How to measure fully differential amplifier input impedance in pspice?

[electronx]

Before the differential ADC, I have a fully differential amplifier as the front end. I want to simulate and measure differential input impedance.

My working range is max 20 khz

Why am I using the 500/2k=1/4 ratio? because I want to include the incoming signal in the range.

ADS131E08S is the adc I use and in the datasheet I read that the input impedance is 200MΩ. For this reason, I installed a 200MΩ resistor as a load.



(V(Vin+)-V(Vin-) ) / (I(E2:3)- I(E1:3))

Generally, there is a concept of high input impedance at the input of boards that read ADC. Let's assume the differential opamp input goes to the connectors. In this case, is the opamp input really a value like 2k ohm?

here are the possibilities.

1.The conclusion impedance is correct (around 2k ohm ), you should do something to increase the input impedance
2.I misunderstand the concept of differential impedance.
3.I make an input calculation error
4. Everything is wrong

Can someone help sort out the confusion in my mind?

[Marco]

You shouldn't use FDAs if you want high input impedance. You want either an instrumentation amplifier or amplifiers like max4208/ad8129.

Or just rely on the PGA in the ADC.

[tszaboo]

In this case, is the opamp input really a value like 2k ohm?

here are the possibilities.

1.The conclusion impedance is correct (around 2k ohm ), you should do something to increase the input impedance

Opamps and FDAs will change their outputs to keep their inputs at a fixed voltage. Therefore the input resistance really is the resistor that you have before the input. This only applies to the frequency range the opamp has enough GBW, slew rate and voltage range to operate correctly. For the purposes of resistance calculation, you can just connect the input of the opamp to a fixed voltage, what's on the other pin. FDAs are a bit more complicated, but not by much.

[iMo]

You may buffer the two inputs of your diff amplifier with two opamps, each wired as the none-inverting buffer, for example.

[electronx]

The thing is, the input impedance of my fully differential amplifier is determined by the r2 and r3 resistors. So here it is 2k single impedance (differential impedance is 4 kOhm? right).

The system has a feedback structure.

(only in the non-inverting configuration, if there is no resistor or passive filter on the opamp leg, the impedance directly at the input of the opamp will be seen as high impedance).

   Additionally, if parallel series resistors and passive filters are used at the input, when the input impedance is taken as equivalent, Resistance (real) - parallel impedance - parallel impedance and equivalent impedance may decrease. Here, the low resistance value always wins.

And the currents that the sensors can give are limited. If the maximum current that it can give is not enough to keep the voltage constant (the current drawn is determined by the input impedance of the ADC), not enough current can be pumped and the signal voltage amplitude starts to decrease.

FDAs have lower input impedance? I think it is used in high-speed applications and these applications can provide high current from source  with low impedance. (I am still not sure in which ADC front end circuits the FDAs are used, I think a source that can provide high current is needed.)


Another issue is that if there is a passive filter, it significantly reduces the impedance. Even if we increase the input resistances of the FDA, the impedance still decreases with the logic of low gain. At the same time, as the resistance values ​​increase in the FDA, the noise increases and makes the output noisy.

In this case I will place a high impedance INA at the Input. and then I will send it to adc with high impedance.

I will add termination resistors to the input. In this case, I can make measurements from high-speed sources that have low impedance and provide high current.


 What I'm wondering is, if I add a passive differential noise filter or common mode passive filter, the impedance will decrease again. In this case, should I use an active filter?

If there are any points I think are wrong, please correct me.

[Marco]

FDAs have lower input impedance? I think it is used in high-speed applications and these applications can provide high current from source  with low impedance. (I am still not sure in which ADC front end circuits the FDAs are used, I think a source that can provide high current is needed.)

At high speed 4k differential impedance is huge already, standard transmission line is 100. Plenty of room for some transformer gain.

[Kleinstein]

The FDA amplifier has a relatively low impedance, similar to an inverting or simple differential amplifier. If needed one could likely increase the resistors a little (e.g. 2 x or 5 x), at least for the relatively low frequencies of interest here.
A point with the FDA is that one can set a gain of less than 1. The INA circuit is good for high gain, but already not great for low gain like x4 or  x1.
If high impedance is wanted would would add buffers in front.

Filtering not necessary decrease the input impedance. The simplest form are capacitors in parallel to the 500 ohm resistors and this has no effect on the low frequency input impedance. The other possible position for a filter may be between the amplifier and ADC. It is also possible to get low pass fitlering with additional series inductors.

[electronx]

Ah, I figured out how to keep the cut-off frequency of the passive RC low pass filter at the input constant and increase its impedance.decrease capacitance, increase resistance   

However, the point to be considered here is that the r value in the rc filter must be added to the r2-r3 values in the scaling ratio. When we optimize the values, here is the result. Increasing the resistances too much increases the noise. So it is best to experiment.
Thus, the impedance of the filter increased to the megaohm range and   connected  100 k (parallel low resistance) versus,

And we have a 100k resistor parallel to the impedance  to 7.5Megaohms. low resistance always winner
impedance remained constant at higher frequencies.