AD8220 output voltage

I am running the AD8220BRMZ from +/-5V, with a gain set to 100 and the REF pin set to 2.5V. I notice that the output swing limits at 4.70V despite the datasheet indicating that I should get 4.85V as a minimum. Am I interpreting the datasheet correctly? This happens with a difference input of 22mV which is as expected i.e. (0.022V x 100) + 2.5V = 4.7V. The common mode max is 2.5V.


Another question I have is as follows. If I continue to increase the difference input beyond 22mV, the output remains clipped at 4.70V as expected but then above 40mV, the output begins to come down. At 45mV the output is 4.22V and at 49mV the output is 3.86V. Does anyone know what might be going on here?


Any comments appreciated. Thanks.

  • 0
    •  Analog Employees 
    on Jan 17, 2015 12:07 AM

    Hi pauly,

    AD8220 is based on the traditional 3 op amp instrumentation amplifier. In addition to the input range and the output range, you also have to consider the internal nodes that can saturate, especially the preamplifier outputs (node C and node D in the datasheet Simplified Schematic).

    Take a look at figure 30 from the data-sheet. It matches your gain, +Vs voltage, and REF voltage exactly. We're not running into negative supply limits in the example above, so it's okay that the -Vs voltage is different.

    You can see that at 2.5V common-mode, your AD8220 can be saturated above about 3.5V output. In fact, this is padded for worst case pinch-off voltage of the FETs. In reality, you might get a little bit closer to the rails. Also, you mentioned your maximum common-mode voltage, but I don't know what the common-mode was in the measured case, which would also effect the saturation voltage.

    We also have a web tool where you can check different configurations or select another in-amp:

    Instrumentation Amplifier Tool - Common Mode vs. Output Voltage (Diamond Plot)

    As for part 2 of your question, it most likely depends on the configuration of the rest of your circuit and how it will behave when the AD8220 is saturated, which can potentially have a few effects such as higher supply current and higher input current. To comment further, I would need more circuit info, such as: what is connected to the inputs and what is the source impedance, how are you driving the REF pin, is there any series resistance at the supply pins, and what is connected to the output. It would also help to probe the other nodes to see if there is any obvious reason, such as supply voltage droop or increase in the common-mode voltage.

    Maybe the more important thing for us to focus on first is fixing the conditions so the part doesn't saturate.

    I hope this helps.

    Best regards,

    Scott

  • Hi Scott,

    Thanks for your comments. I confess I've never paid attention to the diamond plots before (not being too sure how to interpret them). From your comment I can now see the answer to part 2 of my question. With a difference input of 40mV (when my output starts to drop), my common mode is actually 1.90V and figure 30 shows that's exactly what should happen. With a difference input of 49mV, my common mode is actually 2.50V and my output then measures 3.86V and figure 30 also confirms this.

    Going back to part 1 of my question, at the point that my output saturates (4.70V), the common mode input is actually only 1.10V. For these conditions, figure 30 suggests that I should be able to get up to 4.9V. I'm actually not too concerned about losing this 0.20V. To answer your questions, the 2.5V for REF comes from an ADR4525 reference, the -IN/+IN are driven from an AD8624 output, there's no series resistance on the supply pins and the output drives into a 2K load.

    Best regards,

    Paul

  • 0
    •  Analog Employees 
    on Jan 21, 2015 3:12 AM

    Hi Paul,

    I believe figure 30 was measured with no load on the output. You can see in the +5V specifications table that the output swing goes up to 4.85V, but also that is for a 10k load. 2k is a heavier load and the output will saturate earlier. Figure 35 shows the approximate behavior, but the part is loaded to 2.5V. In terms of output current near the top rail, 2k to ground looks like 1k to 2.5V. The 25°C value at 1k looks pretty close to 4.7V to me.

    I hope this helps.

    Best regards.

    Scott

  • Hi Scott,

    It was all there in the datasheet after all! Thanks for taking the effort to answer my 2 questions.

    Best regards,

    Paul