Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.
Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.
It would appear you can use the AD5933 to directly measure the bio impedance if it's between 1kΩ to 10MΩ. You will need to interface it to a microprocessor or computer with a serial I2C interface to read the measurement data.
Hi crutschow, thanks for your reply. Do you have any circuit in mind so I can try to make it? Do I need a software too? I don't mind using a different circuit also.
Hi crutschow, thanks for your reply. Do you have any circuit in mind so I can try to make it? Do I need a software too? I don't mind using a different circuit also.
The easiest would likely be to use a USB-to-I2C adapter such as this or this. That should allow you to retrieve the data from the chip with the available software and view it on a PC.
If your in school, you can see if your school has a copy of Labview which is one possibility. VB.net might be another and so might Python. In any event the software you use on a PC will largely depend on what's available.
An I2C to COM port will almost inevitably be in the mix.
Since you are talking about connections to the human body, a USB isolator will have to be used where the circuit side would be battery powered.
I know nothing about bio-impedance such as frequency related, values, a function of time. Nothing. I suppose I could do that, but that's part of your research.
About the only thing it does tell me is that it's a resistivity and may be AC based and not a resistance. Resistivity is based on R = pL/A where R is the resistance, p == resistivity in units such as ohm-cm. L is a length and A is an area.
There are techniques such as Van der Pauw that can measure the bulk resistivity of a sample with 4 electrodes, but I would imagine that the voltages would be high. You end up taking 2 I-V curves and then rotating the probes (electrically or physically) by 90 degrees and then applying a correction. The correction depends on the symmetry.
Another technique called the 4-point probe requires one to know the thickness of the thin film.
You have to do some research and find out what you need. Bio-electrical stuff, unless done on dead things require the utmost care.
Initially, I was thinking that your meter might use two probes and give a number, but thinking about it, that doesn't make sense. A 4-probe or 4 electrode technique might but then what voltages and currents are needed or expected.
You need to do some research and then ask some intelligent questions. You need to define the problem and not something like "I need a light bulb". "for my fish tank" won't cut it either. Metal Halide, UV, Florescent, Grow lamps, timer whatever.
So, far your problem reads like "I need a light bulb". or maybe I need to use this chip. Same difference. There's nothing substantial and my best response is "So....."
LabVIEW is sometimes available at Universities. It's a language that is designed to interface to instruments. There is a student edition. It's a very unique language because it is "Data-flow" based. Routines called "Virtual Instruments" execute when all the data is made available to them. The language make muti-processing cake, but it makes error handling much different. The language is ICON based. You wire components like you would a schematic.
KeepItSimpleStupid, thanks for the advise. What I thought is: a circuit based on ad5933 for example the evaluation board and a I2C to usb to connect to pc and a circuit to use a 4 electrode system to measure the bioimpedance of the mucle of an arm. Is that more accurate do you think? Do I need a special software for that? and the circuit for the electrodes?
Thanks
For one:
Programmable output peak-to-peak excitation voltage to a maximum frequency of 100 kHz
Programmable frequency sweep capability with serial I2C interface
Frequency resolution of 27 bits (<0.1 Hz)
Impedance measurement range from 1 kΩ to 10 MΩ
Capable of measuring of 100 Ω to 1 kΩ with additional circuitry
Internal temperature sensor (±2°C)
Internal system clock option
Phase measurement capability
System accuracy of 0.5% 2.7 V to 5.5 V power supply operation
I've placed in bold a few things which I can't answer. You also have the AD5934 to compare to.
Then there is isolation to deal with. I have one of these: **broken link removed** and you probably want battery power on the person side.
Now, the analyzer has only three leads. How does one turn that into what I would think of is needing 4 required probes? I-, I-, V-, V- similar or something like an LCR meter has HIpot, HIcurr, LOWpot, LOWcur. How much resistance can be between the leads?
Do you need to do sweeps? What frequency? Is the voltage high enough? Might this be a 4 electrode patch? Can it be done with 3?
The 4 point probe https://www.electro-tech-online.com/custompdfs/2012/08/4ptprb.pdf method gives sheet resistance. Here there is 4 points in a line and the current is passed through the outer ones. Then a voltage is measured on the inner ones. Then the current is reversed to detect for non-linearities. This method finds the "sheet resistance". You need thickness to get resistivity. And again, the force and sense pads are different.
I'm missing a lot of things because I don't really know how bio-impedance is measured and I don't know how contact resistance is handled, nor do I know what to do with 4 probes and 3 wires from the IC?
Hi KeepItsimpleStupid, thank you for all the information especially the
I am a bit confused with all the details you gave me. Can we start from a circuit? Can I send you some files with email or message?
Thanks
The link isn't valid. You have to do a "Copy link location", but I now what you referenced.
You see, I'm confused too. For one, the use of ankle and hand points are almost like use the extremeties for 4-point probe measurements. They also look like they are using a 4 point method needing HIcur, HIpot, LOW pot and LowCUR.
You mentoned 4 points.
1.....d......3
2.....d......4
Now, it might be possible to use an arrangement like the above where d is a constant, 1 & 2 are closely spaced and 3 and 4 are closely spaced.
In-line is also a possibility: I+....V+.....V-...I-
But, again you need 4 separate probes. I'd probably prefer the latter arrangement.
Just like there is the Kelvin measuring technique for measuring low resistances, there needs to be a 4-wire technique for 4 terminal Z measurements.
The AD device is 3-wire and therefore if there is one electrode which would only get one wire and therefore that wire would not be compensated.
In the case of the 3-wire RTD, contact resistance is not a problem, but for very long wire runs, the wire resistance may not be the same.
I would think that you basically would need a custom skin pad where the spacing of 4 linear electrodes are controlled AND you need a 4-terminal measuring technique.
Right now, I don't have a clue to create that 4 terminal system.
The Ad5933 apparently uses Vdd/2 as a reference voltage from the evaluation board schematic/. I'm not sure what that means.
I might be as easy as buffering the two voltage feedback points on the evaluation board. Buffering to the extent that the source/return is the outer electrodes and the feedback is the center electrodes referenced to Vdd/2.
How you were thinking to arrange the electrodes and how you were expecting to connect them?
My suggestion is to have 4 linear spaced electrodes and use the the 4 point probe technique to find resistivity. I would consider a custom sort of probe. One method might be to use a spring probe such as these: **broken link removed** allied and newark should have them and find a way to sterilize them I don't think you want a large area probe, but you should apply some electrode gel to the tips. You could drill a small piece of PCB for the probes and attache via a Velcro band. Spacing is your responsibility.
It would be really nice if most of the electronics were "wearable" except for the USB connection. e.g. just make the contact block replaceable. You would have to address repeatability.
So, the real trick is: "How to make the AD network analyzer device be 4 terminal" with a fair amount of resistance between force and sense. PDF page 1, here: https://www.electro-tech-online.com/custompdfs/2012/08/UG-364.pdf gives a better clue as to what may be needed.
I'm almost thinking that you can buffer the Feedback loop on the top OP amp and have that be HIpot and the output being HIcur and use the lower circuit as is, where the mesurement point is LOWpot. Now, you may be able to force the last of the inline dots to be LOWcur, by just forcing it to be Vdd/2.
You always want to be able to change the direction of current flow and compare the results. 50-100 kHz wouldn't be a problem for the AMP-02, although you may need a DC-DC converter to get a bipolar supply.
The bigger trick is what modifications are necessary to turn the evaulation kit into a 4 independent contact system.
I might not have the answers, but I do know what questions to ask. part of that came from troubleshooting scientific instruments that I did not understand nor could operate myself, but I was able to fix them by using the knowledge of the researcher.
Well, it definitely addressed some of my concerns, but it left the Analog Front End as a black box. What I would suggest is to use the email in the article and ask nicely if he can provide more info on the analog FE. If this is for non-commercial purposes you may get a more favorable result.
The hardest thing to understand is the offset compensation, so I'll start there. A sine wave integrated would be zero. Whn OP-amps are involved there is always an offset current. So, if you assume the input is an AC signal with some offset, the integrated part will be the DC offset.
There is a monitor which is easily understood.
Then the whole point is to use a miniscule amount of current from the desired current, so the feedback loop is buffered. So, if you wanted say 1 nA of current and the bias current of the amplifier is in the fA region, you only taking a very tiny amount from the desired current.
Watch terminology like buffer. This means x1.
They managed to pull this off to 100 Mhz which is impressive.
Does this have anything to do with tooth impedance?
Thank you, I know the book, it is very good but it is ore theoretically. So if a evaluation board ad5933 is used how the 4 electrodes part can be made using op-amps or something similar? And then how we get the results? with Labview?
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
