• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Subscribers 46 subscribers
  • Views 3665 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.

24V Battery operated system (INA114, XTR110) picking up noise

Jayant Deshpande
Expert 1495 points
Other Parts Discussed in Thread: INA114, XTR110, INA333, OPA171

Reference circuit:  http://www.ti.com/lit/an/sboa025/sboa025.pdf

Hi,

I have the above circuit operating on 24V Battery. The only difference is that the load cell has a 3m long shielded cable.

Whenever there is any Mains operated switching in the vicinity, there is fluctuation of output current. Probing further, noise is observed on the power rails. Adding a filter before the INA114 and decoupling the power rails doesn't seem to help.

There is no earthing available as the unit is mounted on a mobile platform having rubber wheels. The load cell shield is terminated to DC Ground near the amplifier.

Can anyone help solve this problem?

Thanks,

JayantD

  • 0
    TI__Genius 12325 points

    Jayant,

    I can make a couple of recommendations from the schematic. The first change I would make would be to add a capacitor across the 10.7 kOhm resistor that is being used to set the reference voltage. From your description of noise appearing on the power rails, I am wondering if this noise is being fed back into the reference pin of the INA114 by the reference buffer. Adding a capacitor across the 10.7 kOhm resistor should help prevent this from happening. I would go with a value greater than 1uF.

    Next, it may be advantageous to use a shield driver for the cable between the load cell and the amplifier. The basic configuration for this is shown below:

    An op amp is used to extract the common-mode of the input signal, buffer it, and drive the shield at this potential. This can help to eliminate the degradation in common-mode rejection that is caused by the cable's distributed capacitance. Also, do you have bypass capacitors on the supply pins of all the integrated circuits? The schematic above is a simplified diagram and doesn't include power supply filtering, input protection, etc.

    One final aspect to consider is that the load cell itself picking up the noise. This would cause hum from the mains power to appear at the inputs of the INA114 as a differential signal rather than a common-mode signal and would therefore be amplified. If your measurement bandwidth is less than 50/60Hz then it may be a good idea to add low-pass filtering at the inputs of the INA114 to reject this noise. 

  • 0 in reply to John Caldwell
    Expert 1495 points

    Hello John,

    Thanks for your prompt response.

    I will add the capacitor across the 10.7k resistor as this is the simplest to implement. The shield driver calls for modifications which could take weeks. I will do it and keep you posted.

    Thanks again,

    Jayant 

  • 0 in reply to Jayant Deshpande
    Guru 21975 points

    Jayant,

    Given the battery power and other symptoms (no change with added input filter), I'm suspicious that the mains glitch may be imposing a voltage step or glitch on the output terminal (drain of the FET). While the output impedance of the XTR110 is very high at low frequency, it relies on the feedback loop through the output op amp of the XTR110 to control the gate voltage of the FET. The switching of the mains may be "pinging" the FET and its output feedback loop, thus creating the glitch. The output impedance of this circuitry is not so high at high frequency.

    How is the ground of this circuitry returned or connected to the measurement circuitry. This could be the source of an impressed voltage glitch.

    Regards, Bruce.

  • 0 in reply to Bruce Trump
    Expert 1495 points

    Hello Bruce,

    The output current returns to ground through a 250 ohm resistor inside a remote indicator. The two systems are connected by approx 3m long cable,

    Do I need additional circuitry to overcome this problem? As it is the PCB has to be redesigned in order to incorporate the shield driver suggested by John, so any additional protections can also be added to the new version.

    Thanks,

    Jayant

  • 0 in reply to Jayant Deshpande
    Guru 21975 points

    Jayant,

    I was also thinking about how the ground of the two sides are connected. I assume that the ground of the XTR is returned to the measurement system with the current signal, correct? Twisted pair, perhaps. It might help to add a 100 ohm resistor in series with this ground line. Don't allow the XTR circuitry to connect to any other earth ground such as a chassis.

    Sorry, but it's difficult to diagnose from a distance.

    Regards, Bruce.

  • 0 in reply to John Caldwell
    Expert 1495 points

    Hello John,

    For the above scheme, Can I replace the INA114 with a INA333 ?  Attracted by the term "RFI filtered inputs"..!

    Any recommendations for the shield driver op amp?

    Thanks,

    Jayant

  • 0 in reply to Jayant Deshpande
    TI__Genius 12325 points

    Jayant,

    The INA333 is a 5V part and will not work on the 10V supplies that you are providing to the INA114. Also, RFI/EMI filtering at the inputs is really meant to deal with higher frequency noise and won't be effective for low frequency problems. Talking to Bruce I realized that we were a bit unsure of the nature of interference you are experiencing. Is this high frequency noise, transients, low frequency mains hum, or something else? I would really recommend doing some careful investigation of your system before investing in a new PCB. One thing to try would be to remove the load cell and cable and replace it temporarily with a resistor divider directly at the input of the INA114 to simulate the output of the load cell, does the effect persist? This would help determine if interference is coupling into the circuit through the load cell and cable, or somewhere else in the signal path. The shield driver does not need to be a high precision device. An OPA171 would be suitable as long as a resistor is placed in series with the output to isolate the cable capacitance from the op amp as this can cause oscillation. 

  • 0 in reply to John Caldwell
    Expert 1495 points

    Hi John,

    It seems to be transients and high frequency noise somehow getting coupled to the system. I plan to rig up a similar system, create noise with a chattering relay and capture some scope screenshots. Will come back to you with details.

    Thanks,

    Jayant

  • 0 in reply to Jayant Deshpande
    Expert 1495 points

    Problem solved!

    It was related to improper PCB layout. Careful track width selection and decoupling did the trick.

    Thank you for your support, John, Bruce!

    Regards,

    Jayant