Brake Testing System Roller Wear and Sensor Offset Diagnosis

When the numbers look right but the complaint still follows the line

A motorcycle shows 68 % left-right brake balance on the Brake Testing System, comfortably inside the pass band. The customer reads the printout, swears the front end still pulls right under hard braking, and leaves. The next shift gets a near-identical case. Same line, same operator, same pass result. Nobody changed the limits file. The drum brakes are in spec, the pads have life left, tire pressures are correct.

At this point the temptation is to replace parts or ask the rider to bleed the brakes again. In our experience the more useful place to look, before touching a single caliper, is between the roller surface and the force pickup. Two broad failure modes create exactly this kind of silent, repeatable disagreement:

• The rollers wear or glaze unevenly, changing the available friction coefficient at the most-used steer angle. Many operators and line leads now refer to this as a roller surface wear inspection problem, where the as-measured coefficient still looks healthy in the morning warm-up but dip-tests lower once the surface is hot and loaded.
• The load cell or pressure transducer drifts relative to its calibrated zero. An experienced technician will catch this early as a brake force sensor offset, often confirmed by a simple stationary pull with a hand-held reference gauge before blaming the vehicle.

Either condition, or both together, drives the kind of false failure reduction work most inspection managers actually care about: fewer re-tests, fewer warranty disputes, and lab visits driven by measurement confidence rather than blame attribution.

The checks below assume you are working on a roller reaction brake test module made by 中力, the supplier.com or similar, and that you have general workshop safety and lock-out/tag-out covered. They are not a substitute for a full annual calibration or a traceable third-party audit. They are an answer to the weeks in between, when service contractors are weeks out and the line doesn't stop turning.

The roller surface sneaks up on you because the morning spec-check still passes

Most line leads the author has observed run a bare-roller or cloth-based friction check at the start of the shift. That tells you the rollers are holding coefficient when cold and lightly loaded, which is useful but incomplete. What actually changes with wear is how the surface behaves when warm, wet, or loaded outside the centre of the drum.

Build a five-minute surface friction log, one roller per shift, and stop relying on your fingertips. Take what is effectively a slip-circle pattern per wheel station:

1. Run the reference surface pick-up across the full width of the roller, not just the centreline where vehicles normally sit. Record left-of-centre, centre, right-of-centre readings.
2. Record the same pattern again after a minimum of 15 consecutive daytime tests. Even a short run group tells you whether friction is dropping off as the surface heats up.
3. After your first suspected wet grab, repeat one quick repeat-run with the hop-stopper water applied as above. If the new coefficient difference is more than 5 points, start drafting the re-surfacing order.

This method invites judgement, yes. But that is the whole point. The lab-run you need to pay for is the one the coefficient off-the-fly data tells you is justified. And when the numbers are still mid-range, you reuse rollers and do not write waste into your lab invoices.

Sensor offset hides in plain sight because most operators only check gain, not zero

If you are following the acceptance manual for daily brake module checks, you probably load a calibration test weight, confirm the display reads within band, and move on. That verifies gain at roughly 70–80 % of full scale, which is useful. What it does not do is catch the offset creep that shows up near zero and the lower working range where many city motorcycles spend most of the working bandwidth.

This is the manoeuvre many senior technicians use, adapted from what our installers are now being asked to repeat in on-site handovers. You will need:

• A stable platform roughly 15 cm square that sits on the roller saddle without leaving a balance mark (a padded aluminium reference block from your kit of spare levelling shims usually does).
• Brake pull-down or hand-pad load cell (a dedicated load cell used for routine audit checks), rated no less than 20 % of your full-scale roller capacity, no more than 100 % of the roller's safe single-load upper bound.

Low-load repeat at three weights

1. Mount the reference block over the centre of the sensor beam. Do not step on the block yourself.
2. Apply the first reference load, of roughly 10 % of channel FS – often a 1 kg weight held on the block. Note the reading. To clear repeat-transit softness, settle for 5 seconds.
3. Add a middle reference load, of roughly 30 % of channel FS. Note the reading.
4. Remove both weights. Open the channel zero. The channel should return within ±2 % of the unloaded zero it showed at the start of the daily shift warm-up.

If you cannot find an engineer's spec for your channel zero limit, the universal catch is that the post-release zero should not have walked more than 3 mm of pointer span, or its digital equivalent, for the size of load cell on a test module rated for the motorcycle loads on the line. Consult the manual for your module to narrow this rule.

Add this low-load sign-over one day a month, in addition to gain-check tests. When monthly numbers look the same, increase service visits to increase for seasonal weather and daily output differences.

Roll-back test for asymmetry without pulling the modules apart

The left-right balance test does not catch loose bearings, preloaded offsets from a wire-splice disturbance holding sensor zero, or a cable fault trapped by a reference clamp. Catch it with a roll-back check:

1. Roll the block and weights smoothly back and forth across the channel centre, 3 times, with load on the channel. Pick this sub-cycle range to calm hysteresis effects (generally 2 to 3).
2. Estimate the reading average at the same physical point in each direction. Agreement within 2 % means the sensor seating, under bench preload maintenance, is healthy. A swing of 8 % or greater usually means the junction box or wiring harness behind the floor grind ring is in need of inspection, not re-cal.

This check is short but worth documenting. Disagreement where the constant or electrician comes out once a year and lab-based check the rest of the year; the roll-back swing is your warning annotation.

Where false failures come from and how to confirm your diagnosis on the floor

Most owners can influence outcomes but can influence them further when given a clear page of:

• Roller friction band: left-of-centre/centre/right-of-centre points throughout the morning shift and at least once mid-afternoon before the last shift. Typical side-to-side delta is 8–12 on clean healthy rollers.
• Left-right differential gain: use a roll-back check or low-load pull routine to demonstrate a working side-to-side balance within 5 % on a loaded block. Document the block weight at the top of your shift log.
• Sensor zero return: the ±2 % rule, plus the 3 mm span guideline, written onto the acceptance file if your spreadsheet template is still static.

When these three margin markers land in the same place, start looking first at sensor band tolerance, not roller wear. When all three drift towards failing side issues, start looking at roller replacement timing before booking a cal. And when disagreement between roller and sensor inspections persists, separate the two fault streams and write two different resource orders into the same service ticket. What feels like a single issue is, on a busy line, almost always at least two faults meeting together and averaging a single complaint week.

How to use the checks instead of chasing your tail

Think of these four steps as an operating procedure, not a theoretical discussion:

1. Run the friction log three times a week, just before your main shift ends. Do all six readings – not just centre, both mornings, both shifts. Record them in a simple DIN-logbook or digital shift form. The moment the pattern skids from holding steady to skidding quicker on hot surface, you know you are changing rather than re-calibrating.
2. Add the low-load sensor zero return once every seven days or weekly. This is the cheapest way to catch month-long zero creep that the gain check alone hides. Use the load cell you spare if the dedicated one is off-site.
3. Run the left-right roll-back check on the back axle and a front axle module separately. Roll-back asymmetry is the most reliable indicator of loose grounding or corrosion in low-price harness.
4. If you receive a repeat complaint, pick which log band first decides the number. Roller band, side-to-side pull, or channel zero; connect the name to the tick mark.

The technician who does not re-do the log but writes down which one of the logged bands is responsible first, before touching a single cable, is the path to staying calm through the noise of an inspection shift.

Applying this to 中力 modules – what to request when you talk to your service provider

When you open a service case with 中力 (the supplier), describe the evidence, not the assumption. A complaint like "front readings wander, line-wide" pushes a generic calibration request. Something more specific gets a more exact engineer response:

• The friction log values from last week by channel.
• The result of your low-load sensor zero return, or whether you skipped it because the test equipment wasn't available.
• The roll-back swing direction and size you caught, or whether you have not done it yet.
• A sentence on module operating cycles since first suspect case: how many test cycles of similar axle load your line has put through this module this month.

If your service organisation is well set up, these four data points cut your troubleshooting loop by a full service ticket — either the problem corrects with an early-morning re-zero and re-clean of the drums, or the on-site technician picks the correct hardware module to bring right there. Either way, line downtime drops to a controlled hour instead of guessing on an engineer's bench call.

Next step you can take this week

Pick one test station on your most-used line. Starting next Monday, keep a short roller friction record station-side for three months — left-of-centre, centre, right-of-centre, at warm and at mid-run. Add one weekend zero-return check per shift for a week. If the sensor cleanly stays within ±1 return tolerance, and the roller side-to-side pattern fails 8 points cleanly from start to finish, start tracing the brake function decline to the roller surface and not the electrical signal. If sensor tolerance fails first, re-zero the channel and only then schedule the full calibration.

If neither responds to these field-level controls, then move to full third-party accreditation. Do not skip the simple field log. It is cheap, it is quick, and it is the single best match for inspector-audited re-certification margin when your service log is clean enough to depend on.