Coin Comparator Token Rejection Loops: Optical Sensor Calibration Fixes
A coin comparator stuck in a rejection loop is almost never a dead board—it is a blind optical sensor crying wolf. The moment a token drops through the acceptor and bounces straight back to the player, you are looking at a phototransistor that has drifted out of its threshold window, not a mechanical jam. I have torn apart hundreds of these units on the bench, and the pattern is brutally consistent: dirty lenses, aged IR emitters, and factory potentiometers that nobody ever touched after the first year of service.
The core mechanism is simple and unforgiving. Every coin comparator relies on a pair of optical sensors—an emitter and a receiver—that watch the coin path. When a token passes, it interrupts the beam, the receiver voltage drops, and the logic board registers a valid coin. But when that beam is weak, misaligned, or the comparator threshold is set too tight, the board reads every passing coin as an intruder. Result: instant rejection, over and over, until the hopper starves and the machine throws a fault code. 🎯
Why Optical Drift Kills Acceptance Rates
Infrared emitters degrade with age. A diode that pushed 20mA of optical power at install will be lucky to hit half that after two years of continuous duty. The receiver does not care about your intentions—it only sees a weaker signal, and the comparator's analog front end interprets that as a coin that is too small or too thin. This is where coin comparator token rejection loops become a self-reinforcing nightmare: the operator cranks the sensitivity pot to force acceptance, the sensor saturates on legitimate coins, and now the unit rejects everything except slugs.
Dust and oil film are the silent accomplices. Cigarette smoke, grease from the coin path, and airborne grit coat the emitter and receiver lenses within weeks. A 0.2mm film of grime can cut optical transmission by 30 percent or more. You will never see it with the naked eye on a dark lens, which is exactly why so many techs waste hours swapping boards instead of cleaning optics.
Calibration Procedure That Actually Works
Stop guessing. Put a scope on the receiver output and watch the waveform as a known-good token passes. You want a clean square wave with a minimum 2.5V swing—anything less and the optical path is compromised. Clean both lenses with isopropyl and a lint-free swab, then re-measure before you touch a single potentiometer. If the swing is still weak, replace the emitter diode; they cost less than a coffee and solve half of all rejection complaints.
Once the optics are clean and the waveform is healthy, set the comparator threshold with the coin you actually want to accept, not a reference slug. Run ten consecutive tokens and log every accept and reject. A properly calibrated unit should hit 100 percent acceptance on genuine tokens and reject known slugs with zero false positives. If you still see intermittent rejects, check the sensor alignment bracket—vibration from the coin path loosens mounting screws and shifts the beam axis by fractions of a millimeter, which is enough to break the detection window. 🔧
The Potentiometer Trap
Here is the ugly truth most field techs refuse to admit: the sensitivity trim pot is not a tuning tool, it is a band-aid. Cranking it to mask a failing emitter guarantees that optical sensor calibration becomes a monthly ritual instead of a one-time fix. The correct move is to restore the optical path to spec and leave the pot at its factory midpoint. If the unit cannot accept genuine tokens at midpoint with clean optics and a fresh emitter, the receiver or the analog comparator IC is done—replace it and stop chasing ghosts.
Watch for thermal drift too. A comparator that works cold but rejects tokens after an hour of play has a receiver whose dark current climbs with temperature. Log the failure timing against ambient temperature and you will catch it every time. 💡
Signal Integrity and Grounding Faults
Weak grounding turns the receiver line into an antenna for every stepper motor and solenoid in the cabinet. A 50mV noise floor on the sensor output is enough to trigger false interruptions and send the board into a rejection loop that looks exactly like optical failure. Bond the comparator chassis to cabinet ground with a braided strap, verify less than 0.1 ohm resistance, and scope the receiver line with the hopper motor running. Clean signal, clean acceptance.
Finally, audit the token itself. Worn or bent tokens change the optical interruption profile and will trip a correctly calibrated sensor. Measure token thickness variance across a sample of fifty—anything beyond 0.15mm spread means you are calibrating against garbage and the loop will return.