Post Go back to editing

AD8274 -- how is common-mode voltage defined?

Looking at the specs, in particular figures 8 and 10 on page 6, the common-mode voltage seems to be limited to 25V for +/-15V supplies.

My question is:  where is that 25V measured?  At the pins of the IC, or at the inverting and non-inverting input pins of the op-amp?  If the former (which seems reasonable, since the specs are given for the IC), then I cannot understand why if I use it in a gain=1/2 configuration, with +/-15V supplies, the circuit would fail with two identical input voltages of 30V  (the voltage at the input pins of the op-amp would be 10V, which is within the power supplies range).

Can you confirm?  (or clarify if there is something I'm misunderstanding?)

Thanks,

Carlos

--

  • Hi Carlos, 

    You are correct that the 25V in the datasheet will be taken at the input of the pins of the device. The reason it is 25V is because that the absolute maximum ratings at each pin need to be observed. See table below:

    If you have your supplies at +/-15V, then the maximum and minimum voltage at any input pin would be 25V (-15+40), and -25V (15-40), respectively. This is the reason why the input range is limited to +/-25V for a +/-15V supply. You can check page 12 of the datasheet (Input Voltage Range section) for more information. 

    I hope this helps!

    Kris

  • Thanks for your reply Kris,

    I had indeed noticed this detail in the absolute maximum ratings section and page 12.  However, there are two details that don't quite go with this:

    1.  Why is it different in G=1/2 and G=2 configurations?  The maximum should be 25V in both cases;  however, for G=2, the maximum is 20.85V.
    2.  On page 12 (Input Voltage Range), it says "for best long-term reliability" --- that sort of contradicts the "Absolute maximum ratings".

    I guess regardless, I should respect the absolute maximum ratings.  Since my common-mode voltage is in the order of 28V, and the differential voltage is small (below 1V), I guess I could power the IC with something like -5V and +25V or +30V, right?

    Thanks,

    Carlos

    --

  • Hi Carlos, 

    Apologies for the delayed reply. 

    The internal resistor network changes depending on the gain configuration that is used. See below:

    As you will see, the actual input voltage at the amplifier will be a result of the input voltage divided down by the resistor network. Since the resistor network will be different for the two gains, the diamond plot will also be different. 

    For example, for the gain of 1/2, the actual input at the amplifier will be +IN*(6k/18k). The actual input must always be kept within 1.5V of the supply. For a 13.5V input at the amplifier (given a 15V supply), you would be able to apply 40.5V at either IN+ or IN-. However, you have the absolute maximum rating to consider, and in this case, it will be the limiting factor to the diamond plot. 

    For a gain of 2, on the other hand, the actual input will be +IN*(12k/18k). For a 13.5V input of the amplifier, the input can only get to 20.25V at IN+ or IN-. In this case, this voltage is what limits the diamond plot. (By the way, the diamond plot says 20.85V, but as this test checks for the output swing, it will not always be so precise so this difference is acceptable.)

    You should be able to use it with a supply of 30V since you have a small differential voltage, and you plan to use a gain of 1/2 (is that correct?). May I inquire what input you will be using for the AD8274? What is your application?

    Thanks, 

    Kris

  • Thanks again Kris,

    My bad that I didn't notice that detail and did the complete calculations!

    I'm now a bit puzzled that you suggest that I can still use it at 30V common-mode (5V above the limit specified by the diamond plot).  Yes, it is a small differential, but the 25V limit applies when the output (thus, the differential) is 0.

    As for the exact application, I cannot say much, as this relates to a research project with an industry partner, and we are working under NDA.  Suffice it to say, I need to sense current consumption of a device that is powered off a 30V DC (approx.) power supply.  We place a current-sense shunt resistor that is small enough so that the voltage drop is in the order of tens of millivolts --- so yes, I have low differential, and high common-mode.   Let me know if you think this makes any difference w.r.t. what we've discussed.

    Thanks,

    Carlos

    --

  • Hi Carlos, 

    I apologize for the confusion, I meant 30V as supply, not common-mode. If you will be using the AD8274 in a single supply, you can scale up the diamond plot. See below:

    So for a 30V supply, with a 28V input common-mode level, the output swing would be somewhere around 5 to 25V.

    We do have current sense amplifiers, which are designed to be used in these exact applications where high common-mode levels are expected. You could check out the AD8418 if it suits your purpose. 

    Thanks, 

    Kris