Searching \ for '[PIC] Increasing ADC Accuracy' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/microchip/ios.htm?key=adc
Search entire site for: 'Increasing ADC Accuracy'.

Exact match. Not showing close matches.
PICList Thread
'[PIC] Increasing ADC Accuracy'
2005\10\06@001153 by David de Regt

flavicon
face
I'm using a PIC16F88 with an ADXL322 board for accelerometer readings
(yes, I have the suggested filtering capacitors in place.)


I'm trying to figure out how to get maximum accuracy out of this.  Right
now, the best I can seem to do is average over 100-ish ADC samples,
using the $80 flag on ADCON0.  Even this, however, is still getting +/-
7 or so accuracy on the ADC-ing.


Is there any way to make the ADC more accurate, or are the ADCs on the
PICs just fairly crappy?


Thanks,

David

2005\10\06@004143 by Jinx

face picon face
> are the ADCs on the PICs just fairly crappy ?

They're OK

To get the best out of them you need quiet supply rails - Vcc,
Vdd and Vref. Depending on other PIC I/O activities you
might want to time those away from sampling. Use the greatest
resolution available, eg use 5.12V for Vref (Vcc 5.12V or a
little higher) to get 4mV / bit or use an op amp to amplify a
narrower signal range (this would also meet the impedance
spec if the sensor is hi-Z). Although some time ago there was
a thread about higher impedance signals working with longer
sampling times


2005\10\06@010614 by Chen Xiao Fan

face
flavicon
face
Most likely it is not that PIC ADCs are crappy but the analog
circuits or layout is crappy.

Microchip AN688B (Layout Tips for 12-Bit A/D Converter
Applications) may be a good reference for you even though it
dealing with an standalone A/D converter.

The environment of the circuit boards can play a big part
as well. If you do not have a proper layout, the noise
can affect the ADC result by a lot.

Regards,
Xiaofan

{Original Message removed}

2005\10\06@015817 by Wouter van Ooijen

face picon face
> I'm using a PIC16F88 with an ADXL322 board for accelerometer readings
> (yes, I have the suggested filtering capacitors in place.)

Which value are the caps?

And I dunno the 322, but with the 311 the internal resistors are far
larger than the recommended (required?) source impedance for the F88 A/D
(either 10k or 2k5, depends on the part of the datasheet you take).

Wouter van Ooijen

-- -------------------------------------------
Van Ooijen Technische Informatica: http://www.voti.nl
consultancy, development, PICmicro products
docent Hogeschool van Utrecht: http://www.voti.nl/hvu


2005\10\06@073419 by Vasile Surducan

face picon face
On 10/6/05, Wouter van Ooijen <spam_OUTwouterTakeThisOuTspamvoti.nl> wrote:
> > I'm using a PIC16F88 with an ADXL322 board for accelerometer readings
> > (yes, I have the suggested filtering capacitors in place.)
>
> Which value are the caps?
>
> And I dunno the 322, but with the 311 the internal resistors are far
> larger than the recommended (required?) source impedance for the F88 A/D
> (either 10k or 2k5, depends on the part of the datasheet you take).

The biggest shit with PIC's AD is the missing of a good internal
layout for the analogic and digital ground (and analogic and digital
+VDD supply). Everything you can do, is to mix in a convenient way
these two, but outside the PIC package.
If you have played a little with Maxims high speed AD converters you
know how important is to keep the analog voltage slighty above the
digital one.

Vasile

2005\10\06@092716 by Aaron

picon face
Vasile Surducan wrote:

>If you have played a little with Maxims high speed AD converters you
>know how important is to keep the analog voltage slighty above the
>digital one.
>
>  
>

Can you clarifiy this??

Aaron

2005\10\06@193527 by David de Regt

flavicon
face
The caps are 0.1uF caps.

Yesterday evening, I tried using a power supply to provide fixed
voltage, and the ADC stayed spot-on, so I'm thinking you might be right
about the source impedance of the 322.  

So what would be a way around that?  Use an opamp voltage follower?

-David

{Original Message removed}

2005\10\07@052119 by Vasile Surducan

face picon face
On 10/6/05, Aaron <.....aaron.piclistKILLspamspam@spam@gmail.com> wrote:
> Vasile Surducan wrote:
>
> >If you have played a little with Maxims high speed AD converters you
> >know how important is to keep the analog voltage slighty above the
> >digital one.
> >
> >
> >
>
> Can you clarifiy this??

Yes, a serious AD converter have both analog supply pins entirely
separate by the digital supply pins. GND and Vcc. To get a reasonable
SNR at high speed conversion, those two supplies are separed and
usually the analogic voltage supply(3V) is bigger than the digital
voltage supply (2.5V).
The values are just for example, on thickest technology chips are
lower (0.9V-1.8V).

Vasile

2005\10\07@064339 by Aaron

picon face
Vasile Surducan wrote:

{Quote hidden}

Oh, the analog supply....got it!
I was trying to figure out how to keep the analog voltage one was
trying to measure above your digital voltage...

Aaron

2005\10\07@094830 by Gerhard Fiedler

picon face
David de Regt wrote:

>> And I dunno the 322, but with the 311 the internal resistors are far
>> larger than the recommended (required?) source impedance for the F88
>> A/D (either 10k or 2k5, depends on the part of the datasheet you take).

> Yesterday evening, I tried using a power supply to provide fixed voltage,
> and the ADC stayed spot-on, so I'm thinking you might be right about the
> source impedance of the 322.
>
> So what would be a way around that?  Use an opamp voltage follower?

That's what's usually done. While you're at it, you can also use the opamp
also to scale and/or active low-pass filter the signal (if required).

Gerhard

2005\10\07@121559 by David de Regt

flavicon
face
After cleaning up my power supply with the many capacitors and using a
better voltage source, the ADC cleaned itself up dramatically.  I
dropped my number of ADCs to average over to 10 instead of 100, and it's
now always right +/- 0.005g or so, which is all I can ask for.

-David

{Original Message removed}

More... (looser matching)
- Last day of these posts
- In 2005 , 2006 only
- Today
- New search...