XTR111: VIN floating input, not connected question

A high-impedance input should never be left floating (with or without OD). In particular, the input voltage might go above the valid range.

A weak pull-down resistor would help. (The input impedance is 2.4 GΩ, so 1 MΩ would be enough.)

Correct.

Hi Davide,

too high source impedances at the inputs of OPAmp and XTR111 can cause stability issues. So, I would decrease the pull-down resistances to 100k.

Then I would add a series resistance of 2k5 between the pull-down resistor and the inputs of OPAmp and XTR111 (see figure 46 of datasheet of XTR111). And finally a low pass filtering cap should be added behind this series resistor to signal ground. This heavily decreases the source impedance at high frequencies and massively enhances stability. But it also serves as a low pass filter to remove HF-contents (coming from the DAC) at the inputs OPAmp and XTR111 and by this at the output cabling. Without these filters common mode noise can make the board fail the CE radiation testing.

Kai

Hi Davide,

The corner frequency of the filter will depend on the frequency range of the EMI signals that the circuit will be exposed on the application, how much EMI and noise needs to be filtered out, and/or the EMI test/certifications that your module will be submitted to.  For example, EMI testing generally has a low band range of 30MHz to 200MHz and a high band range of 200MHz to 1GHz.  Conducted RFI is often tested in the low-frequency range around 150kHz to 80MHz.  

 For example, let’s say you decided to set the filter corner frequency at around ~50-KHz, per your application EMI filtering requirement.

The OPA188 is a zero drift or chopper amplifier. Since the OPA188 is a chopper, the input bias current does exhibit periodic switching transients. This input bias current transient become transformed to a voltage offset when the bias currents flow through the high value circuit resistors. Therefore, chopper amplifiers like OPA188 may need to match the equivalent input impedances between the op-amp positive and negative input terminals in order to eliminate the commutation of the IB spikes across unmatched input impedance.  In the example circuit below, I matched the equivalent impedances in the inverting and non-inverting terminal to 10.7k | | 20.5K | | 22.6k = 5.36k on each input terminal to reduce the effect of the chopper amplifier IB spikes.  The capacitances seen at the non-inverting and inverting input terminals are also matched.   I tentatively set the capacitor to 550pF for a filter corner frequency to 53-kHz in this example:

From the XTR111 perspective, the device input has relatively low input bias current, 15nA typical, so you could use a resistor from 2kΩ to 10kΩ range without adding significant error.  For example, if you decided to use the same filter as above, you could use the same 5.36kOhm and 550pF for a corner frequency at ~54kHz.

Thank you and Kind Regards,

Luis