Bio-impedance circuit inputs

Ask QuestionAsked 2 days agoActive 2 days agoViewed 60 times3

Recently, I wanted to build a replica of a bio-impedance circuit that I found in one of the NIH.gov papers. As obvious, bio-impedance circuits are usually used to measure body composition. They are using two electrodes in the design instead of four which is commonly used, meaning, the current will be inserted through the same electrodes as well as the voltage readings. I am still an undergrad student, so I had some questions regarding the inputs voltages and frequencies usually used to generate current or to be specific in this circuit. .
I am trying to generate signals with a function generator for Vin- and Vin+ but the paper hasn’t discussed what inputs are needed, for example, the voltage amplitude, frequency, and Vin+(-) phase difference. Also, they say they are doing a no-load current test in which they measure the current with a DMM. But where should I measure the current in a no-load case?

Question references:

1. Input signal characteristics?
2. Where is the no-load current measured?

Link to the paper: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3777733/#R7circuit-analysiscurrent-sourceinstrumentation-amplifierbiopotentialsignal-generatorShareCiteEditFollowFlagasked 2 days agoast5jx3122 bronze badges New contributor

• Welcome and nice to meet you. (1) Your proposed Bioimpedance Analysis project is very interesting. I skimmed the paper and made a quick and dirty reading log. (2) You question is this: How come we can measured the current if there is “no Load“? My brainstorming reply is this: (a) The power/signal source is actually a variable/constant voltage/variable constant current source, / to continue, … – tlfong01 2 days ago    
• (b) This means that, (I am only 50% sure), even if no load (ie, no physical human brain as load), there is still a source/series resistor/impedance (sort of current sensing resistor) which is used to measure voltage and therefore indirectly also current (by Ohm’s Law). / to continue, … – tlfong01 2 days ago    
• 1(3) I am thinking of writing up a simple answer as an introduction to bioimpedance analysis for electronics hobbyists/newbies. But first thing first, I need to know your electronics background knowledge and skills. You might like to let me know if (a) you know how to use a multi-meter and oscilloscope, (b) How to use Arduino or similar MCU, or Rpi or similar SBC to blink a LED, (c) How to use a CCS (constant current source) to light up a 12V RGB LED stripe? – tlfong01 2 days ago    
• 1(4) In case you are not familiar with AC circuits and impedance etc, you might like to read the following tutorial to get a rough idea: (a) AC Resistance and Impedance – Electronics Tutorials electronics-tutorials.ws/accircuits/ac-resistance.html, (b) AC Inductance and Inductive Reactance – Electronics Tutorials. Cheers. electronics-tutorials.ws/accircuits/ac-inductance.html – tlfong01 2 days ago    
• 1(5) The Howland Current Source is a bit complicated, and definitely not for for newbies, especially you are playing with AC input. Anyway you might like to some reference materials before you decided to dip your toe into the water. (6) Ref: Application Report AN-1515 A Comprehensive Study of the Howland Current Pump (Source) – TI, 2013apr (6) ti.com/lit/an/snoa474a/… – tlfong01 2 days ago   
• 1@tlfong01 Thank you for the comments. Yes, I think I have sufficient experience using multimeters and oscilloscopes, also I have used the TI MSP430g2553 launchpad to blink LEDs and other relevant projects, but I am new to the concept of current-controlled sources in applications and understanding calculating the complex impedance in a multi-frequency circuit. – ast5jx yesterday
• Thanks for letting me know your background knowledge and skills. I am glad to hear that you are using TI MSP430 which I think is more professional and advanced than the hobbyist grade Arduino and Rpi. And as you said, you are weak in (a) CCS, (b) complex impedance concepts. So I would suggest you to google and wiki introductory tutorials on these two topics. You might need a couple of days to weeks to get a rough idea to catch up. It is only after this learning step then you can understand the Howland Current Source which is the crux of the matter. – tlfong01 14 mins ago   
• I have no practical experience at all in Howland CS. My only relevant experience is on weight sensor using load cells and HX711 24-bit ADC to amplify and analyse very weak Wheatstone Bridge signals. Perhaps I would come back in a couple of days to let you know my thinking of how you can start off you bio-impedance project. Cheers. – tlfong01 8 mins ago   
• I am skimming some of Howland CS app notes I googled and found them a bit hard. Below are two examples. (1) Application Report AN-1515 A Comprehensive Study of the Howland Current Pump – 2013apr ti.com/lit/an/snoa474a/… (2) Application Report Analysis of Improved Howland Current Pump Configurations – TI 2020oct ti.com/lit/an/sboa437/…. I don’t recommend you to read it now. – tlfong01 2 mins ago   Edit   

Add a commentStart a bounty

1 Answer

ActiveOldestVotes2

Also, they say they are doing a no-load current test in which they measure the current with a DMM.

The op-amps create a current source. You can imagine the current coming out of pin 7 of the LF412N, through the 51K resistor, across the electrodes and then back to ground.

For a no-load current measurement you can short the electrodes and measure the voltage across the 51K resistor and divide by 51K. You could also use a multimeter between the electrodes. With the function generator operating you’ll be measuring AC volts or currents.

I am trying to generate signals with a function generator for Vin- and Vin+ but the paper hasn’t discussed what inputs are needed, for example, the voltage amplitude, frequency, and Vin+(-) phase difference.

The INA128P is a instrumentation amplifier. The supply rails are 0V and 12V and the resistor RG of 51K sets a gain of about 2 (really 1+50k/51k) as derived from page 12 of the datasheet.

Based on the supply rails, you’ll want to have your input signal centered around 6V — i.e. for a sine wave set the DC offset to 6V. Since the gain is 2 the amplitude should be limited to 3V.

The video Input Range of an Instrumentation Amplifier explains what goes on inside an instrumentation amplifier and how to make these calculations in general.

For the actual connections to the function generator you could do the following:

• connect a constant 6V signal into VIN-
• connect a sine wave with a DC offset of 6V into VIN+
• connect the ground of the function generator to the ground of the circuit

As for the frequency range, it seems they used frequencies in the range of 100 Hz to 100 KHz.

Update: I just realized that VREF of the INA218 is tied to ground so you’ll want to arrange things so that VIN+ is always >= than VIN-. To accomplish this, just adjust the DC offset of the sine wave so that it’s lowest point is at 6V. This means the DC offset you use will depend on the amplitude setting. The output of the INA218 on pin 6 will be (VIN+ – VIN-)*G where G is the gain factor (around 2.0 in this case). Changing this voltage will affect the amount of current sourced from the LF412N op-amp.ShareCiteEditFollowFlagedited 2 days agoanswered 2 days agoErikR1,28066 silver badges99 bronze badges

• 1Appreciate the detailed explanation. 1) For the “No-load current test” you said I should short the electrodes and measure the current of the gain resistor (R_G = 51k), but doesn’t “No-load current test” mean having an open circuit test at the output i.e. leaving the electrodes open. And I see there is a problem in doing this, with an open circuit there is no current flowing to measure. So can you explain why you are measuring the “no-load current” over the gain resistor? 2) For the values you discussed to be used in the function generator why are we trying to maximize the voltages? … – ast5jx yesterday
• 1… I know the I-Amp reads the difference between input voltages but shouldn’t we carefully choose voltages that could affect the current in electrodes in order to not to t hurt the subject? For example, from 15uA to 50uA. – ast5jx yesterday
• 1If the electrodes are open there won’t be any current to measure. You measure the current through R5 not RG, Unfortunately both are 51K in this schematic. The current source output is pin 7 of the LF412N op-amp. – ErikR 22 hours ago
• 1Yes you should adjust the current for an appropriate current. Measure the Howand current output when VIN+ and VIN- are receiving the same signal (like a constant 6V). This will be a the base level of current supplied. When you send a sine wave into VIN+ the Howland current will be modulated. The amount of modulation will depend on the amplitude of the sine wave. The exact relationship is something you should just measure. – ErikR 22 hours ago

Add a comment