Control of Variable Stock Allowance Before the First Finish Pass: A Route

Where variable stock allowance appears

Variable stock allowance appears when the blank axis, the setup axis, and the axis of the future finish surface do not match. The machine follows the path exactly, but the metal around it is distributed unevenly. As a result, the tool removes too much on one side and leaves an extra layer on the other.

Most often, the problem starts before cutting even begins. The blank may arrive with runout on the outer diameter, ovality, a skewed face, or an off-center hole. By eye, the part often looks fine, but after the first roughing pass you can see that the allowance is not a ring anymore, but a “crescent.”

The error is quickly made worse by poor locating. If the operator clamps the part on a surface that has already shifted, the finish axis moves along with that error. Then the allowance ends too early on one side, while on the other the tool barely reaches the required surface. The size may still be controlled, but the roundness, runout, or wall thickness is already off.

There are usually three sources: the blank itself, the setup, and the cutting conditions. Problems come from crooked rolled stock, castings with shift, an angled cut after parting, a dirty datum, a tilt in the jaws, weak support, too much stock removal per pass, tool wear, deflection on long parts, and vibration during roughing.

After roughing, there are signs you should not miss. On one side, a scale layer or dark unmachined band remains, while the metal opposite it is already clean. The tool mark around the circumference has different widths. A shaft shows noticeable runout on the rough journal, and a bushing shows wall thickness changes by sector.

A simple example: a shaft blank is clamped by its outside diameter, which is shifted by 0.6 mm relative to the axis. Roughing goes smoothly, but the allowance gets redistributed. On the first finish pass, one side is already nearly to size, while the other is still “raw.” If you notice this only at the end, the part often goes to scrap or needs an extra pass.

That is why variable stock allowance is best checked right after roughing. And you should look not only at the overall size, but at the symmetry of the metal relative to the future datum.

What to check before the first finish pass

The first finish pass should not be hunting for the problem blind. If the allowance is already wandering, the finish pass will only reveal the defect when it is too late to fix. The check needs to happen earlier, while the part can still be saved by adjusting the setup, re-clamping, or changing the datum.

First, check whether the axis has shifted relative to the datum. This is a common situation on CNC lathes: the part looks clamped correctly, but the actual axis has already moved. Then the allowance is thick on one side and nearly gone on the other.

Next, compare the allowance on the two sides or in two sections. The total metal stock may be enough, but if it is distributed unevenly, the tool works under different conditions within one pass. That creates load spikes, a different tool mark, and the risk that one area will clean up while another will not.

Also check runout on the datum surface. If the datum runs out, the whole downstream geometry is in question. In that case, linear dimensions alone do not tell you much. The part may be in size in one section and out in another.

Before the finish pass, it is also useful to check the shape in the area with the tightest requirements: the fit, the journal, or the face for mating. Here you look not only at size, but also at ovality or taper. Even a small skew during roughing later turns into extra stock removal or a shortage on one side.

Usually, it is enough to catch five signals: the axis does not match the selected datum, the left and right allowances are noticeably different, the datum surface shows runout, the control section already has ovality or taper, and the result changes after re-clamping the same part.

The last point is often underestimated. If measurements start to drift after re-clamping, the process has lost repeatability. The usual cause is a weak locating scheme, a dirty support, an incorrect clamp, or an unstable blank. That is more dangerous than a one-off deviation, because scrap quickly appears in batches.

A good pre-finish check does not collect dozens of numbers. It answers one simple question: is there enough allowance in the right place, and will the geometry hold after the next setup?

Which dimensions are actually needed

When a job collects ten measurements at a time, the shop gets buried in numbers and still misses the fault. For variable stock allowance control, 3–5 dimensions per operation are usually enough. More is rarely needed.

Keep only the points that are tied to the first finish pass. If the finish pass is based on face A and outside diameter B, you need to measure the relationship between the allowance and those same datums. Dimensions taken from a random rough surface often look useful, but nobody makes decisions from them.

Usually, this set is enough: one measurement shows the blank position relative to the datum, one shows the allowance itself in the future finish area, another catches a shift relative to the second datum, plus one shape check. If there is a risk of drift, add a check at the far end.

A shape check is almost always more useful than one extra linear dimension. If you only look at diameter or length numbers, it is easy to miss a skew. One indicator reading for runout often tells you more than three extra dimensions.

Duplicates should be removed right away. If two measurements answer the same question, keep the one that is quicker to take and easier to understand. A simple rule is: if the result does not change the action of the operator, setter, or inspector, that measurement is unnecessary.

On the operation sheet, it is useful to note not only what to measure, but who does it. After setup, the operator checks the base dimension with a caliper or indicator. On the first part, the setter checks shape with a micrometer and indicator. The inspector confirms only the points used to decide whether to start the batch or correct the offset.

A normal route is simple: 3–5 control points, the same datums that will be used in finish machining, and one shape measurement. That is usually enough to spot the problem before the next operation.

Measurement route step by step

On the shop floor, the best solution is a short five-step route. It catches stock shift early, while the part can still be saved by roughing correction, and the operator does not waste time on extra records.

1. Check the blank before setup. Look at the outside size, obvious ovality, impact marks, and curvature. If the rolled stock or forging already has a large spread, further measurements will only hide the source of the problem.

2. Take the first measurement right after locating. The datum and one working diameter are enough. You only need to understand one thing here: did the part sit true, or did the shift appear right from the start?

3. After roughing, check the remaining stock in two sections. Usually, a point near the chuck and a point at the far end are enough. This check quickly shows skew: one place still has stock, while in the other it has almost disappeared.

4. Check runout on the datum surface. Measure the same datum from which the first finish pass will be taken. Otherwise, you may get a good diameter in one spot and a coaxiality problem on the finished part.

5. If the difference exceeds the limit, stop the batch. Do not wait for the next part to somehow become straight. First look for the cause: crooked blank, locating error, worn jaws, tool drift, or a program issue.

This route does not overload the shop. In practice, the operator only records four things: what the blank looked like before setup, how the part sat on the datum, how much stock remained in two sections, and how much runout the datum produces.

On CNC lathes, this sequence is especially useful for long parts and batches with unstable blanks. If you move the steps around or skip the first measurement after locating, the problem is usually detected too late.

How to set reaction limits

A reaction limit is not for reporting. It is there so the shop does not wait for scrap. A part may still pass roughing size, but that does not mean the process is under control. With variable stock allowance, the fault often shows up before the finish tolerance is lost.

First, separate normal variation from process drift. After setup, take a series of 10–15 parts under steady conditions and see how the measurements move. If the size stays within a narrow band and does not creep from part to part, that is the normal background. If each next reading moves in one direction, the process is already drifting.

Two limits for two different problems

Do not use one common limit for everything. Datum shift and allowance change are different things.

For the datum, the limit is usually set tighter. Even a small shift quickly eats into the stock before the first finish pass. For the allowance itself, the limit can be a bit wider if the spread is not growing across the batch.

In practice, three zones work well. Green means the process is within the normal band and no action is needed. Yellow means the size is still safe, but it has moved beyond the usual background, so the next part is measured again. Red means there is a clear datum shift or the remaining allowance is approaching the minimum, and the setter stops the process and looks for the cause.

Do not wait for the finish tolerance to be fully exceeded. It is better to react when the allowance is already visibly shrinking. If one side has lost so much metal that finish machining can no longer true the part without risk, stop the process at once.

When every part should be measured

Full inspection is not always needed. It is used after changeover, tool replacement, datum correction, the start of a new batch, and after any adjustment that already changed the measurement pattern. Once a series has stabilized and 5–7 parts in a row hold the same range, you can move to spot checks.

It is better to write the frequency directly into the route: the first 5 parts are measured every time, then every 3rd or 5th part until the process shows a new shift.

One more practical point: everyone on the shop floor should know who has the authority to press stop. If that is not written down, the operator sees the risk, the inspector hesitates, and the machine keeps running. It is better to state it clearly: the operator stops at the red zone, the inspector confirms it, and the setter looks for the cause and decides on restart.

Example: a shaft with uneven allowance

On a shaft after forging, the allowance along the length often varies more than a single size suggests. On the left, the metal may still have 2.5 mm per side, while on the right there may be only 0.8 mm. If such a part goes to the first finish pass without checking, the tool will cut normally at one end and quickly run out of stock at the other.

A typical case looks like this: a shaft about 600 mm long arrives after forging with a slight bend and an uneven outer surface. On the first setup, the operator locates the part on the rough face and the journal after roughing. If the face is not cleaned up evenly or the journal sits at a skew, the part axis moves away from the datum by a few tenths. That is already enough to make the allowance uneven from end to end.

You do not need a long report here. A short route is enough: measure the remaining allowance at the left end, then at the right end in the same section, check runout of the outer surface relative to the current datum, and see where the minimum allowance is closest to dropping below the finish limit.

If the difference between the ends exceeds, for example, 0.6–0.8 mm per side, it is better not to send the part straight to finish machining. Two end checks catch the skew before it becomes scrap. A single measurement in the middle often does not show this defect, because the stock there may still look fine.

Runout checking helps explain the cause, not just the symptom. If runout grows on the outer surface while the datum remains stable, the source is often the forging itself. If runout appears from the locating journal or face, look at the setup: clamping force, support, datum cleanliness, and the clamping scheme.

The fix is often simple. Adjust the locating, add a mandatory face check before setup, remove clamping on a doubtful rough surface, and the variation from end to end becomes repeatable. After that, the operator makes two measurements, checks runout, and immediately knows whether the shaft can go to the first finish pass or needs correction first.

Where people most often go wrong

The most common mistake is taking many dimensions without looking at shape. For variable stock allowance, it is not enough to know that the diameter after roughing is “about right.” You need to see where the metal remains and where it is almost gone.

People are often reassured by an average number. They measure the part in two or three places, calculate the average size, and decide that the pre-finish check passed. But variable stock allowance is not caught that way. It is much more useful to compare sides and sections: top and bottom, the area near the chuck and the far end, one section and the next. The difference between points says more than a neat number in one box.

Another mistake is blaming the tool when the error is in the setup. The cutting tool is often accused first because it is convenient. But if the blank sits skewed, the jaws pull unevenly, or the datum was chosen poorly, the tool is only repeating someone else’s error.

You can usually see this from the pattern of the deviation. If the size does not wander randomly but shifts in a clear direction or along the length of the part, the cause is more often clamping, runout, or locating than tool wear.

The problem gets worse when the result is recorded too late. The part has already moved on, the next batch is in the machine, and the measurements are entered into the log an hour later or at the end of the shift. Then the check loses most of its value. If a number does not change a decision immediately, it is reporting, not process control.

Confusion also appears when different defect sources are mixed together: the blank arrived shifted, the setter chose a weak setup scheme, the operator seated it on the wrong datum after re-clamping, or the inspection took too few points and smoothed the problem out with an average. The result looks similar, but the causes are different. If the defect repeats across different setups on one batch of blanks, look at the incoming material. If the deviation appears after a specific clamp or re-clamp, check the setup and locating.

Worst of all, by the first finish pass the allowance is already uneven. After that, there is almost no time left for a calm analysis.

A short check before the first finish pass

Before the first finish pass, the shop usually loses time not on machining itself, but on confusion in inspection. If the operator does not have the same datum, the same points, and a clear reaction limit, variable stock allowance is noticed too late.

A short check works better than a long form. It is enough to confirm a few things. The part is placed on a clean and stable datum. Measurements are taken not “somewhere around here,” but at predefined points. The instrument is chosen for the real tolerance, not because “it happened to be nearby.” The stop line is known to the whole shift in advance. The first parts of a new batch or after changeover are checked more closely than the rest.

In practice, it is simple. The operator takes the first part, cleans the datum, sets it in the same way every time, takes measurements only at marked points, and immediately compares them with the reaction limit. If the second and third parts show the same pattern, the route can move into the normal rhythm.

There is also the opposite mistake: the shop honestly measures many dimensions, but none of them answers the main question — is there enough allowance before the first finish pass on the required surface? Extra numbers create the illusion of control. Only the measurement route that quickly shows whether the part can continue or must stop now is useful.

If that routine is posted by the machine and understood the same way by the operator, setter, and inspector, the first finish pass becomes calmer and has fewer surprises.

What to do after the route is launched

After launch, do not try to collect every number at once. For the first batches, a simple form is enough, one that shows not only the measurement but also the decision. Usually, it is enough to record the part number, machine, datum, 2–4 control dimensions, deviation, and the action taken after the check.

That sheet quickly shows where variable stock allowance control is helping and where the shop is just wasting time. If the size drifted but the decision never changed, the measurement itself should be reviewed.

After a few batches, it is worth reviewing the route. Keep the measurements after which people actually changed the setup, locating, or stock allowance on the operation. Remove the points that never affected a decision. Mark the places where the shift repeats batch after batch, and see on which machines and with which datums it happens most often.

If the drift always appears on the same side, do not try to fix it with a single size correction. First check the locating. A repeatable shift almost always leaves a clear trace.

After each found scrap case, it is best to update the control plan right away while the cause is still fresh. Long explanations are not needed. A short note is enough: what was caught, where the cause was, which measurement was added or removed, and which reaction limit was changed.

When a new CNC lathe is introduced on the shop floor, the measurement route should be checked not only against the process plan, but also against the real setup. For equipment launched with EAST CNC, it makes sense to discuss this already during commissioning and service: which datums should be treated as working datums, where a typical shift may appear, and which checks are really needed before the first finish pass.

A good route often becomes shorter after a month. That is a normal result if it catches the problem earlier and gets in the way of work less often.