• TI Thinks Resolved

OPA625: Voltage Reference Buffer

Part Number: OPA625

Dear Forum Members,

I'm writing to you regarding a design problem:

We'd like to use the ADS124S08 in one of our products (i.e. sensor signal conversion) and besides converting the analog output signal of a DMS sensor we want to use the internal voltage reference of the ADC as excitation voltage for the mentioned sensor. As the circuit would draw a larger current than the ADC can deliver, I thought about buffering the voltage using one of your operational amplifiers. I want to maintain the high precision of the internal reference, hence I chose the OPA625 because the description stated its use as a "Precision Voltage Reference Buffer".

When testing the OPA625, unfortunately I didn't get a satisfactory result. I'm assuming it's a design fault, that's why I need your help.

I use the 3.3V and ground from the ADS124S08 Evaluation Board as supply for the OpAmp, non-inverting unity gain (output -> negative input shortcut) and the ADC internal reference voltage from the EvalBoard as the positive input. The output voltage is not as precise as expected and depending on whether I use a load resistor or not, it varies significantly. What am I doing wrong? Or do you have any other suggestions on how to buffer the reference voltage precisely?

Thank you very much and best regards,
Manuel

PS: Find the schematic and 3 photos (Measured Internal Voltage Reference, Measured OpAmp Output Voltage with and without load) in the attachment.Attached Photos.zip

  • Hi Manuel,

    The majority of the error you are seeing is likely arising from the 100kOhm resistor you have at the non-inverting input of the OPA625. The OPA625 has a relatively large input bias current (2uA typical). If we assume a typical value then you could expect to see up to 200mV offset at the output because of the bias current. 

    Regards,

    Zak Kaye

    Regards,

    Zak Kaye
    Precision Amplifiers Applications 

  • In reply to Zak Kaye:

    Manuel,

    Additionally, you are operating outside of the input common mode limitations of the OPA625. With a 3.3V supply, your upper common mode limitation is 2.15V. With a 2.5V input, you may still be able to get decent DC accuracy, but your transient response will degrade significantly, and this could also be part of why you see the output droop with a load. The easiest solution would be to increase your supply voltage to 5V if you are able to. If not, then you may consider using something like the OPA320 or OPA350 for your buffer.

    Best,
    Zak Kaye

    Regards,

    Zak Kaye
    Precision Amplifiers Applications 

  • In reply to Zak Kaye:

    Hi Zak,

    Thank you very much for the quick response! Unfortunately, I couldn't solve the issue with your suggestion. The 100kOhm resistor was only added later when I tried to troubleshoot the problem. I saw this design practice in in the datasheet of the OPA625 but it didn't change anything in the output voltage.
    On the other hand, changing to a 5V supply (which we have available on our main board) helped to solve the load behaviour, thanks!!

    Could the output voltage offset originate from a damage of the IC (during the soldering process or so)?

    Best,
    Manuel
  • In reply to Manuel Schmuck3:

    Manuel,

    What is your measured output voltage with the supply set to 5V and the 100kOhm resistor removed? It seems unlikely that the offset would shift because of the soldering process. Have you tried replacing the device with another OPA625? Do you have any capacitance between the reference output and buffer input? Have you tried looking at the input/output of the device on an oscilloscope?

    As a side note, when the datasheet says the OPA625 is a "Precision Voltage Reference Buffer", it means it works well for driving the reference pin of an ADC from some external voltage reference. However, it looks like you are using an internal reference for this function. As such, I believe the OPA625 may be more than you need for simply buffering an excitation voltage. Nevertheless, it should still work. You may consider using a lower bandwidth part like the OPA333.

    Regards,
    Zak Kaye

    Regards,

    Zak Kaye
    Precision Amplifiers Applications