How to Accept a Screw Removal Machine: Twelve Tests and Their Pass Criteria Across FAT and SAT

The step most often skipped on delivery day is acceptance. Many factories power the machine on, place one workpiece, watch it move, sign, and move into volume production. Three weeks later they find takt time is twenty percent short, torque curves do not match, and one screw size keeps slipping, only to discover that no traceable pass criteria were ever recorded. Acceptance is not a formality. It decides whether the next three years of conversations with a supplier are based on data or on impressions.

Two Stages: FAT for Capability, SAT for Site Conditions

FAT happens at the supplier's plant and answers one question: can this machine reach its rated capability with standard parts in a standard environment. SAT happens at the customer site and answers a different one: with your workpieces, your air pressure, your power grid and your operators, how much of that capability is left. The two stages cannot share test items, because merging them makes responsibility impossible to assign. When the site result falls short, the supplier can point at unstable air supply and the customer can point at a weak machine, and neither argument can be settled.

The workable arrangement is that FAT tests only machine capability with the supplier's own reference workpiece, and SAT tests only integration and environmental fit with the customer's real workpieces. Ship after FAT passes, sign final acceptance after SAT passes. If the project cannot afford two stages, at minimum the contract should state which metrics are the machine's responsibility and which belong to site conditions.

Six Tests That Belong in FAT

The first is positional repeatability. Remove twenty consecutive screws at the same hole location using a reference workpiece, and record the deviation of the bit center each time. Judge on the maximum, not the mean, and for the M0.8 to M2.0 range the usual requirement is within plus or minus 0.02 mm. The second is torque and axial force control accuracy. Measure ten cycles each at twenty, fifty and eighty percent of full scale, and look at both deviation and repeatability between setpoint and actual value. The third is angular resolution. Rotate the bit through the angle corresponding to one thread pitch and check whether the system can distinguish a thread that has released from one still engaged. This is directly tied to the breakage rate.

The fourth is takt time, and the definition has to be stated. Pure removal time per screw, full cycle time including load and unload, and first-article time after a changeover are three very different numbers. A supplier who quotes only one number is usually quoting the best one. The fifth is pick-up and collection success. Run two hundred consecutive removals and count how many screws are carried away without being dropped onto the workpiece, since a single drop can scratch a product. The sixth is exception handling. Deliberately create a jam, a stripped head and a misloaded workpiece, and observe whether the machine alarms and stops or forces its way through. That behavior decides whether a workpiece gets scrapped.

Six Tests That Belong in SAT

The seventh is stability under site air and power variation. Vary inlet pressure across the specified range and see whether axial force drifts, and by how much. The eighth is fit rate on real workpieces. Run fifty pieces each on three representative customer workpieces and count the share that requires manual intervention. The ninth is changeover time, measured with a stopwatch from workpiece A to workpiece B, including fixture swap, program recall and first-article confirmation. Many projects exceed their target here.

The tenth is first-pass rate of vision alignment, tested separately on vision-equipped models. Run fifty attempts each under normal lighting, a reflective part and a contaminated part, and count how often recognition fails and requires a retake. The eleventh is the line interface, covering I/O handshake timing, MES data fields and completeness of alarm reporting. This item is the most often omitted and the most likely to force rework after go-live. The twelfth is safety and guarding: emergency stop response, door interlock and residual voltage, checked item by item against the site safety standard.

Three Criteria That Cause the Most Disputes

The first is the definition of takt time. If the contract says only that cycle time is under five seconds, the two sides will almost certainly disagree at acceptance about when those five seconds start and end. The correct wording names the start and end events, for example from the workpiece-in-place signal to the machine's completion output. The second is the definition of yield. Is it the intact rate of removed screws, the no-damage rate of the workpiece, or the no-manual-intervention rate. The three numbers differ widely and point to different improvements, so the contract must say which one applies.

The third is the definition of accuracy. Repeatability, absolute accuracy and following error are three different things. A supplier may quote repeatability while the customer measures absolute accuracy at acceptance, and the results cannot match. Pass criteria belong in writing at contract stage, together with the measurement method and the instruments used, ideally including how many measurements are taken and which value is reported.

What the Acceptance Record Must Contain

Keep at least four categories. First, raw measurement data with time, conditions and actual values for every measurement, not just a summarized average. Second, torque and axial force curve files, which are the only objective basis for judging the process window later. Third, exception handling records: how the machine reacted to each deliberately induced fault, captured as screenshots or video. Fourth, the signed judgment table, with each item marked pass, conditional pass or fail, and every conditional pass stating the corrective action and re-verification date.

Archive those four and you have a basis for discussion when the machine misbehaves later. Honred Technology ships its desktop and vertical small screw removal equipment with FAT and SAT judgment templates that spell out the measurement method for positional repeatability, force control accuracy, pick-up success rate and changeover time, so the customer can re-measure on the same basis. For a buyer, the value of an acceptance table is not the moment of signature. It is three years later, when machine behavior starts to drift and you still know where it started from.

If the project includes vision alignment, give item ten its own annex: retake counts under different reflective conditions, a breakdown of recognition failure causes, and the cleaning interval for lens and light. Those three determine whether the line needs a dedicated operator after go-live. The hard part of accepting an automatic vision screw removal machine is never the recognition rate itself but its stability as workpiece batches change, and that can only be tested across multiple batches. A criterion that cannot be tested for stability should not receive full marks on the acceptance table.

Once these twelve criteria are fixed as a template, they can be used in reverse at the selection stage. Send the six FAT items to candidate suppliers before quotation and ask each to state measured values up front, and the bids immediately become comparable. A complete M0.8-M2.0 screw removal solution should carry those metric commitments during evaluation rather than discussing whether they can be added after the machine is built.