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.

problems implementing TI PH application note AN-1852

Other Parts Discussed in Thread: LM4140, LMP7721, LMP7711

I am following this TI application note for PH meter design and I was advised by Sem from TI to use LM7711 chip in place of LM7721 (which was originally in the design note) for price consideration.

my circuit is on pcb , and i have kept the tracks short   but I am getting strange behaviour from the circuit.  Its putting out random voltages at the output and seem to loop between values from 0.0 to 50.0 . 

I measured signals at two points and i have put the pics . PH1 shows the measurement at the junction of R1 and R2.

PH2 shows the signal at Vout.

why am i seeing an AC form instead of a DC
signal ?    does using 7711 instead of  7721 requires some
circuit changes ?

Iam using the PH probe from a working PH meter so i know the probe is good.



please help me out with this since this is the only thing holding my project at
the moment.

I am also attaching the Eagle CAD circuit and board layout for this.



thanks



  • Hi Sami,

    It appears that there is a stability issue associated with the LM4140. The LM4140 requires a capacitor at the output to be stable, please see the Application Hints on page 9 of the LM4140 datasheet. I recommend choosing a capacitor from Table 1. If you choose to use a different capacitor be sure to choose a capacitor that has an effective series resistance (ESR) within the Stable Region of Figures 24-26. Or since ceramic capacitors typically have a very low ESR, you can add a 2 ohm resistor in series with a 1uF ceramic capacitor. This will ensure the ESR requirement of the capacitor is above its minimum requirement.

    Another recommendation would be to add decoupling capacitors to all the devices. Here is a blog explaining why these are necessary.
    http://e2e.ti.com/blogs_/b/precisionhub/archive/2013/08/13/the-decoupling-capacitor-is-it-really-necessary.aspx

    -Tim Claycomb

  • Hi Sami,

    Add the bypass cap to the reference as Timothy suggested. To reduce noise from the reference, I would also recommend a series resistor (~50K)  between the input and the reference, and a large capacitor (~10uF) from the input to ground to add R-C filtering.

    The buffer driver U1 may also be oscillating due to the heavy capacitance of the cable. Op-Amps do NOT like directly driving capacitance more than 50pF, especially in the follower configuration.

    Add a 200 ohm resistor in series with Pin 1 and the PSGND connection. This should effectively isolate the capacitance.

    You should also add a series resistor to the U2 output to also isolate capacitive loads. DMM's can have hundreds, or even 1000's of pF capacitive load. Add a 200 ohm resistor between U2 pin 1 and the connector.

    It is good practice to use a parallel feedback R and C to allow tuning of the bandwidth, instead of directly connecting the op-amp input and output together.

    The LMP7721 is a specially modified version of the LMP7711, in where the input ESD protection is bootstrapped to provide low input bias currents.

    Electrically, they are interchangeable, but to maximize the LMP7721's low input current, the pinout was modified to allow guarding and to keep the power supply pins away from the input, so it is not a "standard" pinout.

    The LMP7711 has a standard single pinout, but is not pin-to-pin compatible with the LMP7721. So you cannot interchange the devices on the same layout (though the basic circuit can be the same).

    Regards,

  • hi PAul!

    i have modified the design as per yours and Tim's advise except one "a series resistor (~50K)  between the input and the reference"  ,  where do I place it ? can you please point it out on the circuit diagram iam uploading and also see if it looks good now?

    Also the LM4240 datasheet page 10 mentions the safe ESR range as 1.000 - 10.0000 , is it one thousand to 10 thousand or one to ten ?  the part TI recommends on page 10   Nichicon F931C105MA has an ESR of 7 as per digikey site.

    thanks

  • Hi Sami,

    Actually, you could increase the R1 & R2 values to 100K and place a 10uF cap (or as large as possible) between the U1 +IN and ground. That way the R1/R2 divider does double-duty as the "R" part of the RC reference filter.

    The feedback should be taken from the output side of R3, otherwise you will run into instability problems. To do what you have done requires a small (~10pF) feed-forward cap between -IN and the output to maintain stability.

    You should also add a 200ohm isolation resistor between U2 output and the output connector.

    The LM4140 recommended ESR stable range is 0.4 to 7 ohms for 10uF.

    Regards,

  •  hi Paul !

    I have moved the feedback to outside of R3 and also added a 200ohm resistor ,do I still need that 10pf feed-forward cap ?  would it be in series with the R3 or directly connected between -IN and OUT ?

    here is the circuit dia so far

     

  • Hi Sami,

    Whups! Close... Move the U1 pin 1 connection to the right side of R3, so that -IN and OUT are tied together,

    R4 looks good, as does C4.

    You should be able to patch these in to your existing layout to see if they do make a difference, before committing to a new layout.

    The most critical path is the PSSIG and U2_3 connection. This needs to be as short and compact as possible. It looks like you are going through a via to the bottom. I would run a trace on the top layer directly from the connector. You will have to re-route the V- traces. It is better to via the V- trace than the signal traces.

    Regards,

  • hi Paul !

    I made the change as per your suggestion.I have kept the traces from PSSIG to U2_3 very very short as you see from the board layout.

    I have removed the via to pin-3 . please confirm if all looks good now. please check the value of the capacitors if they look ok.

    appreciate all your help

  • hi Paul I am anxiously waiting for your answer !

    regards

  • i got the pcb built and put it together but i am getting very unstable results .

    what i noticed is that the Vph voltage is not getting added to the second op amp .

    how can i debug this circuit ?

     

     

  • Hi Sami,

    Please post the final schematic and layout.

    Regards,

  • attaching the schematic and board layout .

    thanks

  • Hi Sami,

    You should see 0.5V between circuit ground and the outer shell of the connector. If you do not see the 0.5V, then the U1 buffer circuit is not working.

    If you short the connector center conductor to ground, the output of U2 should be near zero.

    If you short the center conductor to the connector shell, then the output of U2 should be 0.5V.

    Does your ph probe have a triaxial connector??

    Is your board inside a shielded enclosure? Did you clean the board with rubbing alcohol? This is a very sensitive circuit.

    Regards,

  • Hello all. Maybe AN-1852 needs more detail.
    I also have interest in this topic.
    I read the topic a lot of times and need ask a few questions:
    1-There will be serial capacitor in U1's feedback (won't be shorted)? If yes, why?
    2-There will be 50 k serial resistor between lm4140's output (VREF) and U1's +IN?
    3- I want to add active low pass filter after U2's out pin (U2 out signal will pass through filter). May it cause any problem?

  • and from the AD797 data sheet: i.stack.imgur.com/iZzJl.png
    In this 100 ohm R2 and R1 resistors have used in different way from your advices?