• State
  • Locked Locked
  • Replies 1 reply
  • Subscribers 29 subscribers
  • Views 281 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

power supply ADC offset

Brian_Scot
Prodigy 60 points


I am using an ADC channel of the LM4F232H5QC to measure the output of a high gain current to voltage convertor (OPA129 op amp with 100Mohm gain). When I run the circuit using a bench top power supply, I get good readings with a noise floor of 20 ADC counts, which I am happy with. When I run the circuit using an isolated plug in wall adaptor I get a D.C offset in the ADC readings with the noise floor increasing to around 120 ADC counts.

If I connect the ground clip of an oscilliscope to the circuit ground (plug in power supply negative), the offset is removed, so I can't actually probe the circuit without affecting the problem. The circuit is on a 4 layer board with a power and ground planes, the plug in power supply is supplying a 5V and 3.3V linear regulator (TPS71550 and TPS71533) and the power supplies, op amp and microcontroller have decoupling capacitors as recommended by their associated datasheets. It seems there is still some ground loop somewhere. I think since the op amp is such high gain any small offsets associated with its inputs might be the cause of the problem and was thinking that perhaps using the ADC in differential input mode might be a way forward. I would be interested to hear any thoughts on this approach and any alternative.

  • 0
    Guru 117855 points

    Hate to see your post languish - here some quick comments:

    a) success w/bench supply - not so much w/plug-in wall adaptor - yet unsaid/unstated is how (or if) bench supply also supplies the dual voltages - as does the plug-in.  Might the difference you note result from an imperfect, common ground connection between the 2 plug-in supply outputs?  (and the bench supply's common ground - far more robust?)

    b) 4 layer pcb - power/ground planes - generally good - but you still must follow "best practice" in establishing (and maintaining) a solid, central ground tie-point - and minimizing unwanted current flows through sensitive analog signal paths.  (books are written on this subject - not for faint of heart)  That ground layer should not be relied upon to "do your work" - you must properly plan, power-steer and guard-band your board.

    c) probing the circuit - might you consider a safe, proper, isolated ground measurement device?  If the act of measuring disturbs - that's not good.

    d) differential mode - cuts your measurable signal level in half - and was your signal designed up front to be truly/properly balanced-differential?  And - does the output impedance of your amp match the input impedance of the MCU's ADC?

    e) properly managing such a high-gain amp - w/any MCU in close proximity - usually challenges.  Are the tiny, op-amp input signals properly "guard-banded" - so that they're not impacted by nearby, switching digital signals?  You may take a look @ how much effort the "pros" expend in isolating/shielding/routing their sensitive circuitry.

    Suspect that such a design must begin with your power supply - and power distribution.  If the "offending" plug-in is not intended for final/production use - trash it and replace with one more appropriate.  And - when high-gain amps are in play - how do you prevent (or significantly reduce) the intrusive effects of RF - which these days is everywhere?