• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 46 subscribers
  • Views 476 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

ADS1256: Is it possible to use External PGA on ADC with its own internal PGA?

Ashesh Shrestha
Prodigy 50 points
Part Number: ADS1256
Other Parts Discussed in Thread: PGA116, ADS1255, INA849,

I need to convert low voltage differential signal (bridge circuit), some in 10mV range and some in 3-4 mV range, at about 500 samples per channel. Due to low voltage, to reach near FS, I need gain of more than 128 which is not available as integrated PGA as far as I know. Also, to reduce development time, as much as possible, I am looking to incorporate more widely used ADCs with existing libraries.

And ADS1255-56 seems like a good choice because it is widely available in breakout board form as well as has good sampling speed. But the big issue with ADS1255-56 is that the gain is only 64x. So can I use external PGA such as PGA116 with MUX as frontend to ADS1255-56 and run the ADC as single channel? First the PGA with MUX will convert differential signal to single ended ones. Then the semi amplified single ended signal will be fed to ADC where it wll be further amplified to achieve it's FS and processed.


What would be the drawbacks of such design? Would I introduce more noise or even if it is not exactly feasible?

  • +1
    TI__Mastermind 35141 points

    Hi Ashesh,

    Yes, you can certainly put an external amplifier in front of the ADC. You would always want to make sure the external amplifier is lower noise than the ADC however, or otherwise you will just be applying a larger gain to more noise. This will not help you achieve a high resolution system.

    You also want to make sure that when convert from differential to single-ended that you do not end up losing a resolution bit in the process. The ADS1255 is a differential input device, as is the coding scheme, so if you are applying a single-ended input you are only using half of the ADC's available code range. Most INAs have a REF pin that allows the output to be referenced to a specific voltage. You can then tie that same voltage to the (-) input pin of the ADC, while the INA output is tied to the (+) ADC input pin. Then, the INA voltage swings around the REF voltage, allowing you to take advantage of more of the full ADC code range.

    I am not as familiar with the PGA116, but I will point out that we just released the INA849, which is a very low noise INA that would be ideal for a bridge measurement. This does not have programmable gain, but I believe you could easily use 1x fixed gain to get the requisite noise performance. In most cases you do not need to use the ADC's entire full scale range in order to meet the system performance requirements.

    Also, here is a link to our Precision Labs content that focuses on ADCs. In particular, check out section 6 on ADC noise, which discusses some of the topics you are asking about here: https://training.ti.com/ti-precision-labs-adcs

    -Bryan

  • 0 in reply to Bryan Lizon86
    Prodigy 50 points

    Yes, you can certainly put an external amplifier in front of the ADC. You would always want to make sure the external amplifier is lower noise than the ADC however, or otherwise you will just be applying a larger gain to more noise. This will not help you achieve a high resolution system.
    I just hadn't thought about this in such manner. I will keep this in mind.

    You also want to make sure that when convert from differential to single-ended that you do not end up losing a resolution bit in the process. The ADS1255 is a differential input device, as is the coding scheme, so if you are applying a single-ended input you are only using half of the ADC's available code range. Most INAs have a REF pin that allows the output to be referenced to a specific voltage. You can then tie that same voltage to the (-) input pin of the ADC, while the INA output is tied to the (+) ADC input pin. Then, the INA voltage swings around the REF voltage, allowing you to take advantage of more of the full ADC code range.
    If I understand you correctly, a normal bridge circuit with say 5V CMV can fluctuate between say 4.9V and 5.1V. With 100x gain, if I don't use REF, I would just get 0-1V of analog signal because we can't go below common ground. But if I shift it by say 2.5V, the voltage would fluctuate from 1.5V to 3.5V. Hence higher resolution. Is my understanding correct?

    but I believe you could easily use 1x fixed gain to get the requisite noise performance.
    Sorry I didn't exactly get this. I am essentially using external gain because 64 isn't enough and higher gain with low noise amp would increase my noise free readings. What would be the benefit of using 1x gain?

    Also, here is a link to our Precision Labs content that focuses on ADCs. In particular, check out section 6 on ADC noise, which discusses some of the topics you are asking about here: training.ti.com/ti-precision-labs-adcs
    I will check this out.

    If I may, I also have one more question. I got to learn about fully differential amp but I am not sure I understand the benefits about it in my circuit. Because most of the gain is already carried out at front end, and the amplified signal is close to reference range, can it help in my design?

  • +1 in reply to Ashesh Shrestha
    TI__Mastermind 35141 points

    Hi Ashesh,

    Typically the bridge common-mode voltage (CMV) under a no-load condition will be about half of the excitation voltage. So if you are exciting the bridge with 5V, then the bridge output will be ~2.5V. When you apply the max load that generates a 10mV signal out of the bridge, then one side of the bridge is ~2.5V while the other is 2.51V, and you are measuring and amplifying this difference. As you said, if you have a gain of 100, then your output signal is from 0-1V. If the INA REF = GND, then the ADC is also measuring a 0-1V single-ended signal. If you set REF to 2.5V, then the output of the INA is 2.5-3.5V and this is what the ADC is measuring. For this example, the bridge output only moves in one direction (gets more positive relative to CMV or gets more negative relative to CMV), and the signal is single-ended. In the real world, this would be like measuring weight with a scale, where the weight only results in a positive output signal from the bridge since there is no such thing as negative weight. In this case, it  would make sense to bias the REF pin to 2V and the (-) pin on the ADC to 2.5V. Then, then INA output will swing from 2V to 3V relative to GND, and you will apply this 2-3V input to the (+) pin on the ADC. This will be measured against the 2.5V on the (-) pin, resulting in a signal that looks like +/-0.5V to the ADC. You will have to check the INA datasheet to make sure all of these inputs/outputs/CMV are within the operating conditions, this is purely hypothetical.

    A load cell on the other hand might experience both tension and compression, resulting in a bipolar output signal (one that swings above AND below CMV). This is where it makes sense to bias both the ADC (-) pin and the INA REF pin to the bridge CMV

    For the point about 1x gain: that is my error, I did not mean a gain of 1, I mean only using one gain. In other words, you probably don't need programmable gain and instead could just use a single gain value e.g. gain of 100, for all of the bridge measurements. It seemed like you might use the PGA116 to apply one gain for the 10 mV signal and another, higher gain for the 3-4mV signal. My point is that this is likely not necessary.

    The benefit of a fully-differential amplifier (FDA) is the ability to keep a fully-differential signal as a differential signal, instead of converting it to single-ended. There is actually a whole Precision Labs training module about FDAs in the op amps section if you want to review this material to learn more.

    -Bryan

  • 0 in reply to Bryan Lizon86
    Prodigy 50 points

    Thanks a lot for taking time and helping me sort out the issues. I do greatly appreciate your help.

  • 0 in reply to Ashesh Shrestha
    TI__Mastermind 35141 points

    Hi Ashesh,

    You are welcome, if you have any additional challenges with or questions about the ADS1256, please start a new thread and we will support you there.

    -Bryan