Driving the large number of coils used in a tactile display is not particularly difficult to do, but how to do it well and still keep cost down? Note that the discussion below is going to be a bit technical…
For the mobile (low power) version, multiplexing is an attractive solution, made possible by the low duty factor of the actuators. This type of driver has actually been used in the proof-of-concept display. Using a relatively small number of control signals and components it allows for high currents, high efficiency, and PWM (Pulse Width Modulation, providing per pixel intensity levels). All proved to work just fine for a low power display.
Large full power displays are another matter. Only limited multiplexing is possible (with an up to 50% duty factor) so component savings are going to be small and may not be worth the added complexity. Also, limited (two-way) multiplexing in combination with the large number of actuators means a lot of them are going to be switching at the same time, which will have a negative impact on signal integrity and EMC.
Taking this into consideration, driving each coil with its own separate driver appears to be a decent approach but it remains a costly one, as it will take 144 outputs and drivers for a standard Optacon resolution. A simple and obvious implementation would use a microcontroller with 144 GPIOs, but a controller with that many IOs will come in a large high-density package and require a relatively complex PCB, which is not attractive for low volume production. Also, properly separating logic (sensitive) and display (high switching noise) can be cumbersome with this approach.
A classic solution to these problems is to separate the outputs/drivers from the controller using shift registers. These are generally available in suitable packages, and varieties with integrated drivers exist (which may well be the best option for a final version when all design parameters are known). For now I prefer to work with external drivers, allowing for more design freedom, larger currents, and higher reliability.
Suitable shift registers and supporting components for the upcoming prototype are on order.