..is in the implementation of the "receptors", this could be costly..
xanadunow
.... Right .... getting the sensors to detect enough of a pulse signal to register .. without being activated by spurious light noise, is the first problem to look at.
Maybe place the photosensors at some modest depth in a well or some type of light shield.
The one significant dimension to look at is the linear distance along the great circle path that is required for the spool to make one rotation on its own axis.
You would want to have the optical sensor array span, or maybe overlap this distance.
... It would be required to have at least one 'solid' sensor detection per spool revolution ...all zeros or all ones being the machine halt condition... If going analog, then you need some value that is significantly greater than zero.
... Increase the signal detection level by placing more than one reflective facet on the spool, maybe in pairs for symmetry.
Consider using some sort of 'spool arrival' sensor, to produce a signal to initiate the accumulation process, and maybe a 'spool departure' sensor signal, to start the comparison calculation ....
If you manage to get the initial configuration to perform as intended, you might even be able to eliminate all the laser/sensor arrays except for one.
That is, you would sense the passing by of each of the four spools, in sequence. This would probably depend on your hardware choice .... An all analog version might be fast enough .... You would have to have some numbers for the spool rotation rate about its own axis, as well as that of the 'great' circle axis, and then calculate the time delay between each spool.