accuracy of the LT1298

Archiver

Hello,

Since I got the suggestion of using this handy little AD from you fine
folks, I'm wondering if anybody has any reports on it's accuracy. I
have good reason to think that my voltage samples taken with this chip
are fluctuating randomly with upwards of 6 bits worth of noise. I
have measured the noise on the signal lines to the ADC with a scope
and it is certainly not that high.

Can anyone with experience using the IC comment on the accuracy? Have
people successfully received repeated, accurate 12 bit data from the
device?

Thanks for your time.

Archiver

Without more information about your setup, nobody is going to be able to
target specific answers at ya.

Generally, remember that there is more than one voltage source at play with
an ADC. Not only must the signal be clean, so must the reference voltage AND
the ADC's digital supply. Noise on either of these can result in bouncy
readings.

Also at play is the sampling rate. How often are you asking the ADC for a
value? Can the ADC supply the answers at that rate?

What about conversion times? If you're single converting, are you giving it
enough time to complete a conversion before asking it for a value?

What else can you describe about your ADC circuit?



>

---

Original Message

---

> From: George Bush [noparse]/noparse]mailto:[url=http://forums.parallaxinc.com/group/basicstamps/post?postID=hyj8pPJ4Ydtg6WLBHwA8aud-49vdxf_l_4JLePXrX99ED7UQ6NeGbclIRuIdliGylRz1uOkKd5HPE5zT]capthoohoo@y...[/url
> Sent: October 15, 2003 5:34 PM
> To: basicstamps@yahoogroups.com
> Subject: [noparse][[/noparse]basicstamps] accuracy of the LT1298
>
>
> Hello,
>
> Since I got the suggestion of using this handy little AD from you fine
> folks, I'm wondering if anybody has any reports on it's accuracy. I
> have good reason to think that my voltage samples taken with this chip
> are fluctuating randomly with upwards of 6 bits worth of noise. I
> have measured the noise on the signal lines to the ADC with a scope
> and it is certainly not that high.
>
> Can anyone with experience using the IC comment on the accuracy? Have
> people successfully received repeated, accurate 12 bit data from the
> device?
>
> Thanks for your time.
>
>
>
> To UNSUBSCRIBE, just send mail to:
> basicstamps-unsubscribe@yahoogroups.com
> from the same email address that you subscribed. Text in the
> Subject and Body of the message will be ignored.
>
>
> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/
>
>
>

Archiver

>Hello,
>
>Since I got the suggestion of using this handy little AD from you fine
>folks, I'm wondering if anybody has any reports on it's accuracy. I
>have good reason to think that my voltage samples taken with this chip
>are fluctuating randomly with upwards of 6 bits worth of noise. I
>have measured the noise on the signal lines to the ADC with a scope
>and it is certainly not that high.
>
>Can anyone with experience using the IC comment on the accuracy? Have
>people successfully received repeated, accurate 12 bit data from the
>device?
>
>Thanks for your time.


There is no reason that you should have that much noise fluctuation.
Check you power supplies. Be sure the wiring to the LTC1298 Vss is
neat and not hanging out all over the place. Put a bypass capacitor
right across the power supply at the chip. Put resistors of 1k in
the data lines between the Stamp and the LTC1298.

I have had good stable results from that chip. There is one subtle
point you should be aware of, if you are using it with a BS2. This
was the subject of a long thread a couple of years ago as we were
trying to determine why a full scale input could never give a count
of 4095. The LTC1298 AD conversion is timed by the stamp SHIFTOUT
clock line, and at about 16 kilohertz, it is just a little too
slow--The signal voltage droops on the internal sampling capacitors.
There is no such problem with the BS2sx or BS2p, which run the
SHIFTOUT clock much faster.

-- Tracy

Archiver

--- In basicstamps@yahoogroups.com, Tracy Allen <tracy@e...> wrote:
> >Hello,
> >
> >Since I got the suggestion of using this handy little AD from you
fine
> >folks, I'm wondering if anybody has any reports on it's accuracy.
I
> >have good reason to think that my voltage samples taken with this
chip
> >are fluctuating randomly with upwards of 6 bits worth of noise. I
> >have measured the noise on the signal lines to the ADC with a scope
> >and it is certainly not that high.
> >
> >Can anyone with experience using the IC comment on the accuracy?
Have
> >people successfully received repeated, accurate 12 bit data from
the
> >device?
> >
> >Thanks for your time.
>
>
> There is no reason that you should have that much noise fluctuation.
> Check you power supplies. Be sure the wiring to the LTC1298 Vss is
> neat and not hanging out all over the place. Put a bypass capacitor
> right across the power supply at the chip. Put resistors of 1k in
> the data lines between the Stamp and the LTC1298.
>
> I have had good stable results from that chip. There is one subtle
> point you should be aware of, if you are using it with a BS2. This
> was the subject of a long thread a couple of years ago as we were
> trying to determine why a full scale input could never give a count
> of 4095. The LTC1298 AD conversion is timed by the stamp SHIFTOUT
> clock line, and at about 16 kilohertz, it is just a little too
> slow--The signal voltage droops on the internal sampling capacitors.
> There is no such problem with the BS2sx or BS2p, which run the
> SHIFTOUT clock much faster.
>
> -- Tracy

You and I both agree there is no reason to see that much fluctuation.
The power supply lines are at least as clean as than the signal lines,
and the chip has been placed on a circuit board so the wiring is about
as neat as it will get. There is a bypass capacitor and 220 Ohm
resistors in the data lines. Is there a reason you suggest 1k or is
that just a more conservative resistor to protect the Stamp pins from
over-current? I'm using a BS2p. I only retrieve a value from the
chip every .5 seconds.