• State TI Thinks Resolved
  • Locked Locked
  • Replies 11 replies
  • Answers 4 answers
  • Subscribers 47 subscribers
  • Views 8654 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

20Mohm input impedance with OPA277

Jatin Khatri
Prodigy 120 points
Other Parts Discussed in Thread: OPA277, INA128, INA116, INA111, OPA130, OPA140, INA121

I am trying to build unity gain buffer for field source of (-5V)~(+5V) DC (reference cell) and need min. 20Mohm input impedance at input of this buffer.

Here I have used OPA277 in differential mode configuration with Rs=Rf=10Mohm. Supply to Op-amp is +12V & -12V with 0.1uF & 10uF bypass capacitors. O/P goes to LPF (R=10K, C=0.1uF) further to feed analog multiplexer IC and then ADC.

- Reading is not stable at o/p and fluctuating about 50mV, also get affected if I keep multimeter or DSO probe at source terminal though Multimeter / DSO shows stable source reading. It is also floating if no source is connected.

Not sure if LPF at input will help because I also need fast response time of about 50ms, i.e. input from (max to zero) or (zero to max) within this time durating so it is DC input but should support up to 15~20Hz approx. Also not able to undertand how if Ib can affect its performance with 10Mohm (may be not possible with OPA277) because when I decrease Rs=Rf=10K in this arrangement, it works fine thoug not tried intermediate range of resistors.

 

Thanks

  • 0
    Guru 21975 points

    Jatin,

    Scaling the resistors of a difference amplifier to 10M ohms creates a variety of problems. I suggest that you change to a three op amp instrumentation amplifier such as the INA128. This type of amplifier has buffered inputs that provide very high input impedance. For a discussion of this type of amplifier please check out this design note:

    http://e2e.ti.com/support/amplifiers/precision_amplifiers/w/design_notes/1777.aspx

    Regards, Bruce.

  • 0
    TI__Guru 64290 points

    Jatin,

    I am not sure whether I correctly understand your dilemma - you claim to need a unity-gain buffer with minimum input impedance of 20Mohn and that is why you implement a differential configuration with 10M resistor network around OPA277.  But the input impedance of OPA277 is 100Mohm (see below) so why you do not simply use OPA277 directly in a follower configuration?

    Using large 10M resistors at the input of OPA277, introduce significant noise due to thermal voltage noise of the resistors themselves and the current noise converted to voltage noise as it passes thru 10M resistor.

    The thermal noise of 10M resistor is about 400nV/rt-Hz since Vn=sq-rt(4*K*T*R), thus voltage noise integrated across 1MHz bandwidth of OPA277 results in:

    Vn = 400nV/rt-Hz*(1e6*1.57)^.5 =~0.5mVrms where 1.57 is a brick-wall filter correction factor - see attached presentation on noise basics

    Similarly, the current noise integrated across 1MHz bandwidth will result in 0.2pA/rt-Hz * (1e6*1.57)^.5 = ~250pArms  which will be convertedto ~2.5mVrms across 10M.

    Taking into account the entire resistor network you use, these will somewhat increase the total voltage noise but not to 50mV level you claim to experience - my first guess would be that you gain the overall noise (10x probe) or the 10M resistors use utilize contribute to pick up of environmental noise.  All in all, using large resistors like 10M should be avoided if possible as it may also cause instability in the circuit close-loop configuration - use a scope and NOT a multimeter (rms reading only) to check for any signs of instability that would explain the output voltage fluctuation.

    Noise basics.ppt
  • 0 in reply to Marek Lis
    Prodigy 120 points

    Hello Marek, thanks for your detailed explanation.

    One thing I forgot to mention was, the source is coming from field (few meters distance & really not aware of field conditions) and hence don't prefer to terminate any of source terminal direct to circuit GND. Can say this is why want to keep it differential configuration to have some isolation. See below system block diagram for clarity. Let me know if I am mistaken here.

    Case-1: Simple follower configuration - Just to check, initially tried with single op-amp but issue observed was floating o/p when no source is connected. I tried with LPF at input (Rs=10K, C=0.1uF) but didn't solved the purpose. Here one of the source terminal common with circuit GND which is again not required.
    Case-2: Differential configuration with 10M ohm - Reading I observe on display of our control module which can read ADC every 100ms and I see such approximate fluctuation (50mv) in reading on display. With this arrangement, I captured below images keeping DSO at source and o/p of buffer stage. (For this testing used 10M ohm resistors with 1% TOL). RED line indicates source & Yellow line indicates buffer o/p
     
    Image-1: Buffer o/p connected to existing AI module input
    Image-2: Buffer o/p left open

    Case-3: Differential configuration with 10K ohm - Checked by reducing resistors value and readings are stable (1~2mV fluctuation & acceptable) on display. Below captured images,
     
    Image-3: Buffer o/p connected to existing AI module input

    Image-4: Buffer o/p left open


    Case-4: Source input direct to existing AI module - In display there is no such fluctuation (except 1~2mV & acceptable, same like case-3 with 10K combination)

    It seems result gets worst while I connect buffer o/p at existing AI module. Not sure apart from buffer, there also needs modification in existing AI module section.

    Please provide your comments.

    Thanks,

    Jatin

  • 0 in reply to Bruce Trump
    Prodigy 120 points

    Hello Bruce,

    One question regarding resistor in bias current return path if I use any instrumentation amplifier, minimum value I assume should be 20M ohm if I need min. 20M ohm buffer input impedance (for DC input like reference cell here). Will it work with INA128? or what parameter I should consider while selecting this value? as some of the datasheets I have referred (except INA116) demonstrate 1M ohm at there which is shown 100M ohm in case with INA116 or it is just for representation purpose and can go beyond that

    Thanks,

    Jatin

     

  • 0 in reply to Jatin Khatri
    Guru 21975 points

    Jatin,

    These bias current return resistors are only required if your signal source does not inherently have a path for bias current. This is the case with a ceramic transducer, as shown or with capacitive coupling. If your source signal is a voltage source of some type that can be "loaded" with a few nano-amps, then these resistors are not required.

    What is your signal source? I suspect that these resistors are not required in your case.

    If they are required, an appropriate value might be 100k for the INA128. Much higher values could be used for the FET input INAs such as the INA111 or INA128. But first, let's determine the nature of your source and whether these resistors are even needed.

    Regards, Bruce.

  • 0 in reply to Bruce Trump
    Prodigy 120 points
    Hello Bruce,
     
    It is for cathodic protection application. You might be aware of that where it requires to read potential difference between reference electrode and metallic structure. Just as different metals have different electrode potentials, also known as corrosion potentials or open-circuit potentials, different reference electrode types have different electrode potential values. So here need to measure potential difference between any such reference electrode & Metal structure (Pipe in this case).
     
    So, what do you suggest for these resistors with such source? and how about if I am using three AA cell (4.5V approx) for testing?
     
    Assuming if I don't need to keep resistors there, will the reading be stable '0V' when no source is connected? because this will be same like I tried with voltage follower with OPA277 as I described OR if resistors needed then again question of the resistors value arises.
       
    Thanks,
    Jatin

  • 0 in reply to Jatin Khatri
    Guru 21975 points

    Jatin,

    From your description it seems that resistors on the inputs are not required. It is not clear, however, how you intend to get a ground reference. Are you making three connections to the equipment you are testing?--metalic structure, reference electrode and ground? Perhaps you should provide a diagram of your measurement for confirmation.

    I assume that the three AA cells is just to test the instrument, correct? No resistor is required for this test.

    Though it appears that resistors are not required, you would get an unstable non-zero reading when the inputs are disconnected. If this behavior is unacceptable, then resistors may be required only for this purpose.

    Regards, Bruce.

  • 0 in reply to Bruce Trump
    Prodigy 120 points
    Hello Bruce,
     
    There are only two connections from filed, one from Metallic structure & one from reference electrode to connect at '+' & '-' of AI channel respectively.


    As I said earlier, both have some open circuit potential (known w.r.t some standard reference like hydrogen electrode), so measurement is to know actual potential difference between these two. If still confusing, can refer below link to understand the whole thing.


    Three AA cells, yes for testing in lab only. But in any case it seems need to consider resistors to avoid floating o/p when input is open. Please provide your comments how do you think I can get this done. I already have requested FET based IN121 to check.  

    Thanks,
    Jatin
  • 0 in reply to Jatin Khatri
    Guru 21975 points

    Jatin,

    Thanks for the additional information. I've reviewed the whole series postings and now have a clearer picture of the whole application and nature of the signals.

    Your original solution, a difference amplifier (four resistors around an op amp) may be the best solution. Stability can be solved with a capacitor across the feedback resistor. Other concerns with the very high resistor values are not so applicable in this situation (bandwidth and noise are not a big concern). I believe that you should, however, use a FET-input op amp. There are many possible choices but the OPA130 comes to mind.

    I still have concerns about the ground reference to the whole measurement system. You want two connections made remotely to the measurement points. How is the ground of the instrumentation system referenced to the measurement points? Could the instrumentation system be floating 25V higher or lower than the measurement potentials? This could render the measurement nodes beyond the common-mode range of the input amplifier. With 10M input resistors there is little to pull these two halves of the measurement system within range. This is why I asked about the possibility of a third ground connection.

    Regards, Bruce.

  • 0 in reply to Bruce Trump
    Prodigy 120 points

    Hello Bruce,

    Not sure but I think these two field connections can be treated like coming from any electrically isolated system to measuring point having isolated GND reference. Though in field both the '-' & '+' will have some common reference I believe (soil or sea water - installation location), at potential of few mV and I don't think will be floating high beyond common-mode range. See this link, it represents protection as well as measurement basics with an example, http://www.cathodicprotection101.com/

    Anyway I have requested couple of samples to check - OPA129, INA121, OPA140. Till then please let me know if you see anything I am missing here & must consider.

    Thanks,

    Jatin

  • 0 in reply to Jatin Khatri
    Prodigy 45 points

    Hi Jatin,

    Did you succeed in making the High Impedance buffer? Which Inst. Amp did you use? OR did you use the original configuration of Diff amp with 10Mohm resistors and capacitor filter in feedback path? 

    Thanks,

    Harshavardhan.