pv, min_pv;
static int16 count;
int16 diff;
if (old_pv>pv)
diff=old_pv-pv;
else
diff=pv-old_pv;
if (diff>LEVEL)
{
change_seen=TRUE;
count=0;
}
else
{
count++
if (count>LIMIT_TIME)
change_seen=FALSE;
}
old_pv=PV;
// the RTCC (timer0) overflows (255 - > 0) .
// FOR this program this is apx 76 times
if (--INT_count == 0)
{
// per second.
++seconds;
INT_count = INTS_PER_SECOND;
}
}
|
Here 'change_seen' will go FALSE, when the PV signal stays the same
within 'LEVEL' for 'LIMIT_TIME' counts. |
|
|
|
FAQ 
Search
Register
Profile 
Log in to check your private messages 
Log in
