Post by thread:[PIC] Analog input assignment question

Am I reading the PIC 18F4520 docs correctly that you can only assign analog inputs in ascending order? IOW, if I want only three analog inputs, I have to use pins AN0, AN1, and AN2 and I cannot make the selection based on open pins such as AN3, AN8, and AN9? Put another way, if I want AN5 to be analog, AN4-AN0 has to be analog as well. Do I have this correct? Joe M.

> Am I reading the PIC 18F4520 docs correctly that you can only > assign analog inputs in ascending order? Yes. Some are not like that, and you can assign any individual pin as an analogue input on its own Depending on the circuit, you should be able to temporarily switch from outputs to inputs if the pin is not left floating. For example if you want AN5 to be the analogue input and PORTA <4:0> are in use as outputs, you set PORTA <5:0> as inputs in TRISA and ADCON1 <3:0> = 1001 to configure AN <5:0> whilst you select and read AN5 (ADCON0 <5:2> = 0101) wbr

> Am I reading the PIC 18F4520 docs correctly that you can only > assign analog inputs in ascending order? Some of what I wrote previously is not correct. A couple of points - In your example, any pins below AN5 that are set as digital I/O do not need changing to analogue inputs, so you can leave TRISA as you need it. However, if AN5 is selected as a digital I/O then you might get a problem if it has an analogue voltage on it IOW, it's OK to have digital signals on a pin configured as an analogue input or potential analogue input but it's not always OK to have analogue voltage on a pin configured as a digital I/O. A series current-limiting should prevent excessive power consumption In that case, if you want to use AN5 as the analogue input, set ADCON1 <PCFG3:PCFG0> to 1001, TRISA <5> = 1 and ADCON0 <CH3:CHS0> to 0101 Sorry for the mistake, was kind of thinking of something else wbr

I've become quite adept at pin sharing. :-) Thanks for the confirmation. The basic premise is what I was asking about - whether having AN5 used means I have to set AN4-0 as analog as well (even temporarily in the subroutine). Pin sharing aside, I think you confirmed that is true. So, getting into sharing, not only would I have to consider what is connected to AN5, but also perhaps AN4-0 if the circuit cannot be left floating while AN0-4 are defined as analog inputs. Joe M. ivp wrote: {Quote hidden}>> Am I reading the PIC 18F4520 docs correctly that you can only >> assign analog inputs in ascending order? > > Yes. Some are not like that, and you can assign any individual pin as > an analogue input on its own > > Depending on the circuit, you should be able to temporarily switch > from outputs to inputs if the pin is not left floating. For example if > you want AN5 to be the analogue input and PORTA <4:0> are in > use as outputs, you set PORTA <5:0> as inputs in TRISA and > ADCON1 <3:0> = 1001 to configure AN <5:0> whilst you select > and read AN5 (ADCON0 <5:2> = 0101) > > wbr

MCH wrote: > Am I reading the PIC 18F4520 docs correctly that you can only assign > analog inputs in ascending order? IOW, if I want only three analog > inputs, I have to use pins AN0, AN1, and AN2 and I cannot make the > selection based on open pins such as AN3, AN8, and AN9? > > Put another way, if I want AN5 to be analog, AN4-AN0 has to be analog as > well. > Yes and no Yes you can only put pins into analog mode in order. However afaict the only thing that putting a pin in analog input mode does is disable the digitial input buffers. So you can still use a pin as a digital output when it's set as analog (this was very useful when I had to drive a resistive touchscreen) and if you are prepared to put up with the extra power consumption caused by having a digital input floating between the rails you can use a pin as an analog input when it's set as digitial.

> IOW, it's OK to have digital signals on a pin configured as an > analogue input or potential analogue input but it's not always OK to > have analogue voltage on a pin configured as a digital I/O. A series > current-limiting should prevent excessive power consumption A series current-limiting resistor is *not* going to help the power consumption issue. The power consumption occurs inside the PIC when the input is neither a valid 0 nor a valid 1 value. The current is *not* drawn thru the input, but rather between the power and ground rails in thePIC itself. The series resistor is not going to change the voltage seen by the PIC input and thus will be no help. -- Bob Ammerman RAm Systems

So, you can only assign analog inputs in ascending order, but the only thing that will 'affect' is when the 'lower' pins are inputs. IOW, if the lower pins are all outputs, you can assign ANx as an analog input and only that pin will be treated as such and the only one you will have to worry about. Put a third way, the digital I/O assignments for outputs override the analog input settings. Joe M. peter green wrote: {Quote hidden}> MCH wrote: >> Am I reading the PIC 18F4520 docs correctly that you can only assign >> analog inputs in ascending order? IOW, if I want only three analog >> inputs, I have to use pins AN0, AN1, and AN2 and I cannot make the >> selection based on open pins such as AN3, AN8, and AN9? >> >> Put another way, if I want AN5 to be analog, AN4-AN0 has to be analog as >> well. >> > Yes and no > > Yes you can only put pins into analog mode in order. However afaict the > only thing that putting a pin in analog input mode does is disable the > digitial input buffers. > > So you can still use a pin as a digital output when it's set as analog > (this was very useful when I had to drive a resistive touchscreen) and > if you are prepared to put up with the extra power consumption caused by > having a digital input floating between the rails you can use a pin as > an analog input when it's set as digitial. > > > ------------------------------------------------------------------------ > > > No virus found in this incoming message. > Checked by AVG - http://www.avg.com > Version: 9.0.733 / Virus Database: 271.1.1/2684 - Release Date: 02/12/10 14:35:00 >

ivp wrote: >> The series resistor is not going to change the voltage >> seen by the PIC input and thus will be no help. > > Yes, my bad. There will be a range of analogue voltage > which causes the input FETs to not commit to a '0' or '1' > value and even try to turn on together, making a short > between the rails Except for Schmitt Trigger inputs that are likely to be discussed in this thread. Input wires within the package are of some resistance, so the Schmitt trigger feedback would not let the input to stay stably at that intermediate level.

MCH wrote: > So, you can only assign analog inputs in ascending order, but the only > thing that will 'affect' is when the 'lower' pins are inputs. Almost, but not quite. > IOW, if the lower pins are all outputs, you can assign ANx as an > analog input and only that pin will be treated as such and the only > one you will have to worry about. > > Put a third way, the digital I/O assignments for outputs override the > analog input settings. The input of the digital input latch that is associated with the output (and that is functional even when a pin is set as an output) is also affected. Check out the port block diagrams in the data sheet. They are quite helpful for understanding how this works together. There is no way to disable the input circuitry; there is only a way to disable the output drivers (the TRIS bits). Configuring a pin as analog makes the digital input latch always read 0, and that's it, basically. If you don't read it while it is set to analog, there shouldn't be any other consequences. Gerhard

>> Put a third way, the digital I/O assignments for outputs override the >> analog input settings. That is misleading. When ANx is enabled and TRISxn low (digital output), you can write to the pin *and then read the actual (analog) level using the ADC*. -- Wouter van Ooijen -- ------------------------------------------- Van Ooijen Technische Informatica: http://www.voti.nl consultancy, development, PICmicro products docent Hogeschool van Utrecht: http://www.voti.nl/hvu

But if you write to that pin, the analog voltage present from what you're reading would interfere with the output, wouldn't it? Joe M. Wouter van Ooijen wrote: {Quote hidden}>>> Put a third way, the digital I/O assignments for outputs override the >>> analog input settings. > > That is misleading. When ANx is enabled and TRISxn low (digital output), > you can write to the pin *and then read the actual (analog) level using > the ADC*. > > > > ------------------------------------------------------------------------ > > > No virus found in this incoming message. > Checked by AVG - http://www.avg.com > Version: 9.0.733 / Virus Database: 271.1.1/2692 - Release Date: 02/16/10 14:35:00 >

>> When ANx is enabled and TRISxn low (digital output), >> you can write to the pin *and then read the actual (analog) level using >> the ADC*. > But if you write to that pin, the analog voltage present from what > you're reading would interfere with the output, wouldn't it? You can't successfully present analog voltage to a pin that is an output. What you are reading with the ADC is how well the digital output drivers are doing their work against the load. -- Wouter van Ooijen -- ------------------------------------------- Van Ooijen Technische Informatica: http://www.voti.nl consultancy, development, PICmicro products docent Hogeschool van Utrecht: http://www.voti.nl/hvu