AD8253 PGIA nose issue

Dear Sir,

The PGIA AD8253 datasheet showed noise performance:

Voltage offset RTI Vos +-150+900/G uV.

Voltage nose 0.1~10Hz,RTI,G=1, 2.5uVp-p.

Then I tested AC performance by 6.5 Digital MultiMeter(HP-3304A), which noise performance showed 4.03mV DC offset and 2.6mV AC ripple. The real noise seems huge than specification.

Refer to my testing schematic as blow. Any comments for this issue?

  • 0
    •  Analog Employees 
    on Dec 16, 2015 9:05 PM

    Hi Ray,

    With your DMM, how are you performing your configuration? Are DC and AC measurements separate? Any bandlimiting or filtering?

    Total noise is going to be a function of many things, including NSD Bandwidth, your DMM's own bandlimiting, etc.

    Cheers,

    -David

  • Hi David,

    I separated  AC and DC performance measurement and have not any external passive filter. I checked HP-34401A specification, which bandwith is 300KHz.

    I tried to short this probe before I tested, thus DMM showed 0.000002V ACV-ripple and 0.000001V  DCV offset, that means DMM's background noise very low.

    Cheers, Ray.

  • 0
    •  Analog Employees 
    on Jan 25, 2016 11:34 PM

    Hi Ray,

    The DMM may very well have excellent input noise, but the noise it ends up measuring will not be its own. Instead the question's intent was to figure out how much noise the amplifier is going to contribute. In DC Averaging, most AC components should be eliminated, while AC would report an RMS noise level.

    The AD8253 has a 45nV/rtHz noise at a gain of 1. If you then short the inputs together and ground them, we expect 1uV/rtHz of output noise. You specify that your DMM has a 300kHz 3dB bandwidth, and if we assume it to be a second-order roll off, will contribute sqrt(300,000*1.12) = 579.6rtHz of noise-bandwidth gain, which when multiplied with 45nV, comes out to about 26uVrms of noise. To cover the broadest case, this quantity typically will get multiplied by 6.6 to get peak-to-peak noise, which gets us to 172uVpp. This is obviously an order of magnitude smaller than what you're measuring, but does demonstrate that the noise bandwidth is going to dominate over the low-frequency "near DC" noise spec (0.1-10Hz).

    We can do a bit of a reality check quite easily by installing a simple first-order RC lowpass filter on the output. A 100kOhm resistor with a 15nF cap should yield a 100Hz 3dB bandwidth. In this case, at a gain of 1, the output noise should be 563nVrms or 3.72uVpp. If the system has symmetry of behavior, your 2.6mV should show up as 56.2uV.

    Have you looked at the output with an oscilloscope to see if the noise is truly random, or if it has a specific periodic nature?

    Have you used your meter to measure the noise on your power supplies?

    Cheers,

    -David

  • 0
    •  Analog Employees 
    on Aug 2, 2018 3:18 PM
    This question has been assumed as answered either offline via email or with a multi-part answer. This question has now been closed out. If you have an inquiry related to this topic please post a new question in the applicable product forum.

    Thank you,
    EZ Admin