0
\$\begingroup\$

I'm trying to build something quite challenging and seem to be getting nowhere. Maybe you'll help. I'd like to detect a touch of two naked fingers with no materials or sensors in between (simple capacitive sensors won't work here). It's a design idea and I'm not very skilled engineer so I have no idea if it's even possible.

Here's a simplified sketch of a hand:

enter image description here

The original idea was this: There's an electrode on each finger and small current is running through from A to B. When fingers touch, current starts passing through the SHORTER GREEN path as well (it may not be apparent but it will be always a shorter by at least ~1cm). But it turns out looking for changes in resistance / current seems to be going nowhere. It's not reliable because the biggest changes in resistance come from contacts between the skin and electrodes. It won't ever be perfect, electrodes move, user sweats, temperatures change, etc. Compared to this, the changes in resistance caused by a shorter path are insignificant. But the contact may be good enough for something else (it would be fine for a simple capacitive touch switch).

I'm now thinking that maybe I could measure the distance between the two contacts somehow by sending some signal from A and reading it at B. Just looking at the time of propagation of charge imo won't be possible because I won't be able to do this fast enough to get at least a millimeter resolution.

The problem could be really described as detecting whether the shorter green path is being used or not.

(We could add more electrodes on the fingers somewhere along the blue path but not along the green path.)

Do you think something like EDM (http://jerrymahun.com/library/Distance/a.htm) could help here? Any other ideas?

\$\endgroup\$
  • \$\begingroup\$ Welcome to SE, Mirek. Tell us what the required end result is and we will understand some of the constraints to the question. i.e., What are you trying to make? The quality of answers will bear a relationship to the quality of the information provided and the question asked. \$\endgroup\$ – Transistor Jan 14 '16 at 23:40
  • \$\begingroup\$ Can't you use a webcam and do it via image processing? \$\endgroup\$ – Armandas Jan 14 '16 at 23:44
  • \$\begingroup\$ It's a new wearable mobile controller, apologies for keeping it maybe too short. Not using cameras is the core of my approach, I believe inertial tracking can deliver much better experience and portability. This is actually just a tiny part of the whole solution, but it's been giving me headaches for a while. \$\endgroup\$ – m4pcy5t0 Jan 15 '16 at 0:56
1
\$\begingroup\$

You could do this with three electrodes on the fingers. Two would establish an excitation and the third would be an input to determine if current is flowing down a finger. Conceptually, something like this:

schematic

simulate this circuit – Schematic created using CircuitLab

In this case, V1 is a low voltage AC source, preferably of a frequency much higher than mains frequency so that mains hum can be filtered out and sufficiently low voltage as to not cause problems with the skin or safety in general. R1, R2 and R3 represent the electrode-to-skin impedances (I think that's roughly right for EKG type electrodes). The 10K resistances represent the impedance of fingers and hand.

So, before the top (blue) finger touches the bottom (cyan) finger, there is a bit of AC current flowing through the hand (R5) but very little current flowing through the blue finger. When the fingers touch, an AC voltage appears at A and can be detected by the AC voltmeter VM1. Ideally VM1 is sensitive to the frequency from V1 and insensitive to other frequencies.

There are potential dangers associated with anything connected in this sort of way to a human body, and this answer does not attempt to address those issues. Probably if power is supplied solely from something like a 3V CR2032 battery (no mains connections or adapters or power supplies or computer connections other than wireless) some of the issues will go away, but this aspect requires careful evaluation by someone familiar with the proposed setup and biomedical safety issues.

\$\endgroup\$
  • \$\begingroup\$ Wow, that looks impressive, though I have no other idea whether it can actually work :) I'll try to build it and report the results. Thank you for the speedy help, much appreciated! (it will be battery powered, 5V for the prototype, probably 3V in the final product, your safety concerns are spot on and I'm keeping it in mind, I hope it won't be an issue) \$\endgroup\$ – m4pcy5t0 Jan 15 '16 at 0:51
  • \$\begingroup\$ So I've done some research and this won't work. The biggest problem is that a hand works like a parallel circuit and here you expect resistance like in a series circuit. There's always a parallel link between A/B/C. However, this was quite inspiring and your suggestion to use AC was most helpful (so far I've only thought of DC because batteries). AC is supposed to penetrate human body much easier because it's constantly changing. Also, it led me to this new idea: electronics.stackexchange.com/questions/211957/… \$\endgroup\$ – m4pcy5t0 Jan 18 '16 at 2:51
  • \$\begingroup\$ It may work, the signal will just be small (perhaps too small to be reliable). I got some 10's of uV change in a crude experiment, but I don't have access to proper EKG electrodes. A 4-electrode arrangement would be much less sensitive to skin resistance. \$\endgroup\$ – Spehro Pefhany Jan 18 '16 at 3:18
  • \$\begingroup\$ Well, the contact will always be poor. I can't use EKG electrodes in the design, it will be rather just something touching the skin but presumably moving, with varying pressure intensity. Therefore I can't rely on small changes in voltage \$\endgroup\$ – m4pcy5t0 Jan 18 '16 at 3:25

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.