0
\$\begingroup\$

I need to apply a current in the body by two electrodes and then analyze this signal, I will receive the signal by other 2 electrodes and amplify using a instrumentation amplifier using three Op Amps (TL084CN). But, I can not to simulate the offset by voltage controlled current source in a instrumentation amplifier.

I am using the topology described in the article "Design of an Instrumentation Amplifier - Justin Bauer", where the gain is 10. Example: With Vin = 500mV peak to peak, then Vout = 5V peak to peak. (Gain = 10). Putting 5V DC on Vref, then Vout is oscillating between + 10V and 0V. All right.

My problem is: Using a current source (Mirrored Modified Howland Current Source (MMHCS)) where +Iout = 1mA and -Iout = -1mA. With a initial voltage source where amplitude = 2V and the frequency = 1kHz. I can not to simulate the offset. Vout is always between + 10.5V and -10.5V. enter image description here I tried to apply in Vref 10V, 5V, 2V, but nothing works. The Vout still varies from -10.5V to + 10.5V. The Vref only distorts the voltage signal. I want the Vout varies between (say) + 20V and 0V, it will use the AD converter Arduino Uno, that just read positive signs. I am using Proteus 8.5.

What am I doing wrong?

Sorry my poor english, i don't speak english.

Bibliography:

http://www.egr.msu.edu/classes/ece480/capstone/spring15/group05/uploads/4/7/5/1/47515639/ece_480_app_note_justin_bauer.pdf

https://www.researchgate.net/publication/258327515_High_Accurate_Howland_Current_Source_Output_Constraints_Analysis

\$\endgroup\$
6
  • \$\begingroup\$ what resistance are you modelling for the body. \$\endgroup\$
    – Andy aka
    Commented Oct 30, 2016 at 17:43
  • \$\begingroup\$ I think you don't need two current sources, just one will do. Electrode 2 should be connected to ground. Also, the human body is an antenna for RF signals, so you should have input filters on Electrodes 3 and 4. There could also be DC offsets caused by differing skin-electrode electrolytic potentials, so I would AC-couple the amplifier inputs (unless you want to measure these DC offsets too!). \$\endgroup\$
    – Rich S
    Commented Oct 30, 2016 at 17:45
  • \$\begingroup\$ The amplifier's gain-setting resistor, R7, will probably need to be adjusted in real-world use, to keep the output signal within the range of your power supplies. This is usually done by replacing R7 with a digipot (digital potentiometer). The digipot is controlled by your computer device (Arduino, etc.). \$\endgroup\$
    – Rich S
    Commented Oct 30, 2016 at 17:50
  • \$\begingroup\$ I'm not modeling any resistance to the body. I'm connecting the current source outputs directly to the instrumentation amplifier inputs. As if the body had no resistance. \$\endgroup\$
    – Dyarlen Iber
    Commented Oct 30, 2016 at 19:39
  • \$\begingroup\$ That makes no sense. Those inputs have giaohm impedences \$\endgroup\$
    – Scott Seidman
    Commented Oct 30, 2016 at 22:29

1 Answer 1

Reset to default
1
\$\begingroup\$

I'm not modeling any resistance to the body. I'm connecting the current source outputs directly to the instrumentation amplifier inputs.

That looks like the problem - you are trying to inject a mA into a giga ohm (the input impedance of the InAmp) and not surprisingly the current driver cannot force 1 mA into 1 Gohm. It would be able to if the power rails were +/- 1 million volts but they aren't and cannot be.

Try putting 100 ohm across the output of the two drivers. This then forces a voltage of 100 mV and, with a gain of ten you will get 1V out. Your expectations of 20V out are unreasonable given that the InAmp power rail is 12 volt.

One more thing - you will need a galvanic connection between the driver circuit 0 V and the InAmp 0 V. The human body (a real one) might facilitate this via other "connections" but for your simulation you'll need one.

Try 100 kohm.

\$\endgroup\$
1
  • \$\begingroup\$ Thanks for the help, i will research about galvanic isolation. \$\endgroup\$
    – Dyarlen Iber
    Commented Oct 31, 2016 at 1:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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