WEBENCH® Tools/ADS1248: 4 RTD (PT100) connected to the ADS1248 ?

Reelf,

It is possible to connect 4 2-wire PT100 RTDs to the ADS1248. I show a basic diagram below:

For each measurement, you would set the IDAC current to the top end of the RTD that you'd want to measure.

However, in most industrial environments, most will use some sort input filtering to reduce any EMI that may get picked up. In most designs I've seen, there will be series resistances with common-mode and differential capacitances to act as an RC filter for the input. With the filter resistors, you can't inject the IDAC current into the measurement input because the voltage from the IDAC current across the filter resistance creates a large error.

If you need to use input filtering, then you can only use 3 2-wire measurements with the ADS1248. The IDAC current is sent to the RTD lead, bypassing the input filtering. I have a second diagram below:

If you have questions about either topology, feel free to post back.

Joseph Wu

Reelf,

You may be able to connect four 2-wire RTDs if you share the negative input. I'm attaching another figure that shows how to connect them:

In this case you set AIN6 as the AINN and then cycle through AIN0, AIN1, AIN3, and AIN5 for the AINP inputs. The IDAC output would follow the measurement of each RTD at IEXC1, IEXC2, AIN2, and AIN4 for each respective measurement.

Joseph Wu

Reelf,


Also, if you are evaluating different devices for your application, you should look at the new ADS124S08. It is considered a next generation version of the ADS1248, with more features and input channels.


Joseph Wu

Hello Joseph,

thank you very much for your replies and your assistance !

This (see above) solution looks very good to me. Though "You may be able" confuses me a bit. Sounds like you are not sure if it will work ?!

As all leads to the PT100 are short (less than 1 meter) there won't be much noise and the filtering show in the above schematic should be alright. What about the resisors shown between AINx and IEXCx ? Any recommended value ?

We also had a look into the new ADS124S08. As the ADS1248 seems to do the job and we do not need more input channels, I'd say we'd stick to the ADS1248.

Best regards,

       Reelf

Reelf,


This solution should work fine. It was suggested to me by someone who had seen your post and discussed a bit before my last post. I just hadn't see this topology before.

For the input filtering, I definitely wouldn't go above 10k for series resistance. For low data rates, I'd consider using something like a 1kΩ to 5kΩ and a 0.1uF differential capacitance, with maybe 10nF for common-mode capacitance.


Joseph Wu

Hello Joseph,

thanks for the additional infos !  We'll incorporate the ADS1248 into our design.

Best regards,

      Reelf

Hi Reelf,

I know the ADS1248 looks sufficient for your application from a channel count point of view.
Nevertheless I would recommend to use the ADS124S08. This is the next generation of ADS1248 that incorporates many improvements over the ADS1248.
But that is just my personal opinion. There is nothing wrong with ADS1248. Both devices are very popular for this kind of application.

Regards,