Coin Comparator Token Rejection Loop: Why Your Acceptor Keeps Spitting Out Valid Coins
The coin comparator token rejection loop is not a calibration quirk — it is a hard electrical fault state where the acceptor's sample coil comparison circuit never confirms a valid token, so the gate solenoid fires reject on every single insertion. I have torn down enough chassis to tell you straight: ninety percent of operators who "fix" this by cranking the sensitivity pot are just masking a signal integrity failure that will burn out the comparator board within weeks.
Here is the brutal truth about how these units actually sample. A coin comparator drives an oscillating field through two coils — a reference coil wound against a fixed sample slug and a test coil at the entry throat. When a token drops through, the eddy currents it induces shift the test coil's impedance, and the differential bridge circuit compares that shift against the stored signature. In a coin comparator token rejection loop, that differential output never crosses the threshold window. The relay stays in reject position, the solenoid kicks the token back to the return chute, and the cycle repeats indefinitely. 😤
The Three Physical Culprits Behind Endless Rejection
First culprit: contaminated sample slug. That reference slug is machined to tolerances measured in microns, and a film of oil, nickel dust, or moisture-induced oxidation changes its magnetic permeability just enough to shift the entire comparison baseline. The comparator then rejects genuine tokens because the reference signature has drifted — not the token. Second culprit: coil winding degradation from thermal cycling. Every accept/reject cycle pushes current through the drive coil, and cheap enamel insulation develops micro-cracks that alter inductance. Third culprit: the optical or inductive entry sensor misfiring, which resets the sample-and-hold window mid-comparison and guarantees a reject pulse.
Let me be blunt about the diagnostic order. Do not touch the trim pot first. Pull the comparator, mount it on a bench, and feed it a known-good token with a scope on the bridge output. If the differential waveform is flat or clipped, you have a coil or slug problem. If it swings but never latches, you have a threshold comparator or hysteresis fault on the board itself. 🛠️
Why the Rejection Loop Corrupts Your Audit Trail
Here is the part that costs operators real money. Every entry into the coin comparator token rejection loop generates a reject event, and most modern acceptor boards log these pulses to the metering header. A stuck loop floods your audit data with hundreds of phantom rejects, which destroys your drop reconciliation and makes genuine short-tender disputes impossible to settle. I have seen venues write off entire shifts because the comparator was looping and nobody read the event counter.
The mechanical toll is worse. Each reject cycle slams the gate solenoid against its stop, and the return chute lip takes the token impact at full velocity. Run a loop for a week and you will find hairline fractures at the chute weld and a solenoid plunger worn oval. 💀
Field Fix Protocol That Actually Holds
Swap the sample slug first — it is a five-minute part change and eliminates the most common baseline drift. Then bench-test the bridge waveform with a scope; if the differential amplitude is under 40 millivolts peak-to-peak on a known token, the drive coil is suspect and the whole comparator head needs replacement, not adjustment. Verify the entry sensor output is a clean square wave with no chatter, because a bouncing sensor edge will re-trigger the sample window and lock the unit into permanent reject.
After any slug or coil swap, re-teach the token signature against three sample tokens from different production batches. If the accept window only catches one batch, your comparator's hysteresis band is too narrow and you are one temperature swing away from the loop returning. The threshold window should hold at least plus or minus eight percent of nominal signature shift across all three samples. That is the number to chase.
Board-Level Failure Modes You Cannot Pot-Adjust Away
The comparator IC itself fails in a predictable pattern: the internal reference voltage divider drifts low, which biases the entire comparison window downward and makes every token look "too large." No amount of sensitivity trimming recovers this because the trim pot sits downstream of the drifted divider. Replace the comparator IC and re-bias. Also check the electrolytic capacitor feeding the coil driver — ESR rise from heat exposure causes the drive amplitude to sag during the sample window, and a sagging drive reads as an oversized token, triggering the coin comparator token rejection loop on every legitimate insert.
Measure drive amplitude at the coil terminals under load, not open-circuit. If it drops more than fifteen percent during the sample pulse, the cap is done. Replace it with a low-ESR part rated for the thermal environment, and re-run the three-token acceptance test before the unit goes back in service.