For Dunhill I designed the electronics for a prototype device that was basically a stress-busting feedback system. It was a beautiful, small, silver, shell-like object designed by Andrew Lang. In use it was held between finger and thumb and had a heart-beat throb. The idea was that you were able to learn how to slow down the beats by subconscious means and, in so doing, reduce your stress. This was a very simple electronic design for me but the design problems were enormous and mostly involved the contact resistance of finger and thumb to the silver shell.
When electrodes are connected to patients in hospital the skin is smeared with KY jelly to promote good and consistent connections to the skin. This was not possible with this product. What this meant was that in every different climatic condition and for every different user the device would behave differently and erratically. You would not measure the state of stress of the user, you would measure the state of electrical connection of the device to the skin of the finger and thumb. The stress component was swamped by the large variation in contact resistance finger to silver. For such a simple device it was frustrating that these ordinary practicalities were such a problem. You just can't smear a beautiful silver object with KY jelly! I found, after much research and experiment, several tactics and software techniques that mostly solved this problem.
Firstly, during the first 15 seconds, finger and thumb against silver started to sweat a little bit at the contact point and that was responsible for an exponential decline in contact resistance to reasonable stability - given constant finger pressure. However, person to person differences were large and I had to design-in a system of self calibration. This could only be started once the 15 seconds of stabilisation had happened.
My program, in the device's processor, would then make a series of tests over a given time to discern what normality for that person was and then adjust everything to suit him/her such that the device behaved in a standard way. After calibration it always started off with a beat of about one per second. The act of holding the device between finger and thumb turned it on but unfortunately, the first beat occurred 25 seconds later because of the calibration cycles.
I designed a proximity charger for it, which only had to be brought very close for the battery to start to charge. A maximum distance of about 1mm from the charger's coil to the device's coil would ensure a good charging current. Basically it used a loose-coupled transformer operating at about 100KHz..
(c) Spencer Childs - website by spencer childs spencerchilds spencer.childs@el-design.y
