The interrupter and the coil are joined by an optical fibre, and not for convenience. There must be no galvanic connection between the logic and a machine that throws metre-long arcs.
Which leaves one rule, and it is a requirement rather than a preference.
A broken fibre must turn the coil OFF
Fibres get pulled out, broken, and shaken loose. Transmitters lose power. Controllers hang. In every one of those cases the bridge has to go quiet rather than stick on.
A divider does that: with no signal it holds the input in a known state, and that state is off.
no light -> divider (pull-up ~1k / pull-down ~10k) holds the input at a defined HIGH
-> the comparator's output goes LOW
-> the switch stays closedThe pattern is simple and it is not optional.
Test it physically. Pull the fibre out of a running machine at low power and confirm that it stops. Looking at the schematic is not the test.
Active-high and active-low are not interchangeable
OPF-2412T, HFBR-2412T, IF-D95T and IF-D95C are active-high. Buying an active-low part by mistake is easy, and it inverts the polarity, which stands the entire fail-safe logic on its head: a break starts turning the coil on.
Check the datasheet, not the package.
The pulse width the link can actually pass
What people use instead, and the catch
- Optocouplers,
TLP2345,TLP2362,H11A1. Cheap, compact, fine for a handheld. The isolation is weaker than fibre. - Bluetooth, two
HC-05modules, as paulsimik and LoFoTroFo have done. Send only the potentiometer positions and compute the ramp and the PWM on the coil's own board. Shield it. MIDI over Bluetooth is hard, there is not enough speed.
The symptom, if you ever see it
Pulled the fibre and the machine kept running: either there is no fail-safe divider, or the receiver is the wrong polarity. Stop everything and fix it before the next run. There is no version of that fault which is acceptable to work around.
What the interrupter is doing on the other end of that fibre is bangs per second and ontime.