Switch debouncing

[TaiChipY]

Hardware: 18F452, PICDEM2, ICD2

I have one button with two functions and its working thru if ,switch case plan...

Now I need to "read" faster qlicks on the same button so i can add another function ( basicly, something like mouse )...
This is the example...

*********
first slow click ---> one option
second slow click --> second option
third DOUBLE click ---> third option

***********
Double click function should be enabled in ANY time by quick double pressing.

How to make this?

[Guest]

You'll need a state machine by implementing switch/case statements:

Here are your states...

STATE_IDLE // No button pressed

STATE_BUTTON_PRESSED_ONCE
STATE_BUTTON_INITIALLY_RELEASED
STATE_BUTTON_PRESSED_TWICE
STATE_WAIT_FOR_RELEASE

You'll start off in IDLE state. If you hit the button, then go to state BUTTON_PRESSED_ONCE. Once you release the button, you will want to start a timer (e.g. timer1), and jump to next state STATE_BUTTON_INITIALLY_RELEASED. While in state BUTTON_INITIALLY_RELEASED, you will poll to see if your timer exceeds, say 1/2 second, and also poll to see if the button is hit again before the 1/2 second elapses. If the 1/2 second elapses without a second button press, then this is one function. If you hit the button before the 1/2 second timeout, then you go to STATE_BUTTON_PRESSED_TWICE. This is is your other function. In STATE_BUTTON_PRESSED_TWICE, you simply wait until the button is released, then this is your third function.

Your state machine routine should return an int that will tell the main function whether or not any buttons are pressed. Thus the return values can take on #defines such as:
BUTTON_NONE
BUTTON_CLICK_1X
BUTTON_DOUBLECLICK

So the state machine should just distinguish between a single click or a quick double click. It is up to your other software to do its thing with each successive single-click.

[libor]

Sorry for the off-topic...
are you aware of the circumstance that there is a patent Microsoft holds for double clicking? So I think, you cannot implement double-clicking in your product without paying a royalty to MS.
Read about it here.

Don't you get me wrong, I disagree with patenting such trivial things like this that was already used by the Neanthertal-man to knock on the cave-door two times when he wanted his wife to come and once when he wanted his brother's help.
Don't you think, guys accross the sea, it is time to tidy up your patenting rules, they are ridiculous.

[Sherpa Doug]

Instead of calling it "double click" you could call it a Morse code "i". In Morse code e = dot, i = dot dot, s = dot dot dot, and h = dot dot dot dot. Morse code is long since out of patent.

Sherpa Doug

[Humberto]

libor wrote:

[dave]

This means then that any instrument that has employed double clicking to enable software functions will have to be withdrawn, wonder how many airlines will be sued by MS.......
Dave

[Douglas Kennedy]

The Queen of England had to get Bill Gates permission to double click him on the shoulder as part of dubbing him Sir "something or other" when getting his Knighthood.There is a tension between patents encouraging innovation and thwarting innovation. Here in the US you could get a patent for almost anything. That's where a big company will protect its turf. The big guy patents just about anything and the small guy can't afford the huge legal expense of challenging it. Conversly if Microsoft didn't patent double clicking some small guy might do it and shake Microsoft down. Only if you invent something that is no use to anybody are you sure of a smooth ride.

[Mark]

[TaiChipY]

Now that's the spirit :-))). Like Bi G. has sad: MS has no monopol on the market we just improve old stuff :--) ... On some strange way :-)

*********
After some thinking and after examining the help files ( i got one interesting example how to solve my problem: thru rtcc. Maybe the case was not so good idea ( becuse of "holding in loop part")...

So..
After INT_RTCC stuff i have started with POLL functions in which im using
if's for poll, start_debounce, debounce, end_debounce..

q.)
Im poll state i have this situation:
------------------------------------
if (Gstate == poll)
{
cButton_state = cButton; // confirmation
iButton_state = iButton; // Infos
rButton_state = rButton; // reset

clock = 0;
read_new_state();
clock = 0;
return;
}

void read_new_state() // --- used for defining new changes
{
if (iButton_state != iButtonOLD)
{
iButton_change = TRUE;
}
else
iButton_change = FALSE;

if (cButton_state != cButtonOLD)
{
cButton_change = TRUE;
}
else
cButton_change = FALSE;
....
}
----------------------------------------------
1. press -- activate iButton -- display welcome
2. press -- activate cButton -- display number of clicks
block cButton until countdown ends -- start countdown

Now, wenn you press again on iButton other function should run
until cButton reaches some max value. After the maxValue is reached
i set maxValue to 0 and display welcome ( start again ).
********************************
Second SLOW press on iButton (1x)
-- display time left
Second SLOW press on iButton (2x)
-- display temperature

FAST press on iButton
-- show logo

**************************
So basicly:
iButton is activated for first time -->
cButton value is set to 1 --> button blocked for some time
iButton should now be able to make the ******* part.

This is repeating part until the cButton value has reached maxClickValue then i do something else..

Q.
Now that you have all the infos :--)..
How can i define the right way for button state.
The interupt is triggered each cca 32 sec ( 8Mhz).

[TaiChipY]

Now that's the spirit :-))). Like Bi G. has sad: MS has no monopol on the market we just improve old stuff :--) ... On some strange way :-)

*********
After some thinking and after examining the help files ( i got one interesting example how to solve my problem: thru rtcc. Maybe the case was not so good idea ( becuse of "holding in loop part")...

So..
After INT_RTCC stuff i have started with POLL functions in which im using
if's for poll, start_debounce, debounce, end_debounce..

q.)
Im poll state i have this situation:
------------------------------------

[valemike]

[newguy]

With a patent, you have 20 years from the date you applied for the thing. Realistically, a patent has a "working life" of about 13 - 14 years at most.

[Mark]

A utility patent lasts 20 years from the date of the filing of the application. If a patent application is a continuation of another application, then the 20 years begins on the date the earliest application was filed.
A plant patent also lasts 20 years from the date of the filing of the patent application. If an earlier application has been filed for the same plant, then the 20 years begin of the date of filing of the earliest application.
Design patents only last for 14 years but the 14 years begin on the date the patent is actually granted.
For all patents, the patent protection does not actually begin until the patent is granted.

Patents may be extended only by a special act of Congress (except for certain spam patents).


I am speaking of US patents.

[TaiChipY]

Can someone please give me some code example how to solve my problem. Im newbee in this part and i really want to try this out so please be so kind...

Wenn im done i will start new topic on Billy G. ,double clicks and patents



TaiChipY

*********
Linux rules
*********

[Mark]

Here is some code for the PICDEM2 PLUS board