A/D successive channel sampling Tacq requirements.

Started by demolitron May 4, 2009
Hello,

I am looking at doing some fast ADC conversions and notice that the data sheets specify that CHold is not discharged between samples. So if I leave it on the same channel does that mean that the Tacq is much lower?

It would depend on how much CHold droops during the conversion. Does anyone know the details of this? I am looking at doing this on a PIC16F877 and also a dsPIC30F3013 device.

Thanks!

No I don't think so... You are not going to sample a static voltage at a
high sample freq ?
If you disconnect the analogue signal the charge starts to bleed off.
The specs is drawn out for a source Z of 2.5K if I remember correctly,
so the best way
to reduce Tacq for a fast moving signal may be to drive the pin from a
low Z source (follower)...
should make the S&H cap charge much faster

demolitron skrev:
> Hello,
>
> I am looking at doing some fast ADC conversions and notice that the
> data sheets specify that CHold is not discharged between samples. So
> if I leave it on the same channel does that mean that the Tacq is much
> lower?
>
> It would depend on how much CHold droops during the conversion. Does
> anyone know the details of this? I am looking at doing this on a
> PIC16F877 and also a dsPIC30F3013 device.
>
> Thanks!

--
*******************************************
VISIT MY HOME PAGE:

LAST UPDATED: 23/08/2003
*******************************************
Regards
Eirik Karlsen
----- Original Message -----
From: "demolitron"
To:
Sent: Monday, May 04, 2009 7:32 PM
Subject: [piclist] A/D successive channel sampling Tacq requirements.
> I am looking at doing this on a PIC16F877 and also a dsPIC30F3013 device.

Some of the 16 bit devices have multiple A/D converters, and the
multiplexers can be programmed to alternate between converters, allowing
much faster sampling rates. I think most of the 16 bitters can do 500k
samples/sec, with at least one going up to 2M samples/sec.

The multiplexer on the 16 bit parts can walk and talk while it's chewing
gum, so it is a little tricky to set up. But once it is initialized -- wow!
It does require you to suspend disbelief while you read the datasheet,
however.

72/73 de WB8RCR http://www.qsl.net/wb8rcr
didileydadidah QRP-L #1446 Code Warriors #35

I was looking at experimenting with Oversampling and decimation to increase the ENOB. From what I've read and understand in that case you do sample a relatively slow changing signal very quickly. Plus, even with a low bandwidth signal you'd want to sample higher than the nyquist frequency of the filter before your ADC, right? So you can software filter out the largest components of noise...

I've always found that I wasn't able to get a "stable" 10-bit result let alone 12-bit. But when I put a simple RC filter before the input and sampled at 2x-4x the cutoff I was able to get 10-bit resolution, even with a 25-turn cermet pot as the input.
--- In p..., Eirik Karlsen wrote:
>
> No I don't think so... You are not going to sample a static voltage at a
> high sample freq ?
> If you disconnect the analogue signal the charge starts to bleed off.
> The specs is drawn out for a source Z of 2.5K if I remember correctly,
> so the best way
> to reduce Tacq for a fast moving signal may be to drive the pin from a
> low Z source (follower)...
> should make the S&H cap charge much faster
>
> demolitron skrev:
> >
> >
> > Hello,
> >
> > I am looking at doing some fast ADC conversions and notice that the
> > data sheets specify that CHold is not discharged between samples. So
> > if I leave it on the same channel does that mean that the Tacq is much
> > lower?
> >
> > It would depend on how much CHold droops during the conversion. Does
> > anyone know the details of this? I am looking at doing this on a
> > PIC16F877 and also a dsPIC30F3013 device.
> >
> > Thanks!
> >
> > --
> *******************************************
> VISIT MY HOME PAGE:
>
> LAST UPDATED: 23/08/2003
> *******************************************
> Regards
> Eirik Karlsen
>

I just got my delivery of some dsPIC30's today. I'm like a kid in a candy shop. And yes, the ADC's on these seem to be magical in comparison to the 16F stuff. I've got some frustration ahead of me trying to learn all the new wiz'bang features...
--- In p..., "John J. McDonough, WB8RCR" wrote:
> ----- Original Message -----
> From: "demolitron"
> To:
> Sent: Monday, May 04, 2009 7:32 PM
> Subject: [piclist] A/D successive channel sampling Tacq requirements.
> > I am looking at doing this on a PIC16F877 and also a dsPIC30F3013 device.
>
> Some of the 16 bit devices have multiple A/D converters, and the
> multiplexers can be programmed to alternate between converters, allowing
> much faster sampling rates. I think most of the 16 bitters can do 500k
> samples/sec, with at least one going up to 2M samples/sec.
>
> The multiplexer on the 16 bit parts can walk and talk while it's chewing
> gum, so it is a little tricky to set up. But once it is initialized -- wow!
> It does require you to suspend disbelief while you read the datasheet,
> however.
>
> 72/73 de WB8RCR http://www.qsl.net/wb8rcr
> didileydadidah QRP-L #1446 Code Warriors #35
>