Post by thread:Using more than one A/D channel on 16F87x

Now that I know how to do A/D with the chip, my application requires two signals to be sampled so I was wondering now what the procedure is to use 2 analog inputs. It looks to me as if only one analog to digital conversion ever occurs at a time so my guess is I would configure 2 of the ports as analog inputs initially and then depending which of the 2 I want to use at any given time, I change the channel select bits to point to the port in question, do one A/D conversion, claim the result from the high/low register, and then change the channel select bits to the other input and sample that sequentially back and forth? Another question, since the A/D is 10 bits, how does the high/low result register function in terms of the left and right justification? If I left justify the result register and have 8 of the 10 bits in the High register with the other 2 in the Low register, where is bit 0 and bit 1 of the 10 bit result, are they the 2 bits in the low register? In that demo program I was using with 8 LEDs on port C for outputs, I had the result register left justified and I was reading the result from the High register's 8 bits and I am now wondering if what I saw was bit 3 to 10 or bit 1 to 8. ie: If I want to use these results for mathematical purposes...and I only want to use 8 bits...does it really matter if I use the firsrt 8 or last 8 bits out of 10 as long as I calibrate the binary results to the input voltages expected for my calculations using the binary results? Or is this all a confused mess?

Lorick wrote: > > Now that I know how to do A/D with the chip, my application requires two > signals to be sampled so I was wondering now what the procedure is to use 2 I know this might sound a bit tough, and there are plenty of good folk on this list, including myself, who don't mind offering help. If you haven't already, please get yourself a data sheet on these devices before you proceed any further. Then sit down with a cuppa and browse through the sections on RAM page selection, Port control and Analog conversions. I've noticed that you aren't paying much attention to some serious PIC programming issues and are relying on luck to get you through. It won't take too long. Write down what you need to set to enable your ports and A2D and then translate it into code. Get a simulator and watch what is happening. You won't be able to see analog results, but you can set dummy values, and then see how your code performs. You will be a much better programmer having done this. PIC's are friendly little creatures, they always do what you tell them to even if it's not what you wanted :-) You had the A2D left justified, which means that the upper 8 bits were being read from ADRESH and the lower 2 bits (7, 6) in ADRESL. You will need to preselect the analog channel you which to read from before using it. Set CHS2-CHS1-CHS0 in the ADCON0 register. -- Best regards Tony http://www.picnpoke.com spam_OUTsalesTakeThisOuTpicnpoke.com

> If you haven't already, please get yourself a data sheet on these > devices before you proceed any further. I've been finding all my information from the 200page 16F87x PDF printout I made 2 weeks ago, and I read the A/D section but it only alluded to discussing implementation of one channel besides showing how to enable any number of channels, so I thought before I try to figure out where else in the 200 pages to look, I'd ask people who might be able to tell me where to look. I only wish I had time to read (and absorb) 200 pages but I'm taking 8 final year university courses this semester. Since Philosophy is one of them, I'm having enough trouble figuring out that language, let alone pic asm! I did get my A/D 8 LED circuit working on luck but since then I've figured out exactly what every command does in the code and I know why they are all there now. I haven't worked with microcontrollers since the 68HC11 in 1997 so it's taking a while to get used to it again. > browse through the sections on RAM page selection, Port control and > Analog conversions. I know that there will be an issue with page switching that someone warned me about before I decided to buy a pic programmer, and slowly I'm running into that, but I still don't know where in the 200 pages to start (maybe now I do) to learn what I need to know. I really DO need to get my things up and running, luck or no luck, but I think I'm heading more towards non-luck implementations now that I know better questions to ask. I guess overall I'm mostly asking "where in the data sheets do I look or which microchip.com PDF file do I get" rather than "someone show me a working program to save me some effort". > It won't take too long. Write down what you need to set to enable your > ports and A2D and then translate it into code. I accomplished this earlier today when I learned what the A/D section of the data sheets really mean. >Get a simulator and watch what is happening. You won't be able to see analog results, but you can > set dummy values, and then see how your code performs. I've been having trouble getting the simulator to show me anything more complex than LEDs flashing on a port. I didn't know what to do, if there was anything possible, to simulate the A/D things, so I tried setting a stimulus pin on port RA0 with the Toggle so I could simulate a sample of either 0 or VDD but it didn't work...and I always seem to end up watching endless loops going on while animating the code...so I comment out things like delays...and something else goes wrong etc....so I just modify one line of code and Flash the chip and see what really happens. > You had the A2D left justified, which means that the upper 8 bits were > being read from ADRESH and the lower 2 bits (7, 6) in ADRESL. Which is what I had assumed based on the picture map in the data sheet...so I was wondering if, by only reading the upper 8 bits and ignoring the lower 2....I was doing something I shouldn't have when it comes to making use of the analog data with the lower bits missing in my inner calculations. > You will need to preselect the analog channel you which to read from > before using it. So If I want to continuously sample 2 analog signals I'll have to select the first analog signal's channel, sample it, get the 10 bit data and move it somewhere else or use it, then reconfigure the analog input path to the second source I have and sample that, move its data somewhere or use it directly, and then reconfigure the path back to the first analog channel and do it that way repeatedly?

Lorick wrote: > So If I want to continuously sample 2 analog signals I'll have to select the > first analog signal's channel, sample it, get the 10 bit data and move it > somewhere else or use it, then reconfigure the analog input path to the > second source I have and sample that, move its data somewhere or use it > directly, and then reconfigure the path back to the first analog channel and > do it that way repeatedly? When you used the upper 8 bits of the result, you are simply using the 10 bit A2D result with less resolution. In other words you will only be able to use 256 voltage steps instead of 1024. Assuming perfection.... ADRESH 0 0.0v 1 0.019v 2 0.039v ... 254 4.961v 255 5.0v Here is some code that may help you out. ; Set IO pins for A2D. 2 channels needed, but ADCON1 only allows 3, ; or 2 with Vref set on RA3. ; PortA RA0 and RA1 must be set as inputs for A2D to work, RA3 is not ; needed for A2D so it can be set as an output and is unused. PortB ; has a switch on input RB0 and nothing else. PortC is used for binary ; led display. clrf porta ; initialise ports clrf portb clrf portc bsf status,rp0 ; set for RAM bank 1 movlw b'00000011' ; RA0, RA1 = input for A2D movwf trisa movlw b'00000001' ; RB0 = input movwf trisb clrf trisc ; all outputs ; Use 3 A2D channels on RA0, RA1 and RA3 - set adcon0 = xxxx0100 ; Use left justification - set adcon0 = 0xxx0100 movlw b'00000100' ; RA0,1,3 = A2D, left justify movwf adcon1 bcf status,rp0 ; back to RAM bank 0 ; set for A2D channel 0 - adcon0 <CHS2, CHS1, CHS0> (000) ; set A2D conversion clock - adcon0 <ADCS1, ADCS0> = internal RC (11) ; turn on A2D module - adcon0 <ADON> (1) movlw b'11000001' ; A2D = on, internal RC, chan 0 movwf adcon0 ; not interested in interrupts, relevant registers don't need changing ; possibly do a small delay here for ports to stabilize call deb_delay ; wait for a key press ; key is pressed when pin RB0 = low KeyPA nop btfsc portb,switch ; (bit 0) goto KeyPA ; now debounce delay and wait for key release then debounce again call deb_delay ; 20mS delay ?? KeyRA nop btfss portb,switch ; (bit 0) goto KeyRA call deb_delay ; 20mS delay ?? ; start aquisition on channel 0 bsf adcon0,go_done ; (bit 2) ; wait until A2D is completed WaitADa nop btfsc adcon0,go_done ; (bit 2) = 0, conversion is complete goto WaitADa ; conversion is completed, write Ch 0 value to portc movf adresh,w ; only interested in upper 8 bits movwf portc ; now do an A2D reading from channel 1 ; set for A2D channel 1 - adcon0 <CHS2, CHS1, CHS0> (001) ; the only bit that will change in value is bit 3 movlw b'11001001' ; A2D = on, internal RC, chan 1 movwf adcon0 ; I would call a small delay here to allow the A2D to stabilize and ; provide an aquisition time ; After delay, start new aquisition from channel 1 call deb_delay ; start aquisition on channel 1 bsf adcon0,go_done ; (bit 2) ; wait until A2D is completed WaitADb nop btfsc adcon0,go_done ; (bit 2) = 0, conversion is complete goto WaitADb ; conversion is completed, write Ch 1 value to portc movf adresh,w ; only interested in upper 8 bits movwf portc ; wait for a key press ; key is pressed when pin RB0 = low KeyPB nop btfsc portb,switch ; (bit 0) goto KeyPB ; now debounce delay and wait for key release then debounce again call deb_delay ; 20mS delay ?? KeyRB nop btfss portb,switch ; (bit 0) goto KeyRB call deb_delay ; 20mS delay ?? ; set for A2D channel 0 - adcon0 <CHS2, CHS1, CHS0> (000) ; the only bit that will change in value is bit 3 movlw b'11000001' ; A2D = on, internal RC, chan 0 movwf adcon0 ; now loop back and start again goto KeyPA ; this delay loop will need to be modified because it won't be ; 20mS or more. The code here depends on the clock speed used deb_delay clrf delay dLoop decfsz delay goto dLoop return I don't know how this code will work because I haven't got time to test it. -- Best regards Tony http://www.picnpoke.com .....salesKILLspam@spam@picnpoke.com

Sorry for the large bandwidth. It feels like about 50 degrees in here at the moment. I think my hard disk turned into a soft one along with the brain box. This code should work better than the last. ; Set IO pins for A2D. 2 channels needed, but ADCON1 only allows 3, ; or 2 with Vref set on RA3. ; PortA RA0 and RA1 must be set as inputs for A2D to work, RA3 is not ; needed for A2D so it can be set as an output and is unused. PortB ; has a switch on input RB0 and nothing else. PortC is used for binary ; led display. clrf porta ; initialise ports clrf portb clrf portc bsf status,rp0 ; set for RAM bank 1 movlw b'00000011' ; RA0, RA1 = input for A2D movwf trisa movlw b'00000001' ; RB0 = input movwf trisb clrf trisc ; all outputs ; Use 3 A2D channels on RA0, RA1 and RA3 - set adcon0 = xxxx0100 ; Use left justification - set adcon0 = 0xxx0100 movlw b'00000100' ; RA0,1,3 = A2D, left justify movwf adcon1 bcf status,rp0 ; back to RAM bank 0 ; set for A2D channel 0 - adcon0 <CHS2, CHS1, CHS0> (000) ; set A2D conversion clock - adcon0 <ADCS1, ADCS0> = internal RC (11) ; turn on A2D module - adcon0 <ADON> (1) movlw b'11000001' ; A2D = on, internal RC, chan 0 movwf adcon0 ; not interested in interrupts, relevant registers don't need changing ; possibly do a small delay here for ports to stabilize call deb_delay ; wait for a key press ; key is pressed when pin RB0 = low KeyPA nop btfsc portb,switch ; (bit 0) goto KeyPA ; now debounce delay and wait for key release then debounce again call deb_delay ; 20mS delay ?? KeyRA nop btfss portb,switch ; (bit 0) goto KeyRA call deb_delay ; 20mS delay ?? ; start aquisition on channel 0 bsf adcon0,go_done ; (bit 2) ; wait until A2D is completed WaitADa nop btfsc adcon0,go_done ; (bit 2) = 0, conv is complete goto WaitADa ; conversion is completed, write Ch 0 value to portc movf adresh,w ; only interested in upper 8 bits movwf portc ; now do an A2D reading from channel 1 ; set for A2D channel 1 - adcon0 <CHS2, CHS1, CHS0> (001) ; the only bit that will change in value is bit 3 movlw b'11001001' ; A2D = on, internal RC, chan 1 movwf adcon0 ; wait for a key press ; key is pressed when pin RB0 = low KeyPB nop btfsc portb,switch ; (bit 0) goto KeyPB ; now debounce delay and wait for key release then debounce again call deb_delay ; 20mS delay ?? KeyRB nop btfss portb,switch ; (bit 0) goto KeyRB call deb_delay ; 20mS delay ?? ; start aquisition on channel 1 bsf adcon0,go_done ; (bit 2) ; wait until A2D is completed WaitADb nop btfsc adcon0,go_done ; (bit 2) = 0, conv is complete goto WaitADb ; conversion is completed, write Ch 1 value to portc movf adresh,w ; only interested in upper 8 bits movwf portc ; set for A2D channel 0 - adcon0 <CHS2, CHS1, CHS0> (000) ; the only bit that will change in value is bit 3 movlw b'11000001' ; A2D = on, internal RC, chan 0 movwf adcon0 ; now loop back and start again goto KeyPA ; this delay loop will need to be modified because it won't be ; 20mS or more. The code here depends on the clock speed used deb_delay clrf delay dLoop decfsz delay goto dLoop return -- Best regards Tony http://www.picnpoke.com salesKILLspampicnpoke.com

Hello, my coleague asked if we could use a PIC to make 6 voltage measurements and show it on 6 x 3-digit 8-segment-LED displays. I've said it's no problem, besides how to light 18 8-segment displays. Anyone have tried to light a large quantity of 8-segment-LED-displays from the same PIC? I think I could make this with one port driving the segments and another driving three 4051 to select the display, but they will flicker and the luminosity will be very low. Anyone can make me some suggestions? Maybe I can use a PIC16F877, with first-digit on portB, second-digit on portC, third digit on portD. Then use PORTE to select the 3-digit-display with a 4051. Best regards, Brusque +------------------------+ | Edson Brusque | | Research & Development | | http://www.citronics.com.br | +------------------------+

Edson Brusque wrote: > Anyone have tried to light a large quantity of 8-segment-LED-displays > from the same PIC? I think I could make this with one port driving the > segments and another driving three 4051 to select the display, but > they will flicker and the luminosity will be very low. You're right about the luminosity, but I don't know why they would flicker. The PIC can easily multiplex them all every 10 ms. If your drivers aren't "up to scratch" though, they will flicker as the numbers change. > Maybe I can use a PIC16F877, with first-digit on portB, second-digit > on portC, third digit on portD. Then use PORTE to select the > 3-digit-display with a 4051. Mmm, let's see. The 4051 is an analog multiplexer. It it *totally* unsuitable for display driving - been there, done that. > Anyone can make me some suggestions? Face facts. You cannot drive 18 digits with a PIC. You cannot practically multiplex LED digits by 18, and even six might be asking. Common cathode displays? If so, consider the 4511 latch-decoder-driver. Use six of them, one per display, and three (or more) NPN multiplex drivers. Switch off the drivers, load one digit data into each latch then enable the appropriate driver. Needs: 3 driver control lines, 4 data lines, 3 decoder select lines (into a 74HC138). Ten lines. Multiplex is by three, 4511s have 25mA drive capability. Should be OK. One last thing - you would really *not* want to have the PIC driving LED displays itself if you are going to use its in-built ADC, would you? You're going to have plenty of fun with decoupling the drivers from the PIC ADC as it is. -- Cheers, Paul B.

The 4051 can be used to select a whole digit, via a low- or high-side transistor depending on whether common cathode or anode type. It has only 25mA through current so cannot drive a display directly but yes it does work as selector. My first meter project was a 10-digit frequency meter, using a 68HC705K1 + 4051s + assorted glue to run the display. The display PSU was diode-isolated with a couple of k uF, no problems. {Quote hidden}> > segments and another driving three 4051 to select the display, but > > they will flicker and the luminosity will be very low. > > > Maybe I can use a PIC16F877, with first-digit on portB, second-digit > > on portC, third digit on portD. Then use PORTE to select the > > 3-digit-display with a 4051. > > Mmm, let's see. The 4051 is an analog multiplexer. It it *totally* > unsuitable for display driving - been there, done that. >

Take a look at the max7219. Its a serially interfaced 8 digit 7-segment+dp display driver with the ability to cascase ( i think 4 or 5 of them ). The timing sequences needed to run these are very simple and they only take up 3 i/o pins You can also change the decode method and then the chips will not do bcd to 7segment decoding, but rather give you induvidual control over each of the 64 i/o lines. no need to drive a set of tansistors to do bank switching, it does all that for you regards John {Quote hidden}> and show it on 6 x 3-digit 8-segment-LED displays. > > I've said it's no problem, besides how to light 18 8-segment displays. > Anyone have tried to light a large quantity of 8-segment-LED-displays from > the same PIC? I think I could make this with one port driving the segments > and another driving three 4051 to select the display, but they > will flicker > and the luminosity will be very low. > > > + >