0
\$\begingroup\$

Below is a very simplistic view of differential ended inputs of a data-acquisition channel's inAmp which measures a floating battery voltage V. To make concepts and explanations simpler I use two resistors to mimic the inAmp:

enter image description here

Above R resistances represents the input impedance of the inAmp, and z are representing the line impedances.

Before asking my questions here are my understanding so far:

Differential current id is very small due to the input impedance of the inAmp:

id = V/(2*R+2*z)

Since a diff amplifier will subtract the voltages between the nodes a and b with respect to its own AGND, we can say the voltage across a and b causes by id is:

Va = id*R

Vb = -id*R

Va - Vb = id*(2R)

But due to the common mode currents ic1 and ic2, additional voltages( call it Vad and Vbd) are also created at node a and b as:

Vad = icm1*R

Vbd = icm2*R

Vad - Vbd = (icm1-icm2)*R

The last equation tells us that unbalance in the system(for example if z are not equal) will not reject the common mode voltages entirely.

If my view on the subject correct here is my question:

Another thing I can say is that if icm1 and icm2 large enough so Vad and Vbd becomes large which can exceed the rails of the inAmp. This can cause saturation and erratic readings.

As a remedy a 10k to 100k resistor is suggested to tie between the negative node b to AGND. I guess this is to lower the common mode voltages causes by icm1 and icm2.

But how come a 10k resistor tied between the node b and the AGND solve this issue? How is icm2 reduced by using the resistor? How about icm1? Is it possible to explain this using my circuit example not involving a real inAmp? Could you explain this in a simplistic way possible?

Edit:

I see that there is peculiarity. icm1 and icm2 has no return path, since the source is floating. But why is not having a return path creates big voltages at the inAmp side? If there is no return path for icm1 and icm2 then how does it create a problematic voltage at node a and node b?

Most importantly how does icm1 and icm2 exist at all if there is no return path for these currents to loop?

Btw I'm not sure if my analogy makes sense for id as input currents. Because input bias currents do not loop back to source the way I oriented right?

schematic

simulate this circuit – Schematic created using CircuitLab

Why sometimes the circuit in Figure 1 saturates and how does Rbias in Figure 2 fixes this problem? They say Rbias provides a return path to bias currents but why does the absence of return path causes inAmp to go saturate?

Why non flowing input bias currents saturates the inAmp or diff amp?

\$\endgroup\$
0
\$\begingroup\$
  1. But how come a 10k resistor tied between the node b and the AGND solve this issue?

Schematics are much better than words. Draw it out.

  1. icm1 and icm2 has no return path, since the source is floating.

Correct. So there is no common mode current.

  1. Most importantly how does icm1 and icm2 exist at all if there is no return path for these currents to loop?

It doesn't.

  1. If there is no return path for icm1 and icm2 then how does it create a problematic voltage at node a and node b?

Because if Z << R then \$ V_{ab} = V_{BAT} \$.


schematic

simulate this circuit – Schematic created using CircuitLab

Figure 1. Instrumentation amplifier showing bias currents.

With no return path to the power supply the inputs will be driven to one supply rail (positive with the bias current direction shown) or the other. You can consider these as constant current generators - a bit like your lab bench supply on constant current mode. If there is no significant load the voltage will rise (or fall) to the limit of the power supply.

Given that you have a voltage source between the two inputs it is likely that if the amp is sourcing current that the + input will rise to V+ and the - input will be held at \$ V_{batt} \$ below that.

schematic

simulate this circuit

Figure 2. A constant current power supply feeding a load, \$ R_L \$.

How it works:

  • R1 biases Q3 on. Current flows through R2, Q3 and RL.
  • When the current through R2 rises to about 10 mA there will be enough voltage drop to turn on Q4. This will steal the bias from Q3 and the circuit will settle with 10 mA through RL. The voltage at NODE1 will be 10m x RL.

Now what happens to the voltage at NODE1 if you disconnect RL? The current through R2 will decrease, Q4 will shut off, Q3 will turn hard on and NODE1 will rise to V+.

This is a very crude explanation and I have no idea what the input stage of your amplifier looks like. It might help.

\$\endgroup\$
  • \$\begingroup\$ I dont think you wrote what Im asking about. Maybe because of my circuit does not represent the real one. This is about an instrumentation amplifier coupled by a floating source. The inAmp many times exceeds the rail voltages if you do not provide bias currents to return. Im asking what makes the inAmp to go upto rail voltages in the absence of those bias/return resistors. I dont know since only you replied I guess somthing not okay with my question. \$\endgroup\$ – atmnt Mar 20 '18 at 20:21
  • \$\begingroup\$ Draw the real circuit and the real voltages measured. There's a schematic button on the editor toolbar. Add in the part number and datasheet link for the real circuit. \$\endgroup\$ – Transistor Mar 20 '18 at 20:28
  • \$\begingroup\$ I did now. See the last section of my question. Why sometimes the circuit in Figure 1 saturates and how does Rbias in Figure 2 fixes this problem? They say Rbias provides a return path but why does the absence of return path causes inAmp to go saturate? \$\endgroup\$ – atmnt Mar 20 '18 at 20:48
  • \$\begingroup\$ We're still missing the part number and datasheet link. \$\endgroup\$ – Transistor Mar 20 '18 at 21:52
  • \$\begingroup\$ ni.com/cms/images/devzone/tut/a/af64be30612.gif see top left ... thats it \$\endgroup\$ – atmnt Mar 20 '18 at 21:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.