Condition-Based Tool Change: Where It Beats a Counter

Why a counter does not always hit the right moment

A parts counter is convenient because it gives a simple rule: once a set number of parts has been processed, change the tool. The problem is that a counter only sees quantity. It does not see how the cut actually went or the conditions under which the cutting edge worked.

On a CNC lathe, this becomes obvious very quickly. Today a batch runs smoothly: the stock allowance is consistent, scale barely causes any trouble, the load repeats from part to part, and the insert uses up its normal life without issues. Tomorrow a workpiece arrives to the same drawing, but the actual surface condition is different. The counter shows the same number, but the tool is now living through a different scenario.

Even two workpieces from the same batch can cut differently. One has a slightly larger stock allowance. The other has heavier scale or a harder surface layer. From the outside, the difference seems small, but for the cutting edge it is very real. Load rises, temperature rises too, and wear progresses faster than the counter suggests.

What usually affects the changeover point most is varying stock allowance along length or diameter, casting skin, scale, forging marks, different hardness within the same batch, and interrupted cutting with impacts on entry. Because of this, the “tool change by count” approach often falls into one of two extremes. Either the tool is changed too early and part of its life is wasted, or it is kept until the planned number of parts is reached and problems show up on the finished part.

Late replacement almost always hits two areas first. The size drifts. The operator starts adjusting compensation more often, but that only helps for a short time. Then the surface quality drops: roughness increases, rubbing marks appear, burrs show up, or vibration marks begin. If the part is precise, scrap costs more than the insert.

That is why condition-based tool change works better where the workpiece behaves inconsistently. The counter is useful as a reference, but it rarely works as the only signal. If the material, stock allowance, and scale vary, the number of parts stops being a fair measure of tool life.

In practice, the difference looks simple. The same insert on a stable workpiece may last, for example, 120 parts, while on a variable one it starts ruining size after 80–90. The counter does not notice that. The part size and surface mark notice it immediately.

When a counter works well

A counter performs well where the process changes very little from part to part. If the batch is consistent, the material is the same, and the stock allowance stays within a narrow range, the tool wears in a fairly predictable way. In that situation, tool change by count gives a calm result without unnecessary stops.

This approach works best on mass-produced parts. If a shop turns a large batch of identical shafts from one grade of steel for construction equipment, the difference between the first and the hundredth workpiece is usually small. The load on the insert repeats, and the life can be estimated not “by eye” but from previous runs.

Usually, a counter does not let you down when the material hardness does not vary from batch to batch, the stock allowance stays the same, cutting parameters are not changed in the middle of the run, and the operator does not mix different part types into one tool life cycle.

It is also important to separate roughing and finishing. This is a common mistake: one limit is set for the whole tool station, even though the roughing pass consumes life much faster. For roughing, the counter is usually shorter and more coarse. For finishing, it is longer, but with tighter control of size and surface quality.

Even in a stable run, the counter cannot be left unchecked. If the part size slowly drifts toward the upper or lower tolerance limit near the end of tool life, the limit is already too high. In that case, the life should be shortened, even if the insert still looks “alive.” Otherwise, savings on one insert can quickly turn into scrap on several parts.

In a steady process, condition-based tool change is not always needed as the main method. More often, it works as a safeguard: it helps check whether the planned life has shifted after a new material batch, a different workpiece, or a change in parameters. If the process is stable, the counter remains a convenient baseline. But that baseline only works when the process itself is stable.

Where it is better to watch the actual condition

It is better to look at the actual tool condition where the workpiece does not hold the same level from part to part. If bar stock, forgings, or castings vary in hardness, stock allowance, and surface condition, the same counter starts making mistakes. Today the insert may comfortably reach 120 parts; tomorrow it may start failing after 70.

This approach is especially justified with variable stock. In a stable run, where material, stock allowance, and parameters barely change, tool change by count gives a predictable result. But once the batch arrives with scale, casting skin, or marks from rough pre-machining, a fixed limit is no longer reliable.

Scale and casting skin quickly wear down the cutting edge at the start of the pass. Interrupted cutting adds impact load, and the insert can lose its edge before the counter reaches the target. In such a situation, it is more useful to watch not the number of parts, but how the size holds, whether the cutting sound changes, and whether the surface quality gets worse.

The same happens in runs with varying hardness. On paper, two batches look identical, but one cuts softly while the other immediately causes extra heat and faster wear. If the shop keeps changing the tool strictly by counter, it either throws away still-good inserts or gets scrap because of size drift.

In practice, the replacement moment often has to be adjusted from batch to batch. On a soft batch, the tool may hold 130 parts without size drift. On the next one, after 90 parts the operator already sees roughness increase and begins to adjust compensation. That means the limit for that batch should be lowered instead of waiting for the previous number.

This approach is especially useful on CNC lathes when production works with different metal suppliers, forgings, and castings rather than one uniform series. In these conditions, the real tool condition is a more accurate guide than an averaged counter. Tool life here cannot be calculated once and forgotten.

What counts as signs of wear

Wear rarely shows up as one obvious failure. Usually, the tool gives small signals first, and the operator sees them before scrap or chipping appears.

The first signal is the part size drifting toward one of the tolerance limits. If the batch runs at the same settings and diameter or length begins to slowly move up or down, the reason is often not the program, but the cutting edge. One random deviation does not mean much. But if the shift repeats on several parts in a row, the tool is already losing its previous geometry.

The second sign is the surface. The cutting parameters were not changed, the workpiece is the same, but the roughness got worse. Scratches, dull bands, fine waviness, or the loss of shine in the wrong place appear. This often happens before size goes out of tolerance, so wear can be easy to miss if you check size alone.

Watching the chips also helps. When the tool is working steadily, chips usually behave in a predictable way. If their shape changes sharply, they become longer, break differently than before, darken, or turn bluer, the cutting zone is working differently. That does not always mean the tool must be changed immediately, but it is worth checking.

On CNC machines, many people miss another clear signal: rising spindle load. If feed and depth are the same but the load is higher, the edge is cutting harder. On the screen, this shows up before the problem becomes audible.

There are also signs that are noticed almost right away: extra squealing, chatter, or stronger heating of the part. In turning, these small things often point to wear more accurately than a parts counter.

It is better to watch not one symptom, but a combination. Usually, this checklist is enough:

• the size drifts toward the tolerance limit;
• the surface has become rougher;
• the chips have changed;
• spindle load has increased;
• noise or heat has intensified.

If at least two of these match, condition-based tool change usually gives a more accurate replacement moment than a hard limit by parts. Especially where the stock material varies a little from batch to batch.

How to introduce the approach without confusion

Do not start with the whole shop floor at once. Pick one operation where you most often lose money because of scrap or because the insert is changed too early. It is better to choose a repeatable turning operation on one part, where size is easy to check and tool changes already happen regularly.

First, collect a baseline on a stable batch. You need your own numbers: how many parts the tool lasts without size issues, what the surface looks like, and what load the machine shows in normal operation. If you are turning a shaft with a constant stock allowance, such a batch will quickly give you a clear starting point.

Then add 2–3 simple checks that the operator can perform without extra fuss. Usually three things are enough: the size starts moving toward the tolerance limit, the surface becomes noticeably rougher compared with a normal part, and the load rises on the same program and the same workpiece.

Condition-based tool change only works where the rule is clear to everyone on the shift. At the first stage, it is better to set one clear threshold. For example, the tool is changed before the counter if the size has moved closer to the tolerance limit twice in a row and the load has also become higher than usual.

How to make the rule part of daily work

Write the rule down briefly and without ambiguity. If it exists only as a verbal rule, everyone will interpret wear differently. One person will remove the insert too early, another will keep it until scrap appears.

Next, you need a simple log. It is enough to record the date, the operation, the tool number, how many parts it processed, and the reason for replacement. It is also useful to note separately whether scrap appeared before the tool change or the tool was removed early.

After one or two weeks, compare the notes. If the tool is almost always removed before the counter for the same reasons, the rule is working. If the signals are different every time, look deeper: the workpiece may vary in hardness, the cutting parameters may be too aggressive, or the clamping may create extra variation.

It is better to launch this calmly and on one operation. Once the rule starts hitting the right moment, you can move it to other parts and machines.

A simple shop-floor example

A turning department machines shafts from forgings. The operation is the same, the tool is the same, and the parameters hardly change. By habit, the tool is replaced by counter — after 80 parts.

On a stable batch, this still works. The stock allowance stays close to the planned value, the tool cuts smoothly, and size control stays within norm almost until the end of the cycle. Sometimes the insert lasts 75–80 parts without surprises.

Problems begin when a forging arrives with varying stock allowance. On some shafts, the metal comes off easily; on others, the tool takes extra material. The load jumps, the temperature rises, and wear progresses faster than the counter predicts. On paper, there are still twenty parts left before replacement, but in reality the size starts drifting after 55–60 pieces.

Usually, this is visible not from one signal, but from several. First, the operator notices that the size has to be adjusted more often. Then roughness increases, extra sound appears in the cut, and on inspection parts the tolerance band shrinks to an uncomfortable minimum. If the shop waits for the number “80” at that point, it gets early scrap.

Condition-based tool change works better here because it looks not at the plan, but at the actual tool life in that batch. On the uneven forging, the tool will be removed earlier, for example after 58 parts, and the batch will run smoothly. On the uniform forging, it will not be thrown away on the 60th part “just in case” — it will honestly last its 75–80.

The difference seems small, but in a shift it becomes noticeable quickly. The counter gives one number for all workpieces, even though the workpieces behave differently. A condition-based approach removes two expensive mistakes at once: changing a still-good tool too early and delaying replacement when size has already drifted.

If the shop machines parts that are effectively identical from the same material, tool change by count can still work. If the stock allowance varies even within one batch, it is better to look not only at the number of parts, but also at what is happening during cutting and inspection.

Where people most often make mistakes

The most common mistake is simple: trying to introduce the new rule across all operations at once. On paper, that looks convenient. In the shop, it almost never is. One operation runs smoothly, while another has varying stock allowance, the material behaves differently, and wear develops in a completely different way.

If you try to switch to condition-based changes everywhere at once, the team will quickly get confused. On one machine, the decision will save money; on another, it will cause extra stops and arguments about when the tool should be removed. It is better to choose one operation where scrap is expensive and size is critical, and work out the rule there first.

Another mistake is relying only on sound. An experienced operator really can hear a lot: squealing, chatter, a different cutting tone. But sound does not replace part size control. The tool may work almost normally, while the size is already slowly drifting toward the limit.

On CNC lathes, this is especially noticeable. The surface still looks acceptable, the chips come off evenly, but the diameter has already drifted by a few hundredths. If you rely only on hearing, it is easy to miss the replacement moment.

People also get confused when workpieces are not separated by material and stock allowance. A batch made from one material with even allowance will live by one scenario. The next batch, with higher hardness or varying allowance, will give a completely different tool life. That is why you cannot write one rule for all parts in a row. If the stock varies, split the parts into at least simple groups. Otherwise, it will seem like the method does not work, when in fact different conditions were just mixed together.

Another issue is the habit of running the tool until it chips. It may seem like this squeezes out a few more parts. In reality, the shop often loses more: it gets scrap risk, a re-setup, and an unplanned stop. Condition-based control works better when the tool is removed as soon as signs of deterioration appear, not after a failure.

And one more small thing that later makes it impossible to draw a proper conclusion: nobody writes down why the tool was removed. Without that, a week later it is already hard to remember what the reason was — size, surface, vibration, or edge chipping.

For the log, four points are enough:

• operation number;
• material or workpiece group;
• reason for tool removal;
• how many parts were processed by that point.

This takes less than a minute to record, but later it shows where the rule was accurate and where it needs to be adjusted. Without these notes, the decision again turns into guesswork.

What to do next

Do not try to rebuild the entire shop floor right away. It is much more useful to take one part and one tool where wear is already more or less understood. That way, you will quickly see whether the condition-based approach brings real benefit or just adds extra checks.

It is best to start with an operation where the counter already makes mistakes sometimes. For example, on one batch the tool comfortably reaches the limit, while on another the size drifts earlier because of different workpiece hardness or stock allowance. In such a situation, condition-based tool change often works more accurately than tool change by counter.

Run a simple test on two or three batches. Track one part of the production by the usual counter, and on the other add a check of the actual tool condition. Do not complicate the scheme. You do not need a beautiful table for the sake of a report, but a clear difference in scrap, downtime, and the number of premature replacements.

For the first step, a short record is enough: batch number, part size at inspection, time or number of parts to changeover, reason for tool removal, and a brief note on the chips, sound, or surface finish. That is already enough to see the picture.

If the difference is noticeable, make the rule permanent for that one operation. There is no need to write a general standard for the whole shop right away. First, get a stable result where you can already see the gain in part size, changeover time, and tool life.

If the issue is not only the insert, but the repeatability of the whole process, it is worth looking wider: at the machine, system rigidity, workpiece, and setup. EAST CNC, the official representative of Taizhou Eastern CNC Technology Co., Ltd. in Kazakhstan, works with exactly these tasks: selection of CNC lathes, supply, commissioning, and service. On the EAST CNC blog at east-cnc.kz, you will also find practical materials on metalworking if you need to compare approaches on real operations.

A good start looks simple: one part, one tool, several batches, and an honest record of why the tool was changed. That is usually enough to understand where a counter is convenient and where it is already time to look at the actual wear of the cutting tool.