AD8602 Offset voltage variation

Hi,

My customers are considering using the AD8602.
So I got a question.

They applied 2.0 V to the differential input pin and kept operating at 75 ° C for a long time.

(experimenting with the circuit below.
Connect a 6.8 kΩ resistor to GND and operate for several hours, then release GND and measure the offset voltage between Rload and R 4 with the following circuit.)


When the offset voltage was measured every few hours, the offset voltage decreased as time passed and it was lower by 50 uV in 100 hours than at the start of measurement.

# 1 Is this phenomenon due to NBTI?Or is there any other reason?

# 2 Is there graph data on offset voltage fluctuation measured along with voltage, temperature and time change?

Best Regards,

Yuya

  • Hi Yuya,

    I did not quite understand the test method they use during VOS test, "Connect a 6.8 kΩ resistor to GND and operate for several hours, then release GND and measure the offset voltage between Rload and R 4 with the following circuit"

    Where exactly does the 6.8K is connected? I see that the 6.8K is feedback in the schematic, "the 6.8k to ground" is this another resistor not shown in the schematic? also could you please comment the ground connection on R4 and Rload I assume this is present is the actual test circuit?

    VOS which is a product of small mismatches in the input stage on the opamp is expected to drift over temp and over time. The parameter TCVOs (offset voltage drift) provide the value of the expected VOS drift over temperature and for AD8602 it is 2uV/'C. Figure 8 shows the VOS drift of hundred of units tested from 25'c to 85'c. Ideally the lower the VOS the better, I just could not confirm if  the test method used in getting the VOS is valid based on the circuit above and the drift is ideally from the part and not from other sorce like the resistor, and meter.

    You can refer to this document for discussion on input offset voltage, it also discusses the factors affecting VOS drift, it also has a suggested schematic how to test VOS using in amp.

    http://www.analog.com/media/en/training-seminars/tutorials/MT-037.pdf

    Could you please share more information about your customer and target application of the AD8602. 

    Regards,

    Joven

  • Hi, Joven

    The circuit diagram was wrong. I'm sorry.
    As shown in the figure, the connection to GND is switched by 6.8 kΩ.
    R4 and Rload are input to the + terminal.

    I can see that VOS changes with time.
    Is there a net increase in offset voltage over time(About 100 hour)?

    Best regards

    Yuya

  • HI Yuya,

    Thank you for the additional information.

    I am currently looking for data, if there is any, for the VOS drift over time of AD8602, I will get back to you on this.

    Regards,

    Joven

  • Hello Yuya,

    Joven contacted me for the drift data. Please see attached for the Vos vs time (in hours).

    One issue I see is that the part is in comparator mode when the ground is switched in. I guess it is intentionally designed this way. You have 2V at VIN+ and 0V at VIN-; so the output of the circuit will be railing positive. The article below goes over using the op amps as comparators and their corresponding advantages and disadvantages. In on section, it mentions that the large differential voltages could shift the amplifier parameters.

    http://www.analog.com/en/analog-dialogue/articles/amplifiers-as-comparators.html

    We have not looked at the long term comparator operation, but can you ask your customer to bake the part at 150C, unpowered and retest to see if the Vos recovered. If it recovers, it could be the charge injection/trap at the gates of the input devices.

    Regards,

    Kaung

  • Hi, Kaung

    Thank you very much!

    Regards,

    Yuya