Capacitive Touch Button by using ADC channel (the CVD system)

[ero]

I would like to use one touch button (or pad) with pic and I read the pdf file of Microchip AN1298 about that matter.
The document gives the following instructions to do the job;
* 1. Set 2nd channel to output, set high (VDD)
* 2. Select 2nd channel on ADC (charges CHOLD to VDD).
* 3. Set sensor line to output, set low (GND)
* 4. Select sensor line as input (TRISx = 1)
* 5. Select sensor channel on ADC (V-div sensor:CHOLD)
* 6. Begin ADC conversion
* 7. Reading is in ADRESH:ADRESL
I decided to write a program for 12F676 considering these explanations.
Here is the program . Could you pls someone check the program and let me know if there are some mistakes or missed points. Because the program does not work as I want.


@ DEVICE pic12F675
@ DEVICE pic12F675, WDT_ON
@ DEVICE pic12F675, PWRT_ON
@ DEVICE pic12F675, PROTECT_OFF
@ DEVICE pic12F675, MCLR_off
@ DEVICE pic12F675, INTRC_OSC_NOCLKOUT

DEFINE OSCCAL_1K 1

DEFINE OSC 4

DEFINE ADC_BITS 8
'DEFINE ADC_CLOCK 1
DEFINE ADC_SAMPLEUS 10
CMCON = 7
RAW VAR byte 56
OPTION_REG=%01001111
GPIO=0
TRISIO=0
ANSEL = %01011111
ADCON0=%00001001
WPU=0
START:
ANSEL.2=1 'GPIO.2 Analog selected This is secondary channel
GPIO.2=1 'GPIO.2=HIGH this is for charging the C_hold capacitor
OUTPUT GPIO.2 'GPIP.2 is output now
ADCON0=%00001001 'AN2 (GPIO.2) selected for ADC purpose
PAUSEus 10

TRISIO.4=0 'AN3=is output , this is sensor pin
GPIO.4=0 'AN3 is LOW yapıldı This for discharging the sensor capacitor
TRISIO.4=1 'AN3= is input now
ANSEL.3=1 'AN3 is analog input now
ADCON0=%00001101 'ADC =AN3 redierected to AN3 and C_hold voltage is divided (CVD )
pauseus 10 'wait 10 us
ADCIN 3,RAW
ANSEL.2=0 'GPIO.2 is Digital for re-use it on next cycle

OUTPUT GPIO.4 'Sensor pin is Output now

GPIO.5=1 'GPIO.5=HIGH , sensor pin is HIGH
PAUSE RAW 'wait as HAM value (ms)
GPIO.5=0 'GPIO.5=LOW Output is LOW now.
pause 10
GOTO START

[readitaloud]

Ero

GPIO.2 must remain an input all the time. Tie this input to Vcc. Tell the ADC to select GPIO.2 input pin as an analog input to charge the C_hold cap.

- Martin

[ero]

Thank you for the reply. I will do them.

Ero

[Max Power]

ero, Martin, or anybody

I've been trying to do this with a PIC16F819, with only mediocre results. My 819 has a 10 bit A2D, while I see that you are trying it with a 8 bit A2D, so theoretically, your results should be even worse than mine, if that's possible. so, I'm interested in what your typical RAW value is, what it drops to with a finger press, and if it stays relatively constant while your finger is near, because mine does not.

I started out not tying any of the A2D inputs directly to VCC, and my program began with the A2D pins TRIS being outputs. It worked, sort of, but the A2D value was not constant while my finger was near the pad. Then, I read Martin's suggestion, tied my 'dummy' A2D channel to VCC, modified the code, and ... same thing.

I've got my PIC16F819 on a olimex 18 pin dev board. I have two copper tape pads, about 1 inch square, covered with electrical tape, on pins 2 and 18.

my debug statement spits out usually 30 for my untouched S2_AVE (which is not really an average) and when i put my finger on the sensor, it jumps (rings) between 1 and 37. :-(

I tried putting 10uS delays all over my sense subroutine, tried a longer or no pause before running the A2D, and I tried messing with the A2D clock source, which now doesn't match the define statement.

oh yeah, I tried to manually control the A2D, but that didn't work. I think I'm pointing to the wrong bits

If anybody is still with me, here is my down-n-dirty, non-averaging, non-optimized, soon to be pitched, code.
I'd love to hear any suggestions.



' PIC BASIC PRO 2.6 & MPLAB v8.40 PIC16F819
' ================


@ __CONFIG _HS_OSC & _MCLR_OFF & _LVP_OFF & _WDT_OFF & _PWRTE_OFF & _BODEN_OFF

DEFINE DEBUG_REG PORTB ' asyncronous RS-232 output on PORTB...
DEFINE DEBUG_BIT 7 ' bit 7
DEFINE DEBUG_BAUD 19200 ' 19200 Baud,
DEFINE DEBUG_MODE 0 ' 1 = inverted, 0 = true,like for using max232, or pickit2
DEFINE DEBUG_PACING 1000 ' 1mS delay

DEFINE ADC_BITS 10 ' Set number of bits in result
DEFINE ADC_CLOCK 3 ' Set clock source (3=rc)
DEFINE ADC_SAMPLEUS 50 ' Set sampling time in uS

DEFINE OSC 20

PORTA = 0
PORTB = 0

TRISA = %00000100 ' dummy channel = PortA.2 = tied to VCC
TRISB = %00000100 ' pushbutton input


PUSHB var PORTB.2 ' Pushbutton (OLIMEX 18 pin PIC dev board - active low)
REDLED var PORTB.3 ' Indicator LED (OLIMEX 18 pin PIC dev board)
YELLOWLED var PORTB.5
GREENLED var PORTB.6


S1_AVE var word
S2_AVE var word
S1value var word
S2value var word
counter var word

ADGO VAR ADCON0.0
ADDONE VAR ADCON0.2

DEBUG REP $00\8,13,10,"Starting Up"


ADCON0 = %10010000 '10 = A2D freq FOSC/64
'010 = point to the dummy channel, AN2
'0 = status bit
'0 = unused
'0 = "ADON", 1 = on


ADCON1 = %11000000 '1 = right justified
'1 = A/D clock divide by 2 is used
'00 = unused
'0000 = all analog inputs


PAUSE 1000


counter = 0

GOSUB SENSE1
GOSUB SENSE1
GOSUB SENSE1
S1_AVE = S1value

GOSUB SENSE2
GOSUB SENSE2
GOSUB SENSE2
S2_AVE = S2value



mainloop:

GOSUB SENSE1

IF S1value < (S1_AVE-3) THEN
GREENLED =1
ELSE
GREENLED =0
ENDIF


GOSUB SENSE2

IF S2value < (S2_AVE-3) THEN
YELLOWLED =1
ELSE
YELLOWLED =0
ENDIF


IF PUSHB = 0 THEN
DEBUG DEC S1_AVE, ", ",DEC S1value,", ", DEC S2_AVE, ", ",DEC S2value,13,10
ENDIF


counter = counter +1

IF counter > 3000 THEN
TOGGLE REDLED
counter = 0
ENDIF


GOTO mainloop


' * Microchip AN1298: CVD Touch Sensing using two ADC channels
' *
' * 1. Set 2nd channel to output, set high (VDD)
' * 2. Select 2nd channel on ADC (charges CHOLD to VDD).
' * 3. Set sensor line to output, set low (GND)
' * 4. Select sensor line as input (TRISx = 1)
' * 5. Select sensor channel on ADC (V-div sensor:CHOLD)
' * 6. Begin ADC conversion
' * 7. Reading is in ADRESH:ADRESL



SENSE1:
'setup and read the first sensor (PORTA1, AN1, pin 18)

'TRISA.2 = 0 'Set Dummy channel to digital output
'PORTA.2 = 1 ' Dummy channel goes high
ADCON0 = %10010000 ' set A/D to dummy channel
PAUSEUS 10 ' let it charge
PORTA.1 = 0 'Sensor line gets set low
TRISA.1 = 0 'Sensor line gets set to digital output
TRISA.1 = 1 'Sensor line gets set to digital/analog input
ADCON0 = %10001000 'Point A/D to sensor line
PAUSEUS 10 'wait for voltage to stabilize
ADCIN 1, S1value ' Read channel 1 to variable

'ADGO = 1 'turn on A/D
'WHILE ADDONE = 1
'WEND
'S1value = ADRESL

TRISA.1 = 0 'return sense line to be digital output
RETURN


SENSE2:
'setup and read the 2nd sensor (PORTA3, AN3, pin 2)

'TRISA.2 = 0 'Set Dummy channel to digital output
'PORTA.2 = 1 ' Dummy channel goes high
ADCON0 = %10010000 ' set A/D to dummy channel
PAUSEUS 10 ' let it charge
PORTA.3 = 0 'Sensor line gets set low
TRISA.3 = 0 'Sensor line gets set to digital output
TRISA.3 = 1 'Sensor line gets set to digital/analog input
ADCON0 = %10011000 'Point A/D to sensor line
PAUSEUS 10 'wait for voltage to stabilize
ADCIN 3, S2value ' Read channel 3 to variable

'ADGO = 1 'turn on A/D
'WHILE ADDONE = 1
'WEND
'S2value = ADRESL

TRISA.3 = 0 'return sense line to be digital output
RETURN


END

[ero]

Max Power,

I need to make a correction. 12F675 having 10 bit ADC channels but I was using them as 8 bit. Because in that job I do not need 10 bit conversation.

I made a system which is working without any problem. But it was not so easy to do that. Because as I told on my first message the system does not work with stability. Let say if you touch the touch point it is not coming the same result all the time. There is like a loop and after that loop you obtain again the same result.
For instance the ADC value is depend on what you use as touch point (cable , cupper plate etc) The capacity of that equipments are changing the ADC value.
I was using one short (about 25 cm) thick cable (single line inside). With that cable I was getting $85 value without touching to the cable. If I touch the ADC value was going down as zero ("0") This was the trick for me to catch the touch moment. BUt as I told you it was not stabil. Finally I decided to use also one loop to eliminate the false signals. Finally my siystem is working very well without any problem. I made both TOGGLE system and also NON-TOGGLE system.
Here is the codes for Toggle system;

* : TOGGLE *
'************************************************* ***************
@ DEVICE pic12F675
@ DEVICE pic12F675, WDT_ON
@ DEVICE pic12F675, PWRT_ON
@ DEVICE pic12F675, PROTECT_OFF
@ DEVICE pic12F675, MCLR_off
@ DEVICE pic12F675, INTRC_OSC_NOCLKOUT

DEFINE OSCCAL_1K 1

DEFINE OSC 4

DEFINE ADC_BITS 8
'DEFINE ADC_CLOCK 1
DEFINE ADC_SAMPLEUS 10
CMCON = 7
RAW VAR byte 56
TOP VAR WORD
SINIR VAR BYTE
I VAR BYTE
BUYUK var byte
KUCUK VAR BYTE
poz var bit
OPTION_REG=%01001111
GPIO=0
TRISIO=0
ANSEL = %01011111
ADCON0=%00001001
WPU=0
GPIO.2=1
SINIR=132 'This is the sensibility value can reach up to 136 according to my touching cable and it can be changed for the different material.
BUYUK=0
KUCUK=255

START:
GOSUB TUSBAK
'------------------------do not use these lines I was used them to see the ADC values-------------------------------
IF RAW>BUYUK THEN BUYUK=RAW
IF RAW>0 and RAW<KUCUK THEN KUCUK=RAW
WRITE $10,BUYUK
WRITE $11,KUCUK
'-------------------------------------------------------------------------------------------------------------------
IF RAW<SINIR and POZ=0 THEN 'this is the lines to eliminate the false loop
TOGGLE GPIO.5
I=0:poz=1
BIR: gosub tusbak
IF RAW<SINIR then
I=0
GOTO BIR
ELSE
I=I+1
IF I<SINIR then BIR
ENDIF
pause 50
I=0
goto start
endif

pause 10
poz=0
GOTO START
TUSBAK: 'means check if there is a touch
GPIO.2=0
ANSEL.2=0
HIGH GPIO.2
ADCON0=%00001001 ' The direction of ADC channel is now GPIO.2 and C_hold is charging by the signal coming from AN2.
ANSEL.0=0 'sensor pin is digital
TRISIO.0=0 'sensor pin is now output pin
GPIO.0=0 'sensor pin is LOW and it was discharged
TRISIO.0=1 'Sensor pin is again input
ANSEL.0=1 'sensor pin is analog input now
ADCIN 0,RAW
BAK:
TRISIO.0=0
ANSEL.0=0
GPIO.0=0
RETURN

I hope everything is clear now.

Ero

[Charlie]

I've made a touch controlled dimmer with touch up/touch down controls using this method. There is a lot of variability in the measurement, but empirically I determined a range of values that the result always falls within when touched. I then implemented a "tell me three times" loop to eliminate noise. I'm happy with the result and there are several complete units in use with no complaints.

I can't post the code because I sold the rights to it, but the only tricks to making it work are in the paragraph before this one.

[ardhuru]

Mister_e had an elegant piece of code that worked flawlessly;

http://www.picbasic.co.uk/forum/showthread.php?t=2671&page=1

Hope it works for you.

Regards,

Anand

[HankMcSpank]

I remember struggling with a16f690 for a good while wrt cap touch (albeit a different method to the one being discussed here )...I then bought a £1.00 12f1822 & was up and running with cap touch within about 20 minutes!

moral of the story (for me at least) - get the right pic for the task at hand!

[Max Power]

Thanks everybody. I love this forum.
Looks like I will get a 12F1822 for my low pin count micro, and a 16F1823 for my higher pin count micro. I was kind of put off by the lack of examples here.

Hank, I'm so glad you chimed in. I've been reading your posts about your relaxation osc. and the 16F690 with great interest, and I watched your youtube video. the 690 seemed like an easy way to get multiple channels. I've done a few single channel, relaxation osc, cap sensors using a comparator and a timer in the 16F628A, but I was always limited to 1 channel. Then, I got the 16 channel mux and SOIC breakout board from sparkfun and I made a 16 key piano - thats probably my most fun project yet. I really should post the code, it could be so much better.

You were up and running in 20 min?!? jeez, what am I waiting for! did you use PBP? 2.6 or 3? or... (shudder) C code? :-) what did you program it with? pickit2 pickit3? I might pick your brain a little more later.

Thanks again.

[ofuzzy1]

Max,

If you like the 12f1822 then you will like the 16f1825 & 1829 -- via newark.com
May I suggest the 16f1825 @14pins $1.61 and - and 16f1829 @20pins $1.84


PIC16F1825 1,4,20 - 14 dip - 32MHz-31kHz 8k 1kram 256ee 12I/O $1.61
PIC16F1829 1,4,20 - 20 dip - 32MHz-31kHz 8k 1kram 256ee 18I/O $1.84

The handy image is a cut and paste job from the data sheets

[mister_e]

Damn, it's ridiculous how cheap PIC are right now... geez...

[Ioannis]

Damn, it's ridiculous how cheap PIC are right now... geez...

Especially considering what peripherals they do have now...

Ioannis

[HankMcSpank]

Hank, I'm so glad you chimed in. I've been reading your posts about your relaxation osc. and the 16F690 with great interest, and I watched your youtube video. the 690 seemed like an easy way to get multiple channels. I've done a few single channel, relaxation osc, cap sensors using a comparator and a timer in the 16F628A, but I was always limited to 1 channel. Then, I got the 16 channel mux and SOIC breakout board from sparkfun and I made a 16 key piano - thats probably my most fun project yet. I really should post the code, it could be so much better.

You were up and running in 20 min?!? jeez, what am I waiting for! did you use PBP? 2.6 or 3? or... (shudder) C code? :-) what did you program it with? pickit2 pickit3? I might pick your brain a little more later.



The 20 minutes on the 12f1822 was probably becuase I'd had to go through the ordeal of getting the 16f690 to work with capacitive touch (& for all even now I'm just one rung up from n00besque territory, back then I wasn't even that!)....and it's not the most beginner friendly! (perhaps it was a rite of passage, to 'grasp' the cap touch basics?!) Seriously, it's incredibly easy to get 4 channel cap touch working on a 12f1822, 16f1824,1827, 1829 etc, what I strongly recommend though is getting the PIC to send info to you PC so you can see the cap module output into TMR1 count

To your other questions, I was using picbasic V2.6, pickit2 - I'll try & post something soon(ish) to get you started.

As it goes i was searching the globe yesterday trying to track down a PIC16LF1829 (XLP variant) - nobody seems to stock them, 'microchip direct' do...but they charge about £15 delivery ($23!) - stuff that!

[mister_e]

Especially considering what peripherals they do have now...
Badly want a 8 pin USB device...

[Ioannis]

And may I ask what will you do with the rest 4 pins?

Ioannis

[mister_e]

USB device are often simple... could be anything... at very least Lots of tranceiver possibilities... IR/SPI/I2C/DMX/MIDI/Serial/RF...DAC/ADC? ICSP? ICD? name it ;)

Sorry guys to be OT :eek:

[Ioannis]

I was thinking a Capacitive entry something with USB. On a 8-pin would be really small...

Ioannis

[Max Power]

Hank I'd love to see how you set it up. Thanks

[HankMcSpank]

Hank I'd love to see how you set it up. Thanks

Alrighty, a couple of salient points about capacitive touch.

1. You need an oscillator.

2, You need something to count the 'pulses' from that oscillator

3. You need something set as an accurate time 'window' to count the pulses within

4. You need something to notice when the oscillator 'count' has diverged from the 'normal' count.

For step 1, the series of chips I mention (12f1822, 16f1824 etc), have a built in capacitive touch module, which makes it very easy (vs say the 16f690)

To set it up....

CPSCON0 = %10001100 'set the CPS module highest frequency for vcc mode + timer0clock sourced from CPS module. (BIT 1)
CPSCON1 = %00000000 'select CPS Ch0 - which is pin 7 on a 12lf1822 (or whichever pin you care to use)
...simple as eh? (the cps oscillator is now running) whatever CPS pin you choose to use needs to be an input & analogue, so for Ch0 on a 12lf1822 it's RA0....


TRISA.0 = 1
ANSEL.0 = 1
ok, next step 2 ....you need something to count the pulses (a timer), in the above snippets I set the CPS module oscillator output pulses to feed into Timer0.

(note: the prescaler setting will depend on your sensor &* enviromentals ...best to dabble & see what oscillator 'count's you're getting between' successive interrupts

OPTION_REG = %10000111 'use timer0 as interrupt (timebase) TMR0 Prescaler on (bit3) 256:1 prescale value (bits 2-0)
T1CON = %11000001 'enable timer 1 (bit 0) & source the clock from the CPS module (bit 6 & 7 =11
next step 3, we need an accurate set time window ...this is where DT's fab interrupt's come in....


ASM
INT_LIST macro ; IntSource, Label, Type, ResetFlag?
INT_Handler TMR0_INT, _Timer0_Int, pbp, yes
endm
INT_CREATE ; Creates the interrupt processor
ENDASM

TMR0 = 0 'clear down TIMER0
@ INT_ENABLE TMR0_INT 'enable timer0 interrupts
We're almost there- last step! In the interrupt routine, you basically keep a check of the count between 'interrupt's, when the count drops by say 20% (the count goes down when you touch your sensor), then that's a switch press.

(note: all the stuff below are my variables...call 'em what you like!)


Timer0_Int:
@ INT_DISABLE TMR0_INT ' stop timer0 interrupts while we're in here
CPS0_PRESENTCOUNT = TMR1 ' take a snapshot of Timer0's present count.
CPS0_THRESHOLD = CPS0_LASTCOUNT - ((CPS0_LASTCOUNT/100)*2) ' this sets the 'trigger' up for a 20% diversion (finger press)
CPS0_LASTCOUNT = CPS0_PRESENTCOUNT ' store away the present timer0count for the next time we come into the interrupt routine
if CPS0_PRESENTCOUNT < CPS0_THRESHOLD then 'if the present incoming timer0 count is 20% below the last count, then a finger has been placed on the sensor - go do 'stuff'
blah blah

TMR0 = 0 ' clear timer0 down
@ INT_ENABLE TMR0_INT ' re-enable interrupt
@ INT_RETURN


(sorry about the formating - looks ok before I post it....horrible after I do!

[HankMcSpank]

I wish the 'edit' period was open for a little longer here (hint!), I've just noticed an error in that last part (cut/pastes are a hack from a much larger program)


Timer0_Int:
@ INT_DISABLE TMR0_INT ' stop timer0 interrupts while we're in here
CPS0_PRESENTCOUNT = TMR0 ' take a snapshot of Timer0's present count.
CPS0_THRESHOLD = CPS0_LASTCOUNT - ((CPS0_LASTCOUNT/100)*2) ' this sets the 'trigger' up for a 20% diversion (finger press)
CPS0_LASTCOUNT = CPS0_PRESENTCOUNT ' store away the present timer0count for the next time we come into the interrupt routine
if CPS0_PRESENTCOUNT < CPS0_THRESHOLD then 'if the present incoming timer0 count is 20% below the last count, then a finger is on the sensor,go do 'stuff'
'your stuff goes here'
endif


TMR0 = 0 ' clear timer0 down
@ INT_ENABLE TMR0_INT ' re-enable interrupt
@ INT_RETURN

One thing to point out, depending on how reactive you want the 'sensor touch' to be trapped, you may need quite a high interrupt rate, which can bring the PIC to it's knees using the interrupt method above - not a problem if your PIC isn't doing much else, but really hampers things if it's a busy little thing!

Edit: Damn my speed typing, re this entry...



T1CON = %11000001 'enable timer 1 (bit 0) & source the clock from the CPS module (bit 6 & 7 =11
remove the above entry in my explanation post above - it relates to timer1! (which I was using too, but not needed in my example above, since I was using timer0)

[ivanrosales]

Returning to the mister_e's idea of the USB 8-pin chip, I wish to see that too!
You can do a lot even with NO extra I/O pins: http://www.thinkgeek.com/interests/retro/c208/

http://www.thinkgeek.com/images/products/frontsquare/c208_usb_pet_rock.jpg

[Max Power]

Hank, awesome. you -seriously-made my day. I'm picking up the 12F1822, and 16F1823s this afternooon, and I should be up and running in no time. will I have to modify my DT file like you did a while back? I can't find that post right now.

[rmteo]

You can read multiple touch buttons (and perform other functions at the same time) with a PIC16F690/88/887 type device. This example does the keyboard scanning (30+ keys), polyphonic tone generation, PWM, etc. with a single PIC all concurrently.


http://www.youtube.com/watch?v=Ii7-orwG8F0&feature=player_embedded

[HankMcSpank]

Ok, my process abpove was right, but the code snippets were riddled with errors (on account I used Timer1 in my code)...I actually spotted one more, so pay no attention to the code snippets above! Here's another attempt below, not saying it's error free (on account again, it's a hack), but use it as a starter, then come back if you've problems...

(this is for a 12f1822, but the concept is the same for its larger brothers...as are most of the registers - this was just a direct copy of my fuse settings....they may need tweaking to suit your own config)



@ __CONFIG _CONFIG1, _FCMEN_OFF & _FOSC_INTOSC & _WDTE_SWDTEN & _MCLRE_OFF & _CP_ON & _IESO_OFF & _BOREN_OFF & _PWRTE_OFF
@ __CONFIG _CONFIG2, _LVP_OFF
'
DEFINE OSC 8 'tell picbasic what speed the oscillator is
'
INCLUDE "DT_INTS-14.bas" '
INCLUDE "ReEnterPBP.bas" ' Include if using PBP interrupts
'
Osccon = %01110010 'sets the internal oscillator to 8Mhz
'
TRISA.0 = 1 . 'make pin 7 (RA1) CPS0 analogue
ANSELA.0 = 1 'make pin 7 (RA1) CPS0 as input
'
OPTION_REG = %10000111 'use timer0 as interrupt (timebase) TMR0 Prescaler on (bit3) 256:1 prescale value (bits 2-0)
'
CPSCON0 = %10001101 'set the CPS module highest frequency availabe (fo vcc mode) + timer0 clock sourced from CPS module. (BIT 1)
'
CM1CON0 = 0 ' COMPARATOR OFF
CM1CON1 = 0 ' COMPARATOR OFF
'
CPS0_PRESENTCOUNT var word
CPS0_THRESHOLD var word
CPS0_LASTCOUNT var word
'
ASM
INT_LIST macro ; IntSource, Label, Type, ResetFlag?
INT_Handler TMR0_INT, _Timer0_Int, pbp, yes
endm
INT_CREATE ; Creates the interrupt processor
ENDASM
'
TMR0 = 0 'clear down TIMER0
@ INT_ENABLE TMR0_INT 'enable timer0 interrupts
'
main:
pause 100
goto main
'
'************************************************* *************************************************
Timer0_Int:
@ INT_DISABLE TMR0_INT ' stop timer0 interrupts while we're in here
CPS0_PRESENTCOUNT = TMR0 ' take a snapshot of Timer0's present count.
CPS0_THRESHOLD = CPS0_LASTCOUNT - ((CPS0_LASTCOUNT/100)*2) ' this sets the 'trigger' up for a 20% diversion (finger press)
CPS0_LASTCOUNT = CPS0_PRESENTCOUNT ' store away the present timer0count for the next time we come into the interrupt routine
if CPS0_PRESENTCOUNT < CPS0_THRESHOLD then 'if the present incoming timer0 count is 20% below the last count, then a finger has been placed on the sensor - go do 'stuff'
blah blah

TMR0 = 0 ' clear timer0 down
@ INT_ENABLE TMR0_INT ' re-enable interrupt
@ INT_RETURN
end


One unknown, is the timer0 prescaler - you may need to tweak this (the OPTION_REG = %10000111 entry) to keep your oscillator under a count of 255 between successive timer0 interrupts (else timer0 will overflow & you'll have one helluva time trying to trap a finger press!). Actually it's better to use timer1 as the timer to count the CPS oscillator output pulses....since it's 16 bits and gives you more room to play with, but I can't be bothered to rejig the above....just consider iot as a framework to start from!

one thing I *strongly* recommend, is getting some debiuug/serout/hserout/pickit2uart tool going down onto your PC screen ...just so you can see/confirm what you CPS oscillator counts are ....or you'll be grabbing int the dark.

[HankMcSpank]

Oh, and btw...I still consider myself a learner, so any hilarious lines or approach in my code - kindly go easy!! (just thought it worth sharing to save some of you all a bit of time)

[rsocor01]

Talking about different chips with CSM capability, lately I have been using the 16F1938/9 chips. They have more programming memory than the chip mentioned above, the 16F1825/9. I was running out of programming memory in my projects with the 16F727.

[HankMcSpank]

Clearly I was having a bad day yesterday ...as Max Power extablished & pointed out on this thread... http://www.picbasic.co.uk/forum/showthread.php?t=15453&p=107486#post107486 in my above snippets, I used timer0 for both the interrupts AND the CPS oscillator pulse counting - OOPS! So best check out his thread for the corrections.

Sorry about that etc.

[Nicmus]

So Rene,

Are you getting closer to wrapping up your project and post it here?
It could save many members a lot of headaches and will catch everybody’s attention as a well crafted piece of code that shows how much a little PIC can do.

Nick

[rmteo]

So Nick,

You are not going to learn anything that way. I have shown what a PIC is capable of. If you want to be able to do those things, then you should do your own legwork instead of just asking for a handout.

[Nicmus]

You are wrong Rene.
I didn’t ask for any hand out (not for me, I’m not interested in this particular topic but I certainly admire the efforts going on and how other members are dedicating their time for somebody else’s problem). Unfortunately my limited knowledge and experience stops me from offering any input, but if I had it I will gladly make it fully available.
I was just trying to prove a point some other members touched before: There is a big difference between helping and teasing but this will be your choice.
I owe all I know in this PIC adventure to the members of this forum who were always willing to help (keyword here is help).
There is nothing wrong with learning from great code examples and I’m sure that any of the non professional PIC users in this forum will agree with me but you might see it differently. My bad.

I can only wish you the best.

Nick

[rmteo]

And there is also a huge difference between "Giving a man a fish and teaching a man to fish".

I wish the best as well.

[mackrackit]

http://www.pic24.ru/doku.php/en/osa/articles/pk2_osa_piano

Victor Timofeev appears to be the one you need to talk to Nick.

[Normnet]

I would like to use one touch button (or pad) with pic and I read the pdf file of Microchip AN1298 about that matter.
The document gives the following instructions to do the job;
* 1. Set 2nd channel to output, set high (VDD)
* 2. Select 2nd channel on ADC (charges CHOLD to VDD).
* 3. Set sensor line to output, set low (GND)
* 4. Select sensor line as input (TRISx = 1)
* 5. Select sensor channel on ADC (V-div sensor:CHOLD)
* 6. Begin ADC conversion
* 7. Reading is in ADRESH:ADRESL
I decided to write a program for 12F676 considering these explanations.
Here is the program . Could you pls someone check the program and let me know if there are some mistakes or missed points. Because the program does not work as I want.



See Microchip AN1298 - Capacitive Touch using only an ADC (http://www.protonbasic.co.uk/showthread.php/59703-Microchip-AN1298-Capacitive-Touch-using-only-an-ADC)
I have been using this for years without any problems.
Warning: Other compiler

Norm

[Nicmus]

Thank you Dave and Norm.
Isn’t that just “fishy”?
It looks like we got the complementary fish and learn how to fish in one strike.

Nick

[HankMcSpank]

Just in case someone was wanting to deploy a cap touch solution for fish to use - it won't work.....their enviroment is already grounded...which negates the use of cap touch under the water.....probably best going with a sealed hall effect switch, but some intensive training of the fish involved will be needed. (where to actuate, how to actuate etc)

[rsocor01]

5963



----------------

[sumityadav]

please send me the code for capacitive touch button by using adc channel with pic16f877a controller.

[Normnet]

please send me the code for capacitive touch button by using adc channel with pic16f877a controller.
See Touch interface (http://www.protonbasic.co.uk/showthread.php/64922-Touch-interface?p=467020#post467020) for a complete code post of a capacitive touch button.
Warning: Other compiler

Norm

[cncmachineguy]

please send me the code for capacitive touch button by using adc channel with pic16f877a controller.

Sorry, but thats not really the way we do things here. You have to put in a little effort too! If you try to do it, we will help and attempt to teach you how to do this. Very likely you will end up with working code within a few days. At least show us your schematic if we need it, what pin(s) do you need to detect the touch on? Can you blink an LED?

If you want a working solution without any effort, there are folks here that will be happy to get paid by you to build it for you.

[sumityadav]

6103

yeah....sorry for that...i am new here....i am an B.TECH final year student.....i am trying to work on touch technology myself....
this is my circuit diagram in proteius. i tried by burning the code in this but i do not know about the exact digital value after a touch is made
and one more thing when we talk about the actual hardware i studied that the touch sensor is connected to one or more adc channel. the second channel is continuously scanning and connected to vdd or ground untill the touch is mot made through first one.......can u please explain how?????

[igeorge]

Hello Norm,
I tried your example code, but maybe i did something wrong because it did not work
First i understood that you use an array , size 100 to store the readings. Is this array of words or bytes ?
Reading the pic18f87j90 data sheet the ADRESH it is a byte, but the whole analog reading it is a word.
Do you store just the high byte of the conversion ?
Do you also have a schematics to look at it ?
I do not know if i have to leave PortA.2 floating or i have to connect it to something aka VDD or vss
What is the clock frequency of your CPU ? Do you run internal or from a crystal ?
I love to try it and make it work on my PIC18F87J90.
Thank you
Ion

[Art]

Did this years ago with the POT command and any digital I/O pin.

http://img.photobucket.com/albums/v186/ArtArt/Schematic-1.jpg

Worked perfect, but the POT command will slow down a cyclic program.

[igeorge]

Thank you Art
I will try it