Scrap & Yield Analyser

What this tool does

CM8-91 is a production yield register. You log one row per production run — how many units you started, how many came out good first time, how many were reworked to good, how many were scrapped, and what a unit of material cost. The tool works out first-pass yield, final yield, scrap rate and scrap cost for every run, then rolls them up honestly: by month, by product, by process, by shift and by scrap reason, so you can see not just that you are losing material but where and why.

Everything runs inside this single file. There is no account, no upload and no network request of any kind — product names, costs and shift performance never leave the computer you are using.

Logging a run

A run is whatever your operation treats as one batch through one process: a machining batch, a press run, a moulding shift, a paint rack sequence. Keep the definition consistent, because the register is only comparable run-to-run if a "run" means the same thing each time.

Give each run a date, a product, a process or line, and a shift. The process field drives the per-process summary and the "worst process" tile, so use the same wording every time — "Press line 1" and "Press 1" will be counted as two different processes. The main scrap reason is the single biggest cause of the scrap in that run; if a run had two distinct problems, it is usually better to log it as two rows, each with its own reason and counts.

The four quantities

Each unit started should end up in exactly one of three places:

• Good first time — passed inspection without any rework. This is the number that drives first-pass yield, and it is the strictest measure of how well the process ran.
• Reworked to good — needed extra work but ended up saleable. These units count towards final yield, but they cost you the rework labour and they hide process problems if you only ever look at the final figure.
• Scrapped — beyond economic recovery. These carry the material cost.

The tool warns if good + rework + scrap exceeds the quantity started, because a unit can only be counted once. The sum may legitimately be less than started — units still in the rework queue, held for inspection, or used for destructive testing — and the entry preview shows how many units are unaccounted for so you can decide whether that is deliberate.

First-pass and final yield

If a run has no started quantity, every yield shows a dash rather than a made-up number.

The gap between the two yields is the rework burden. A process with 88 % FPY and 98 % final yield ships almost everything, but one unit in ten is being touched twice — and rework is where cost, delay and escaped defects hide, because reworked units rarely get the same inspection as first-time-good ones. Watch FPY to improve the process; watch final yield to understand what actually shipped.

Why totals, not averaged percentages

Every aggregate yield in this tool — the headline tiles, the monthly trend, the shift chart and the per-process summary — is computed by summing the quantities first and dividing once:

It is never calculated by averaging the percentage of each run. The difference matters. Suppose one run of 1,000 units at 95 % and one run of 10 units at 50 %. Averaging the percentages gives 72.5 %, which describes nothing you actually made. Summing the quantities gives 955 good from 1,010 started — 94.6 % — which is what really happened to your material. Averaged percentages let tiny runs swing the figure; totals weight every unit equally. If you compare this tool's numbers against a spreadsheet, make sure the spreadsheet does the same, or the two will disagree and the spreadsheet will be the one that is wrong.

Scrap and rework cost

Enter the material cost of one unit at the point it was scrapped. A part scrapped after machining has had more value added than a raw casting, so the same product can carry different unit costs at different processes — that is correct, not an inconsistency. What this figure deliberately excludes is labour, machine time, energy and disposal. If you want a fuller loss figure, build those into the unit cost yourself and say so in the report notes.

The rework cost per unit is optional. Leave it at zero and the rework cost column shows a dash rather than pretending rework is free. If you set it, use an honest average — the loaded labour rate multiplied by the typical time to rework one unit is a reasonable start.

Finding the biggest problem

The scrap-cost-by-reason chart is ordered largest first — a Pareto view. In most operations a large share of the cost comes from one or two reasons, and the ranking tells you where the next improvement project should go. Cost, not unit count, is the right ranking: a hundred scrapped washers matter less than three scrapped machined housings.

Resist the pull of "operator error" as a default reason. If operators keep making the same error, the question is why the process allows it — fixturing that accepts a part the wrong way round, a missing first-off check, an unclear standard. The reason list includes setup, design and material precisely so that the honest cause gets recorded rather than the convenient one.

Reading the shift chart

The yield-by-shift chart computes each shift's FPY from that shift's summed quantities. A consistent gap between shifts is a finding, but rarely the finding people first assume: night and weekend shifts often run with thinner supervision, less engineering support, and machines that have been running longest since maintenance. Before concluding anything about the people, check what else is different about the hours they work — and check the started quantities shown under each bar, because a shift with one small run tells you almost nothing.

Setting a target

The FPY target on the Settings tab draws the line on the monthly trend and counts the runs below target on the headline tiles. Set it from your own history, not from folklore: a realistic target is one your best recent months have actually achieved. A target no run ever meets stops being information. The row-level FPY colouring uses the same target — green at or above it, amber within five points, red further below.

Printing and sharing

Print Report produces a report from whatever the current filter shows: headline figures, the four charts, the worst-runs and per-process tables, the full register and your closing notes. Print to PDF to circulate it. The scope line states the filter in force — clear the filters before issuing anything described as the full period. The date range, shift and reason filters apply to every screen, the report and the CSV export alike.

Saving your work

Runs, settings and the report header are written to this browser's local storage as you type, and the toolbar shows the time of the last save. That storage belongs to one browser on one computer: another browser, a private window, a second machine or a clean-up tool that clears site data will not have it.

Treat Export .json as the real save — one file containing everything, which Import .json restores anywhere. Export CSV gives you the register for spreadsheet work and includes every filtered record, not only those drawn on screen. Reset asks twice, then erases everything this tool has stored. There is no undo.

Accuracy & disclaimer

This tool records what you enter and calculates from it. It cannot tell whether units were counted correctly, whether the scrap reason recorded is the real one, or whether the unit cost reflects the value actually lost — and every figure it produces depends on all three. Scrap cost here is material cost only unless you choose to load the unit cost with more.

Yield conventions differ between industries — some count rework against yield differently, some measure at final test rather than per process. The formulas this tool applies are stated in full above so you can check they match your own definitions before comparing its output with anyone else's numbers. It is a record-keeping and calculation aid, not a substitute for measurement system checks or a quality management system.