More & more my firm has been asked about touch screens - in particular the low-cost, simple resistive ones.
This writing aims to provide basic guidance & (hopefully) helpful detail - while describing, 'How a laboratory Rat convincingly BEAT a (hapless) human in an offical, 'Touch Screen Speed Challenge!' (read on for that detail...Tech does not (always) have to be 'Sleep Inducing.')
It is assumed that your '4-wire' is a 'Resistive Touch Screen.' Such Touch Screens employ '2 Orthogonal, Resistive Planes' - and (most always) during operation one plane is 'powered' and the second one - then 'read/examined' for 'Touch.' Often the Screen's resistance is high enough that an MCU's GPIO may drive it. However my firm's preference is to employ a pair of 'small signal FETs' - in that Screen Powering role.
In '4-wire touch' it is normal for power to be impressed across opposite sides of each touch plane. When a touch occurs near to the 'ground rail' edge of that plane - the voltage output is minimal. When the touch nears the positive supply rail (3V3, if possible) the voltage approaches the supply's (+) rail. It is the 'undriven plane' which is 'read' - one wire ties to ground - the other 'feeds' an ADC.
To simplify - note that 'in reality' - each plane becomes 'Regularly Switched between Powered and Voltage Sensing.' Thus - there must be effective and active switching - forced upon EACH (+) supply touch wire - which is switched between the supply (when that plane is powered) and the ADC (when that plane is being read/sensed.) The identical treatment is implemented upon the 2nd touch plane. Fortunately - the 'ground wires' of each touch plane - remain - 'tied to ground.' (it may take several reads for this writing to 'sink-in' - although staff 'nods agreement.')
Now the 'Switching between roles' is properly timed & forced upon each Touch Plane. There exist 'integrated Touch Screen Controllers' - which prove expert at this. Those would be my 'sure' recommendation. Otherwise - you must regularly, 'Test for a 'Touch' - and when recognized - toggle between power & voltage read - at each Touch Plane. Many designs 'always toggle' - which my group believes 'wasteful AND EMI generating' - thus we toggle only upon Touch Detect. (i.e. one touch plane is set to 'Idle in 'Read Mode' - and initiates the process of 'Plane Toggle' upon 'Touch Detection.' Rather clever - (several) clients have noted.)
Our firm's method of 'Screen Calibration' involves, 'Systematic Touches at each Screen Edge and Screen Center - while first one - and then the other Screen Plane is read/sensed. (i.e. one plane is powered - the other is then read - the pressure of the 'Touch' proves sufficient to impress voltage upon the 'undriven plane.' (i.e. the orthogonal planes come into 'contact' at the 'touch site!')
The above proves 'well & good' for one plane - yet you have two. And thus that (same) procedure is now repeated - but for the 'reversing of the roles of each plane. By carefully collecting the voltage levels at each 'touch location' - it becomes possible to 'Map the Touch Location to 'Display Coordinates.'
By convention - the 'Left, Bottom' portion of the Screen (usually) becomes '0,0' Location. Thus the Right, Top assumes 'max value' - and one is free to assign values to that. Note that the 'normal' touch screens may undergo (some) change in resistance with, 'lot, aging, use & temperature.' When 'high accuracy is demanded' - it proves wise to provide 'Touch Screen Calibration' - in which the user will 'touch key targets' - which are generated & placed 'on-screen' - to receive such touch. Software can then make the necessary modifications (if required) to re-establish 'Accuracy of Touch.'
Calibration parameters - usually '5' - should suffice. (4 identify each display corner, 1 marks display center) It is normal/customary to determine the 'reasonable area' over which a 'touch extends' - and then to use that measure to determine 'how many' (legitimate) 'Touch Targets' may be placed. It is wise to include a 'safety margin' (i.e. 'dead zone') between adjacent 'touch points.' The touch targets may then 'slide' - either left/right or up/down (or both) - to more precisely overlap your display's 'pixel target.' (these announce the 'target for the touch') It is believed that these methods - most usually - enable you to generate & detect multiple & efficiently sized/placed 'touch targets.' (alas - only 'one' at a time) Relating the ADC voltage - to each target - should not prove (that) difficult - and may resolve via a (near) linear 'Screen Location to ADC Reading formula.'
It is hoped that (some) benefit results for 'touch users.'
Kindly note that this writing was 'composed quickly' in answer to a post which arrived w/in the hour - and which vendor noted - "was not expected to receive 'quick response.'" Follows - the link to that post:
https://e2e.ti.com/support/microcontrollers/other/f/908/t/838821
Feel free to: 'Dispute, Agree, Delve Deeper (even) **LIKE** (although that - may prove difficult!) It has been 'years' since I worked 'seriously' w/such screens - newer (primarily Capacitive Touch) offer significant benefits - yet are more complex & costly.
It may prove of interest to note that one of our past (volume) Touch Installations saw these Touch Screens (minus any display) deployed as the 'flooring' for 'Laboratory Animal Cages' (many cages) and the animals' movement, activity level & positional 'favorites' were regularly monitored & logged. Note that rats (and pigeons) were quickly able to 'grasp' Touch Screen Usage (these 'off the floor' & supplied w/a Display) and to 'Reward themselves' via the 'Correct Sequence of Touch Presses/Pecks.' (even when - and especially when - that sequence was (deliberately) altered!)
It must be noted that (some) rats & pigeons 'beat' certain humans in 'Speed of 'Sequence Solving!' I'd love to say that 'said human' remains unknown (here) - yet that's no longer true. Note too that 'unkind' and/or 'rubbing it in' commentary from past 'rodent/bird victors' - are 'subject to edit.' (or may have 'gotten lost' (mysteriously) - in the mail...)
