dsPIC ADC help needed

[Nathan House-5]

Hi,

I'm struggling trying to get the analog input module on my dsPIC30F4011
properly configured. Eventually I need to be able to sample two channels
simultaneously at 100Hz, continuously. But right now I'd be very happy with
getting *anything* working.

The configuration code I have so far is posted below. I'm stuck on how to
calculate the ADCS & SAMC registers correctly. I've been trying to calculate
these for quite some time, but for some reason just can't get it right! I'm
running my PIC at 10MHz, with no PLL. If anyone could provide me with some
insight, on how to calculate ADCS and SAMC, I'd really appreciate it.


     ADCON1bits.ADON  = 0; // Turn OFF A/D module

     // B4 (pin 6) - Accelerometer
     ADPCFGbits.PCFG4 = 0; // Set pin B4 to analog mode
     TRISBbits.TRISB4 = 1; // Set pin B4 as INPUT

     ADCON1bits.FORM = 0;   // Buffer output = Integer
     ADCON1bits.ADSIDL = 1; // Stop module in idle mode
     ADCON1bits.SIMSAM = 1; // Sample CH0 .. CH3 Simultaneously
     ADCON1bits.ASAM = 1;   // Auto sample after convert
     ADCON1bits.SSRC = 7;   // Internal counter ends sampling and starts
conversion (auto convert)

     ADCON2bits.VCFG = 0;   // Voltage reference is AVdd and AVss (internal)
     ADCON2bits.CHPS = 2;   // Converts CH0 .. CH3 (inputs are actually only
on 2 & 3)
     ADCON2bits.ALTS = 0;   // Use MUX A configuration
     ADCON2bits.BUFM = 0;   // Use single buffer, 16 word

     ADCON3bits.ADRC = 0;   // Use system clock
     //ADCON3bits.ADCS =    // Conversion clock. 4011 data sheet example
20-1
     //ADCON3bits.SAMC =   // Time between start of sample, and start of
conversion

     ADCHSbits.CH123SA = 1; // Positive input for CH1 is AN3, CH2 is AN4,
CH3 is AN5
     ADCHSbits.CH123NA = 0; // Negative input for CH1, CH2, and CH3 is Vref-
     ADCSSLbits.CSSL4  = 1; // Include AN4 in scan

     ADCON1bits.ADON  = 1; // Turn ON A/D module


Thanks!

Nathan
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Olin Lathrop]

Nathan House wrote:
> I'm struggling trying to get the analog input module on my
> dsPIC30F4011 properly configured.

Then read the manual again, carefully, it's all in there.  That A/D is
complicated compared to other PIC A/Ds, but I've always found it to work as
described.

> The configuration code I have so far is posted below.

I have attached the initialization subroutine for the A/D module of a
30F3010 project that samples two analog input signal simultaneously at
100KHz each.


--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Nathan House-5]

I don't understand assembler, but thanks for helping Olin.

-Nathan


--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Jan-Erik Soderholm]

Nathan House wrote:

> I don't understand assembler,...

I'd concider that a major problem that probably
should be "fixed".

Aren't most/all examples in the datasheets in assmebler ?
At least for the PIC16/PIC18 they are...

Jan-Erik.



>
> -Nathan
>
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Olin Lathrop]

Nathan House wrote:
> I don't understand assembler,

Then learn it.  You need to understand what goes on at the low level on
these small resource limited systems, else you will be forever bumping into
strange symptoms with little chance of figuring them out.  If you think a
compiler will just take care of everything for you, then you frankly don't
belong here.  Go back to Java, web apps, or whatever, and leave tiny systems
to those willing to take the time to understand them.

In any case it should be reasonably obvious what is being stuffed into what
register for initialization.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
[phone].  Gold level PIC consultants since 2000.
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Nathan House-5]

 >Then learn it.

OK, I will :-) Do you have any recommendations on how I should get started
(maybe a book you've found useful, or a website..)?

>You need to understand what goes on at the low level on
>these small resource limited systems, else you will be forever bumping into
>strange symptoms with little chance of figuring them out.

Makes sense.. I decided not to learn assembler, to jump right into
microcontroller programming using C because *I thought* assembly programming
was something of the past, something that was being phased out. Thank you
for correcting my line of thinking. If I take a time-out and learn assembler
now, it sounds like I'll be saving myself many headaches in the long run.

>If you think a compiler will just take care of everything for you, then you
frankly don't
>belong here.  Go back to Java, web apps, or whatever, and leave tiny
systems
>to those willing to take the time to understand them.

I "think" that I'm a senior in high school, self-taught when it comes to
programming and electronics, with very little idea of what to do. I'm
learning as I go, and asking for help when I need it.

Nathan
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Jan-Erik Soderholm]

Nathan House wrote:

> Makes sense.. I decided not to learn assembler, to jump right into
> microcontroller programming using C because *I thought* assembly
> programming was something of the past,...

I do not think that either I or Olin wrote that you actualy
should *program* in assembler, only that it is good to be
able to *read* assembler code... :-)

Jan-Erik.

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Olin Lathrop]

Nathan House wrote:
> OK, I will :-) Do you have any recommendations on how I should get
> started (maybe a book you've found useful, or a website..)?

First you learn the instruction set, then you learn how the assembler lets
you express that instruction set in the source code.  There are probably
books out there that cover the general concepts, but I've never read any of
them.  If you're interested in dsPICs, start with the dsPIC Programmer's
Reference Guide.

> I "think" that I'm a senior in high school, self-taught when it comes
> to programming and electronics, with very little idea of what to do.
> I'm learning as I go, and asking for help when I need it.

It's a lot easier now with the internet and access to computers basically
free.  I learned my first machine language as a junior in high school in
1973 from the DEC Small Computers Handbook and the "orange" book.  The
machine was a PDP-8, which is sortof like a PIC 12F in a rack mounted box
with toggle switches to load the program in.  There was nobody around I
could ask either.  Fortunately the documentation today about PICs is just as
good as the documentation on a PDP-8 was back then.  Just dig in.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
[phone].  Gold level PIC consultants since 2000.
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Olin Lathrop]

Jan-Erik Soderholm wrote:
> I do not think that either I or Olin wrote that you actualy
> should *program* in assembler, only that it is good to be
> able to *read* assembler code... :-)

You're right, I didn't write that, but I do think it's a good idea to do the
first project or two in assembler since that's the only way to really learn
it and the low level details that go with it.  After that you can mix and
match languages as you see fit as then you'll have the knowledge to know
which language is appropriate for the situation.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
[phone].  Gold level PIC consultants since 2000.
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Vitaliy-2]

Olin Lathrop wrote:

>> I do not think that either I or Olin wrote that you actualy
>> should *program* in assembler, only that it is good to be
>> able to *read* assembler code... :-)
>
> You're right, I didn't write that, but I do think it's a good idea to do
> the
> first project or two in assembler since that's the only way to really
> learn
> it and the low level details that go with it.  After that you can mix and
> match languages as you see fit as then you'll have the knowledge to know
> which language is appropriate for the situation.

I think it's more true for the lower-end PICs. I've written thousands of
lines of C code for the 18F and the PIC24, but not a single line of
assembly.

So it's probably a good idea, but certainly not a requirement.

Vitaliy

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Nathan House-5]

Would someone mind explaining how to calculate Tad, ADCS and SAMC?

I'm still stuck trying to get this to work and would really appreciate some
help.

Thanks.
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Rikard Bosnjakovic-2]

On Mon, Jun 22, 2009 at 16:43, Nathan House<nathanpiclist@...> wrote:

> Would someone mind explaining how to calculate Tad, ADCS and SAMC?

Never done dsPICs myself, but perhaps Table 17-9 of the dsPIC30
reference manual (DS70046,
http://ww1.microchip.com/downloads/en/DeviceDoc/70046E.pdf) can be of
any help.


--
- Rikard - http://bos.hack.org/cv/
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Nathan House-5]

Thanks for the suggestion, but I only need to sample at 100Hz and the table
doesn't show a Tad value for that slow of a sampling rate.

That table has always confused me. Would anyone know why the slower the
sampling rate, the higher the minimum Tad time requirement? If I understand
correctly, Tad represents the amount of time that it takes to convert one
bit. Shouldn't it take the same amount of time to convert the bits,
regardless of how fast you're sampling?
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Olin Lathrop]

Nathan House wrote:
> That table has always confused me. Would anyone know why the slower
> the sampling rate, the higher the minimum Tad time requirement? If I
> understand correctly, Tad represents the amount of time that it takes
> to convert one bit. Shouldn't it take the same amount of time to
> convert the bits, regardless of how fast you're sampling?

Yes, as long as all bits get done in time to start the acqusition of the
next value.  There is a valid range for Tad where everything else the
datasheet says about the A/D still applies.  If Tad is too long, then the
sample and hold can drift.  If Tad is too short, then things don't settle in
time to decide what the bit value should be.

Screw the table.  That's only for people that flunked 5th grade arithmatic.
Keep Tad within the specified limits (as with all datasheet parameters), and
everything will be OK.


********************************************************************
Embed Inc, Littleton Massachusetts, http://www.embedinc.com/products
[phone].  Gold level PIC consultants since 2000.
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Nathan House-5]

Is there a recommended Tad time to use? Should a different sampling time be
used on different devices? For example, should I use a different sampling
time when using a potentiometer then on an accelerometer?

The FRM says that Tad should be a minimum of 83.33 ns. But there's nothing
in there (at least that I've found) about a maximum Tad time, or a
recommended Tad time.

Appreciate the help!
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Vicent Colomar Prats]

Minimum means that time is enought to acomplish a conversion, so the more
Tad time you give, the more time you are wasting. So, minimum should be also
recommended in most cases, but that depends on how you do conversion and if
it's a speed critical aplication.

2009/6/29 Nathan House <nathanpiclist@...>

> Is there a recommended Tad time to use? Should a different sampling time be
> used on different devices? For example, should I use a different sampling
> time when using a potentiometer then on an accelerometer?
>
> The FRM says that Tad should be a minimum of 83.33 ns. But there's nothing
> in there (at least that I've found) about a maximum Tad time, or a
> recommended Tad time.
>
> Appreciate the help!
> --
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Nathan House-5]

>Minimum means that time is enought to acomplish a conversion, so the more
>Tad time you give, the more time you are wasting. So, minimum should be
also
>recommended in most cases, but that depends on how you do conversion and if
>it's a speed critical aplication.

So if I get an invalid value calculating the ADCS value using the minimum
Tad time, should I just keep raising the Tad time until I get a valid
answer?

Equation 17-1 in the 30F family reference manual:

 <http://roboticsguy.com/equation/rendered/images/70046E_equation_17-1.png>

Solving for ADCS with my instruction time and the minimum Tad value:

[image:
http://roboticsguy.com/images/misc/70046E_equation_17-1_solving_for_ADCS.png]

Which equals: *-0.58335*

If I make Tad 400ns then ADCS would equal 1. Is that what I should do?

Thanks :)
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Vicent Colomar Prats]

If you are not sure about ad times giving you correct values, you can make
an experiment:
Make some conversions changing tad every time and then display the results
of conversions.
I did that way when I wanted to insure correct values at different times. An
example of this in CCS C is:

unsigned int16 j;
char valor[]="    ";
unsigned int8 p1,p2,p3,p4;

   setup_adc_ports(sAN0|sAN1|sAN2|sAN3|sAN4|VSS_VDD);
   for(j=0;j<1000;j++) {
      setup_adc(ADC_CLOCK_INTERNAL);
      setAN(0);
      delay_us(4);
      p1 = leeAN;
      setup_adc(ADC_CLOCK_DIV_32);
      setAN(0);
      delay_us(4);
      p2 = leeAN;
      setup_adc(ADC_CLOCK_DIV_8);
      setAN(0);
      delay_us(4);
      p3 = leeAN;
      setup_adc(ADC_CLOCK_DIV_2);
      setAN(0);
      delay_us(4);
      p4 = leeAN;

      itoa(p1,10,valor);
      LCD_CursorYX(2,1);
      LCD_Print(valor);
      itoa(p2,10,valor);
      LCD_CursorYX(2,6);
      LCD_Print(valor);
      itoa(p3,10,valor);
      LCD_CursorYX(2,11);
      LCD_Print(valor);
      itoa(p4,10,valor);
      LCD_CursorYX(2,16);
      LCD_Print(valor);

      delay_ms(100);
   }

Take into account, it is not exactly the same you are talking about, as I
use C and fixed delays, but it is the most similar example of my code I
found.
It changes the AD clock and, thru a display, you can check wich values are
different, so wich is giving you incorrect values due to a faster clock.

2009/7/3 Nathan House <nathanpiclist@...>

> >Minimum means that time is enought to acomplish a conversion, so the more
> >Tad time you give, the more time you are wasting. So, minimum should be
> also
> >recommended in most cases, but that depends on how you do conversion and
> if
> >it's a speed critical aplication.
>
> So if I get an invalid value calculating the ADCS value using the minimum
> Tad time, should I just keep raising the Tad time until I get a valid
> answer?
>
> Equation 17-1 in the 30F family reference manual:
>
>  <http://roboticsguy.com/equation/rendered/images/70046E_equation_17-1.png
> >
>
> Solving for ADCS with my instruction time and the minimum Tad value:
>
> [image:
>
> http://roboticsguy.com/images/misc/70046E_equation_17-1_solving_for_ADCS.png
> ]
>
> Which equals: *-0.58335*
>
> If I make Tad 400ns then ADCS would equal 1. Is that what I should do?
>
> Thanks :)
> --
> http://www.piclist.com PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Alan B. Pearce-2]

> So if I get an invalid value calculating the ADCS value using
> the minimum Tad time, should I just keep raising the Tad time
> until I get a valid answer?

Can you be sure that there is not something else affecting the accuracy? I
haven't checked the datasheet, but IIRC these devices require a pretty low
source impedance for the analog signal, something like <2.5k. If you are not
meeting this criteria, and whatever other criteria may be given for the
analog signal, then it doesn't matter how long you make the Tad, you will
not get the right answer.

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist

[Nathan House-5]

When I said "should I just keep raising the Tad time until I get a valid
answer," I meant a valid answer for the ADCS equation, not a result from an
actual analog input. I have a potentiometer hooked up to an analog port to
make the testing easy, but I would prefer to know how to properly configure
the module, as opposed to just experimenting and just sticking with some
random value that seems to work.
--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist