Forums

Programming AD7708 with LPC2000

Started by jdauchot November 15, 2007
Hi All

I Know it is not an LPC2000 problem, but has anyone used an Analog
Divices AD7708? . I apply a steady DC volatge to one of the inputs and
the reading decays while reading the ADC. When I very the input up and
down the reading comes back up and decays.

Can anyone help

Regards

Jean-Jacques

An Engineer's Guide to the LPC2100 Series

--- In l..., "jdauchot" wrote:
>
> Hi All
>
> I Know it is not an LPC2000 problem, but has anyone used an Analog
> Divices AD7708? . I apply a steady DC volatge to one of the inputs
and
> the reading decays while reading the ADC. When I very the input up
and
> down the reading comes back up and decays.
>
> Can anyone help
>
> Regards
>
> Jean-Jacques
>
>>>PLs try Pulldown / Pullup resistor from ADC o/p to ground .
I have used Pull down of value 4.7k .with TI serial ADC with XAG-49
controller .but phenomenon is same as u observed .
pls try .thanks good luck .
SatheBV
Hi Jean-Jacques,

> I Know it is not an LPC2000 problem, but has anyone used an Analog
> Divices AD7708? . I apply a steady DC volatge to one of the inputs and
> the reading decays while reading the ADC. When I very the input up and
> down the reading comes back up and decays.

I haven't used the part. But I plan to use an AD7792 which looks
to be similar.

You haven't really provided enough info.

1. What is the circuit driving the input?

2. What DC voltage?

If you read the data sheet, it says there are programmable
voltage ranges; +/-20mV to +/-2.56V.

3. Does this problem occur on multiple boards?

Cheers,
Dave
Hi Dave

Thanks for reply

>>f you are measuring a unipolar signal with reference to gnd...i.e the
AINCOM is connected to GND, have you cleared the NEGBUF bit ( set mode
register >>to unbuffered )? this will allow you to bypass the input buffer
and alows signals reference to AGND to be measured...refer to page 42 of the
datasheet >>"converting single-ended inputs".

This is the reply I got from Analog Devices. Does that make sense as I am
only a Software Engineer ?

The DC voltage is from a load cell amplifier which delivers 0 to 2V dc. It
measure tank weight on ma feeding system on farms

Regards

Jean-Jacques

Hi Jean-Jacques,

> I Know it is not an LPC2000 problem, but has anyone used an Analog
> Divices AD7708? . I apply a steady DC volatge to one of the inputs and
> the reading decays while reading the ADC. When I very the input up and
> down the reading comes back up and decays.

I haven't used the part. But I plan to use an AD7792 which looks
to be similar.

You haven't really provided enough info.

1. What is the circuit driving the input?

2. What DC voltage?

If you read the data sheet, it says there are programmable
voltage ranges; +/-20mV to +/-2.56V.

3. Does this problem occur on multiple boards?

Cheers,
Dave
Hi Jean-Jacques,

>> If you are measuring a unipolar signal with reference to gnd...i.e the
>> AINCOM is connected to GND, have you cleared the NEGBUF bit ( set mode
>> register to unbuffered )? this will allow you to bypass the input buffer
>> and alows signals reference to AGND to be measured...refer to page 42 of the
>> datasheet "converting single-ended inputs".
>
> This is the reply I got from Analog Devices. Does that make sense as I am
> only a Software Engineer ?
>
> The DC voltage is from a load cell amplifier which delivers 0 to 2V dc. It
> measure tank weight on ma feeding system on farms

Basically the AD response is asking if you have configured the
part - in software - before probing signals on the board.

So, it sounds like you need to setup the part for unipolar mode.
Read the datasheet for the details.

I'm assuming you have the common point connected to ground?

Feel free to send me your schematic page, and I can see if
there is anything weird looking.

Is there anything special about your load cell amplifier that
could be causing this? Have you tried replacing the input
with a lab supply set to say 1V? That way you can eliminate
it as the source of the problem.

That's what a hardware guy would do anyway ;)

Cheers,
Dave
> A real hardware guy wouldn't bother with a power supply when a AA
> battery was laying on the table complete with battery holder. It
> would take to long to dig the test leads out from under the other stuff.
>
> Still, to protect the device, connect a 1k resistor in series with the
> battery. Measurement accuracy isn't the goal at this point.

Ah, but a real hardware guy would have too much stuff on his desk
to see the battery

:)
> That's what a hardware guy would do anyway ;)
>
> Cheers,
> Dave
>

A real hardware guy wouldn't bother with a power supply when a AA
battery was laying on the table complete with battery holder. It
would take to long to dig the test leads out from under the other stuff.

Still, to protect the device, connect a 1k resistor in series with the
battery. Measurement accuracy isn't the goal at this point.

Richard
urement accuracy isn't the goal at this point.
>
> Ah, but a real hardware guy would have too much stuff on his desk
> to see the battery
>

Pour a bagfull of CR2032's over your desk. That way, you'll always be
able to find one quickly when you need to turn on an input port with a
couple damp fingers and a battery.

Rgds,
Martin