• State Resolved
  • Locked Locked
  • Replies 2 replies
  • Subscribers 46 subscribers
  • Views 566 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

ads1247 external Vref question

Harm Verhagen
Prodigy 210 points

We see something unexpected when using the ads1237 with external Vref.

We want to measure a signal between -2.5V and 2.5V.

We do this by connecting AIN3 to GND, AIN0 to our signal, and measure between AIN0 (P), and AIN3 (N).

What we tried first:

As we wanted to measure both positive and negative, we supplied  -2.5V to REFN0, and 2.5V to REFP0.

The output of this clipped (as expected) above 2.5V, and below -2.5V.

However. the output was only half scale.  When supplying 2.6V to AIN0, we read back ~ 3FFF000 (not 100% stable). We expected to read 0x7FFFFF.

So we did not get the full resolution here, and to calculate the Voltage, we had to multiply with a Vref of 5V, instead of our expected 2.5V.

V = 5.0V  * (val/0x7FFFFF)   => gives correct result


(taking care of sign extending 24 bit two-complement val)

 

What we then tried (and seem to work)

We supplied  0V to REFN0, and 2.5V to REFP0.

Now we do get a full scale readout (0x7FFFFF) on >= 2.5V

And our voltages can be calculated with a Vref 0f 2.5.

V= 2.5V * (val/0x7FFFFFF)

the whole range -2.5V/2.5V works, and clips at the expected voltages.

So the seconds setup seems the fix for us, still I'm posting this as we don't understand it fully, maybe we oversaw something.

Questions:

1. Shouldn't give the first setup give a full scale readout on AIN0 = 2.5V ?

2. Why gives the second setup result in higher resolution than the first one ? (as the difference between REFN0 and REFP0 is lower?)

MUX1 = 0x0

SYS0 = 0x8

  • 0
    TI__Guru* 97755 points
    Harm,


    First, I want to make sure that you know that the analog input should be tied to ground for only bipolar supplies. In a PGA gain of 1, the analog input range is basically from GND+0.1V to 4.9V (assuming an AVDD of 5V). This is because the PGA is similar to the front end of an instrumentation amplifier and the input amplifier range does not span the entire range of 0V to 5V.

    If you have +/-2.5V supplies, then the GND is directly in the middle of this input range and should be ok. I think this is how you have your supplies set up, but I wanted to make sure you understood this input range.

    For the reference, the reference voltage is going to be REFP-REFN. If you have REFP=2.5V and REFN=-2.5V, then your total reference voltage is 5V. If AIN3 is at GND, AIN0 is 2.6V, and the reference is 5V, you get something about 428xxxh or something about half scale.

    In the second case, your reference is from REFN=0 to REFP=2.5V. This is closer to what you would use in normal operation. REFP-REFN=2.5V, so you would be able to measure inputs from -2.5V to +2.5V. You don't need to supply a negative reference for measuring negative voltages. That's why the internal reference is only a single +2.048V reference.

    There are two important things to note. First is that the reference input voltage should be no larger than AVDD-AVSS-1V. In this example the reference should not be larger than 4V. Second, the reference should be a low-noise reference. Any noise in the reference is reflected in the noise of the measurement.


    Joseph Wu
  • 0 in reply to Joseph Wu
    Prodigy 210 points
    Clear, thanks.