configure 28377D 16bit differential ADC to uniploar

Hi Jianwu,

If you do not provide the complimentary half of the differential signal, or if you connect the negative ADC input to ground, the ADC will simply not function correctly (you don't lose a bit).

If you have a single-ended signal that you want to sense, you should use a fully differential op-amp (e.g. THS4531) or a pair of normal single-ended op-amps to create the differential compliment about the common mode voltage (Vcm = VREFHI / 2).

Note that all voltages applied to the ADC pins should be between VSSA and VDDA at all times...differential mode of the ADC does not allow negative input voltages.

Hi,

I understand what you are saying. Here is the sitation: ADCINP: 0 to 2.5V, ADCINN: 0 to 2.5V. Vcm is 1.25V. But ADCINP - ADCINN is from -2.5V to 2.5V. Am I correct? If you look at ADCINP - ADCINN, it can be negative and positive. I am saying if we only use 0 to 2.5V of that range. We are not exactly 16 bit.

Thanks
Jianwu

Hi Devin,

28377d technical reference manual Table 10-3 says that ADCINyP - ADCINyN will be in the range of -VREFHI to VREFHI. So if we only ADCINyP - ADCINyN in the range of 0 to VREFHI, then we don't use full range of 16 bit ADC. Is this correct?

Thanks
Jianwu

Hi Jianwu,

I think your understanding is correct.

If you only use 0 to 2.5V of the differential range, that would correspond to the positive ADC input always being between 1.25V and 2.5V, and the negative ADC input always being between 0V and 1.25V. The solution to this would typically be in the signal conditioning circuitry; you need to increase the gain by 2 and adjust the offsets of each input such that they both have an offset of 1.25V.

Is this output directly from your sensor, or do you already have some signal conditioning circuits?

Hi Devin,

You are right. It is exactly what we have right now. We do have conditioning circuitry between sensor and ADC. However, our team have concern of adding offset to the analogue circuitry mostly because of accuracy and interference on the sensor signal itself. Therefore, we are trying to contact TI to see if we can re-configure the ADC to only use 0-2.5V differential range and still achieve 16bit, which is what I mean by "unipolar" at the very beginning.

Thanks
Jianwu

Hi Devin,

With that being said, we can't do much on ADC internally. If we are going to use 16 bit differential ADC, we have to give -2.5V to 2.5V differential input. Just want to confirm with you.

Thanks
Jianwu

Hi Jianwu,

Confirmed. We do not have any way to adjust how the ADC reads the differential voltage.

One way you can increase effective resolution is by over-sampling and averaging. Every increase in the number of samples taken by a factor of 4 increases the resolution by 6dB/1 bit. This process is usually limited by the SFDR specificaiton. This may or may not be appropriate depending on what you are sampling and how much time you have to collect an ADC sample and take action from it.

It may also be worth posting on the precision amplifier e2e forum if you have questions about how to get the signal into the right range while introducing the least possible noise/distortion/error.