How can body impedance measurements work with the drastic changes in skin ohms?

[corrado33]

I've been looking into accurate scales lately, reading up on technology and what not. So, I know how impedance works, I'm an electrochemist and I use it in my experiments. But my question is this. With the drastic changes in skin resistance (wet, dry, humid, not humid, calluses, etc) how the heck can a body impedance measurement be semi-accurate? Heck, I can get my skin resistance to be anywhere from megaohms to kohms. Do they simply throw out the bulk resistance? (Resistance due to wires, contact resistances, etc?) What part of the spectrum do they look at? I can't seem to find examples of body impedance measurements, and I'm not about to hook up low impedance electrodes to my body to find out. (I'm much more confident in my skills as a chemist than my skills as an electrician/hobbyist.)

(Oh I'm not buying a scale because of accurate BMI predictions, I'm pretty lean, so BMI predictions are ALWAYS wrong for me.)

EDIT: Let me be a bit more specific. I'm talking about the body impedance measurements performed by scales, not by trained professionals with sticky electrodes and conductive gel.

[Andy Watson]

With the drastic changes in skin resistance (wet, dry, humid, not humid, calluses, etc) how the heck can a body impedance measurement be semi-accurate?

If you want to get a reasonably reliable measurement you use a four wire method. Typically, you would inject a known current and then measure the voltage (somewhere else). As you have discovered, the biggest variable is the contact resistance of the electrodes - you need to get around that.

Do they simply throw out the bulk resistance? (Resistance due to wires, contact resistances, etc?)

Depends what you want to get at! The four wire method eliminates most of the "external" variables, then you have to have some sort of 3D model of the part that you are trying to measure - or make measurements relative to previous readings.

What part of the spectrum do they look at? I can't seem to find examples of body impedance measurements,

Generally we were interested in a region of between 1kHz and about 500kHz, but it varies with the type of tissue that you want to measure. Look up Electrical Impedance Tomographic Spectroscopy (EITS). Try http://discovery.ucl.ac.uk/9615/  it used to be a repository and jumping-off point for all things EITS.

The consumer weighing scales are usually little better than random number generators. If you are lucky they will make a multi-frequency measurement and use the relative impedances to try to differentiate between lean resistive tissue and fatty, capacitor-like tissue. They use this ratio to estimate the relative proportions of lean and fat material. As you have sussed, the biggest variable is between people - they're just not made consistently

Edit: The link that I provided appears to have changed significantly - it used to be a hub of everything that pertained to EITS.
Try this link. Towards the bottom of the page there is a plot that shows the typical frequency spectrum for soft tissue. Obviously the details will be differnt for fatty tissue - but the idea is the same:
https://zilico.co.uk/research-development/electrical-impedance-spectroscopy.html

[Galenbo]

...I know how impedance works...With the drastic changes in skin resistance ... body impedance measurement be semi-accurate...skin resistance ... megaohms to kohms...Resistance due to wires, contact resistances, etc?) What part of the spectrum do they look at?...body impedance measurements... low impedance electrodes ...body impedance measurements...

Please unmix impedance and resistance. Like unmixing sound properties and battery voltage.

Impedance isn't expressed in Mohms or Kohms, and sometimes->often doesn't get influenced by wet/dry contacts or wire resistance.

[corrado33]

...I know how impedance works...With the drastic changes in skin resistance ... body impedance measurement be semi-accurate...skin resistance ... megaohms to kohms...Resistance due to wires, contact resistances, etc?) What part of the spectrum do they look at?...body impedance measurements... low impedance electrodes ...body impedance measurements...

Please unmix impedance and resistance. Like unmixing sound properties and battery voltage.

Impedance isn't expressed in Mohms or Kohms, and sometimes->often doesn't get influenced by wet/dry contacts or wire resistance.

Sorry, this is incorrect. Impedance is a complex number, the real part of which IS resistance and is measured in ohms. Taking an impedance spectrum of a system with increasing bulk resistance will show very obvious changes in the x intercept of a nyquist diagram.

[Galenbo]

I should have said impedance isn't expressed as a real number in Ohms.
For the rest, the statement still stands. Unmix.

You can of course use the word impedance when you know it only contains the real part, and resistance when you know it's a complex number.
But that won't help you.

[Galenbo]

Taking an impedance spectrum of a system with increasing bulk resistance will show very obvious changes in the x intercept of a nyquist diagram.

Of course, but what is the relation you discovered in the changes on a pole plot to witch part of the impedance at witch frequency of the skin resistance ?

[Karel]

Have a look here: https://www.eevblog.com/forum/projects/high-sensitivity-bioimpedance-measurement-analog-front-end/msg847764/#msg847764

[TimFox]

In general, for a linear analysis with alternating current, one measures a complex quantity:  impedance (Z) or admittance (Y).  Both components of a complex quantity must have the same units.
The magnitude of complex impedance is measured in Ohms.  The real part of the impedance (resistance) is measured in Ohms.  The imaginary part of the impedance (reactance) is measured in Ohms.
Sometimes the parallel circuit is more useful.  The magnitude of complex admittance is measured in Siemens.  The real part of the admittance (conductance) is measured in Siemens.  The imaginary part of the admittance (susceptance) is measured in Siemens.