I am trying to replicate the circuit in this paper: http://chrisharrison.net/projects/electrick/electrick.pdf

I have followed the schematic in Figure 7 and utilized Howland current source as described in the prior post: Building a VCCS with AD8220 Instrumentation Amplifier

Currently, I have an AD9850 with a 1 Vpp 100 kHz output. This wave is passed through Howland current source, which should yield a current < 1 mA.

My question now is how is the EIT field established? I am trying to understand based on Figure 1 in the paper, but I do not know how that translates to my setup. Are all the adjacent electrodes connected to each other? When I establish a current receiving pair, is Iout going to one electrode while the adjacent electrode is connected to ground?


The caption from figure 1 of the paper you link to

Figure 1. Left: we insert current from one pair of electrodes and measure voltage from other pairs.

This means that the current is injected between an adjacent pair of electrodes. Voltage is sensed between all the other pairs of electrodes.

I would expect this to be iterated, so that each pair gets connected in turn through the multiplexer shown in figure 7 to the current source, while all the other pairs are sensed for voltage.

It doesn't matter if one of the current driving pair of electrodes is grounded or not, all that matters is that a current flows between them, and that the differential voltage is sensed between adjacent pairs of voltage sensing electrodes.

I've no more than skimmed the paper for figures 1 and 7, so don't know whether the authors go into detail about how to reconstruct the finger position from the readings. That is the tricky bit. The obvious naive route would be some form of pattern matching, whereby a measurement set is taken with no finger present, and then one with a finger touching, and the difference, suitably scaled for the drive level, is stored. The same thing would be done with fingers in different positions. Then you'd examine the differences and see what systematic alteration was happening. I am sure there are more sophisticated mathematical techniques available.

  • \$\begingroup\$ Thank you very much! I have attempted what you suggested and connected my Iout and ground to one pair. I am not detecting any voltage from the other electrode pairs so I might have an issue with the material itself. For my purposes, I dont need exact position, just need to know if the platform was touched \$\endgroup\$
    – KIchigo
    Oct 26 '18 at 13:34
  • \$\begingroup\$ I have the whole thing setup but am having another issue. I have the appropriate current/voltage (42 mA with 0.5 Vpp). However, when I connect the current source to my platform, the signal disappears. My platform is a flat surface that has sprayed with copper conductive paint. I had one electrode of my current receiving pair connected to I(out) and the other electrode connected to ground \$\endgroup\$
    – KIchigo
    Oct 30 '18 at 7:48

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.