Turning Flanges with Uneven Stock Allowance Without Shock

Where shock comes from on a cast flange

Shock on a cast flange almost always starts not with the cutting tool, but with the workpiece itself. On cast parts, the allowance around the circumference is rarely even: one area may already be close to size, while another is noticeably thicker. Add ovality, a shifted center, and slight runout after clamping, and cutting becomes uneven right away.

The tool does not remove metal equally through the whole revolution. On one part of the circumference it barely cuts or takes a thin chip, and a moment later it bites into a thicker section. That transition is usually the first thing the operator hears: the sound changes, the machine starts working harder, and the surface shows signs of unstable material removal.

The problem gets worse if the first roughing pass is too deep. On paper, 2 mm per side may look fine, but on the thick sector the actual allowance is sometimes twice that. Then the load rises sharply on the clamping setup, the tool, and the spindle at once.

That is already unpleasant on a rigid short part. On a large-diameter flange or a thin-walled one, the effect is more noticeable. The part flexes slightly on the thick sector, then springs back, and the size changes in steps rather than smoothly. You may not always see it on the indicator right away, but the surface quickly develops bands, local scoring, and a repeating pattern around the circumference.

A typical mistake looks simple. The process engineer sets the first pass at 2 mm per side, assuming the allowance is roughly even. But on one sector the actual stock is 4 mm, not 2. For the machine, that is a completely different condition. The tool enters the metal as if the feed had suddenly increased in the middle of the revolution.

After that kind of entry, it is not only the surface that suffers. The part may shift a little in the jaws, the tool wears faster, and the outer diameter or face dimension starts to drift. Then extra passes are needed to remove the marks and recover the geometry. That is why shock on a cast flange should be treated not as a minor nuisance, but as the first warning sign: the allowance needs to be broken down carefully, not removed in one bold pass.

What to measure before the first pass

A cast flange rarely gives a uniform allowance around the whole circumference. If machining starts based on an average size, the tool will first meet the highest point and take a shock. Before the first cut, you need to understand not only the size, but also the shape of the blank relative to the datum you plan to use.

Set the part up the way it will sit in the real clamping arrangement. Then check the outer diameter, the face, and the locating surfaces with an indicator. What matters is not only the number on the dial, but the shape of the deviation: where the rise starts gradually, where the needle jumps sharply, and where the surface simply wanders.

It is best to find the highest point around the circumference separately and mark it immediately with a marker. Turn the part slowly, without jerks. The maximum indicator reading will show the point where the tool will meet the largest allowance. In practice, that is often what decides whether the first roughing pass goes smoothly.

An average value is not very helpful here. It is much more useful to measure the actual allowance at least in 4 to 8 points around the circumference. On a large flange, it is convenient to check every 45 or 90 degrees. Then you see the real spread, not a nice average number that explains nothing.

Usually it is enough to record the minimum and maximum allowance on the outer diameter, face deviation around the circle, offset of the fit or center hole relative to the chosen datum, the most protruding sector, and areas with suspicious scale, porosity, or signs of warping.

Separate warping from casting offset. They are different things. If the face runs in a wave while the outer diameter stays more or less stable, the part is warped. If one side is consistently thicker and the opposite side thinner, the casting is simply shifted relative to the datum. For rough machining, the difference matters a lot: in the first case the tool meets a changing height, in the second it faces uneven stock around the circle.

A simple example. The outer diameter is supposed to finish to size after removing 3 mm per side. When measured in eight points, the spread turned out to be from 1.5 to 6 mm. The highest area is roughly at the 11 o’clock position, the face shows a wave of about 0.7 mm, and the fit surface is offset by almost a millimeter. After a check like that, it is already clear that a normal even first pass will hit metal hard and may leave almost no allowance on the opposite side.

This check takes only a few minutes. But it immediately shows where the tool will meet metal first, where the allowance will end sooner, and how reliable the chosen datum really is. That is cheaper than replacing an insert after the very first shock.

How to choose the datum and clamping

On a cast flange, the datum often matters more than the cutting parameters. If the part is referenced to a surface that already moves around, the tool will catch shock on every revolution even with cautious feed.

For turning flanges with uneven stock allowance, the datum should be chosen not from the most convenient place, but from the surface that truly holds the axis. The cast outer diameter is a poor choice for that: it often drifts sideways and varies from part to part. It is more reliable to use the center hole, the cast hub land, or the face together with the internal fit, if those zones repeat across the batch.

If you do not yet have a clean datum, it is better to create one with a short preparatory pass. Often it is enough to lightly face the part and remove a narrow land that can later be used to clamp it again with confidence. That is faster than chasing runout through all the roughing passes.

The clamp must hold the part, but not distort it. A thin flange is easy to pull out of shape with the jaws. On the machine it may look fine, but after unclamping the size will drift. If the metal is soft or the wall is thin, reduce clamping force and increase the contact area. Soft jaws bored in setup usually give a more predictable result than random three-point contact.

If the part will need to be turned over later, do not remove all the stock from the future datum in one setup. Leave a land or a face that can be used to grip the part again. Otherwise, during the second clamping you will be relying not on a proper datum, but on traces of the previous machining, and the axis will start wandering again.

There is another common cause: the runout comes not from the flange, but from the fixture. The faceplate may be mounted slightly off, soft jaws may have been bored in the wrong position, or the chuck may pull unevenly. Before starting, it is worth checking chuck runout without a part, jaw runout after boring, the part’s contact with the supporting surface, and whether tightening shifts the part in the last few degrees of clamp rotation.

If the flange is large and heavy, do a test clamp without cutting and rotate the part at low speed. The sound and the indicator quickly show where the issue is: in the casting, the datum, or the fixture.

Roughing sequence, step by step

On a cast flange with wandering stock allowance, it is not a good idea to start with a normal roughing pass across the full circumference. If the layer is uneven, the tool will cut easily in one place and then suddenly bite into a thick section. That leads to noise, shock load, edge chipping, and vibration marks. It is better to open up the part’s real geometry first, then divide the metal into manageable layers.

The logic is simple: first find the highest areas, then remove them locally, and only after that move on to normal rough machining. At first this may seem slower, but in practice it often saves time. One ruined insert costs more than a couple of short local cuts.

Start with a light trial pass and a small depth of cut. Its job is not to remove volume, but to show where the flange already touches the tool and where the metal has not been reached yet. After that contact, the protruding arcs and the empty zones are easy to see on the surface.

Do not try to go straight around the whole circle at the same depth. Remove the highest areas with short cuts along the arc. If one sector has 2 to 3 mm more stock, remove that sector first until the difference becomes smaller.

Once the big jumps are gone, the remainder can be split into 2 to 3 roughing passes with a predictable load on the tool. At that point, full-circumference passes are acceptable, because the layer is much more even. It is better to leave a small allowance for the next pass than to take too much at once.

Change the feed not along the entire path, but before entering the thick sector. On the thinner area the machine can move faster, and where the allowance grows again, the feed should be reduced. That keeps the cycle from stretching needlessly and prevents the tool from taking a shock at the hardest point.

After each cut, stop and look at the surface. The contact marks immediately show where excess material is still left. If one sector is still higher than the others, remove that one locally again rather than across the whole diameter.

In practice it looks like this: the first pass only sketches the shape, the second and third cut down the raised areas, and after that the machine runs more evenly, without sudden jumps in load. On large flanges, this is especially noticeable, because an early mistake quickly turns into vibration across the whole part.

If the tool still enters hard after leveling, do not rush to increase the speed or change the insert. First check whether one local hump remains and is still breaking the evenness of the layer. Usually that is the reason.

Example of a flange with wandering stock allowance

Imagine a cast flange with a diameter of 420 mm. After measuring, you see the usual picture: on most of the circumference the allowance is nearly even, but from about 11 to 2 o’clock the casting stands out more. In that quarter there are 2 mm of extra stock compared with the rest of the circle.

If you immediately use the same depth of cut across the entire diameter, the tool will not run smoothly everywhere. On most of the circle it will remove a normal layer, but in the thick sector it will suddenly hit the extra metal. That is where the shock comes from: the load changes, the machine jerks, the edge heats up, and marks remain on the surface.

In that situation, a calmer sequence works better. On the main circle you may have, for example, 3 mm per side, while the problem quarter has 5 mm. The rough size needs to leave 1 mm for the next pass. So it makes sense to enter only the problem sector first and remove about 1.5 to 2 mm per side with a lower feed. Not all the extra metal at once, only part of it.

After that sector pass, the difference around the circumference is no longer 2 mm, but about 0.5 to 1 mm. For rough machining, that is a completely different situation. Now you can make a circular pass around the whole circumference. The tool cuts almost evenly everywhere, without sharp changes in load, the sound becomes steadier, and the part stops knocking the tool at the same angle.

If the allowance varies a lot, do not be greedy on the first sector removal. Two short passes on the heavy area are usually more reliable than one aggressive one. This is one of those cases where a calmer start really saves time.

If the cutting sound still changes at the same angle after the first sector pass, do not continue the full circle at the same settings. Measure that area again. Most often, a local bump is still left there and needs to be removed separately.

Where mistakes happen most often

The most common mistake is simple: the operator looks at the drawing, sets the depth of cut based on the nominal value, and assumes the casting is close to the finished shape. That rarely happens on a cast flange. The stock around the circumference can vary noticeably, and the tool cuts calmly in one place but gets an extra load half a turn later. The problem is not that the cutting conditions are bad by themselves. The problem is that the real part does not match the calculated one.

The second mistake comes right after that: only the outer diameter is checked, while the face is ignored. If the face runout is larger than expected, the tool gets a shock even at a moderate depth. From the outside the part may look fine, but the face is already creating a poor picture on the first pass. Then people start changing speed and feed, when the real first step should have been checking the datum, the clamping, and the actual runout on both surfaces.

When feed does not save you

Reducing feed helps, but it is not a cure for everything. If the stock is much thicker in some places, a lower feed only makes the shock quieter. The cause is still there. Sometimes an overly cautious setting even gets in the way: the tool starts rubbing where the metal is already removed, while still taking a hit in the thick spot.

In practice it looks like this: on the first part the operator hears a shock, lowers the feed, the pass sounds quieter, and it seems the problem is gone. Then the tool wears out quickly, and the size drifts after a couple of parts. That means the issue was not only the feed, but also the roughing strategy.

Another common mistake is leaving different stock for the finishing pass. After rough machining, the flange already looks better, and it is easy to relax. But if one sector has only 0.5 mm left for finishing and another still has 1.8 mm, the problem simply moves to the next stage. The finishing tool does not like that kind of spread.

The mistake that hurts the whole batch

After the first successful part, it is easy to think the method is solved. That is a dangerous moment. On cast flanges, two neighboring blanks often behave differently. If you start the batch without checking again, you can end up with the same mistake across the whole lot.

Before the batch, it is better to quickly verify three things: where the thickest allowance remains after roughing, whether face runout changed after re-clamping, and whether you are leaving the same stock for finishing around the circumference.

That check takes only a few minutes, but it saves tools and nerves. In CNC flange machining, the mistake rarely lives in a single number. It is usually a combination of uneven casting, an incorrect runout assessment, and too much confidence that the first part already proved everything.

Quick check before startup

Before the cycle, it is better to spend two minutes on inspection than later to change an insert and chase shock marks through the entire batch. The biggest mistakes are usually visible before the first cut.

If the flange is cast, first rotate the part slowly and find the thickest point around the circumference. You can often see it even without cutting: from runout, the casting trace, the difference in surface color, or the indicator reading. If you do not find that area in advance, the tool will, and usually too abruptly.

The clamp has its own trap. The jaws may hold the part tightly but still slightly crush a thin flange. After tightening, check the face and outer diameter again. If the readings changed more than before clamping, you have already introduced an error into the datum.

Before starting, it is useful to do a short check: mark the zone of maximum allowance, verify whether the part shifts after final tightening, estimate how even the stock is for finishing, rotate the part and listen at the same angle, and start the first run at reduced feed or in single-block mode.

A finishing allowance does not have to be large. It is much worse when one area has almost nothing left and another still carries a heavy roughing step. Then the finishing pass will not correct the geometry; it will just repeat the problem in a neater form.

Pay attention to the sound from the very first revolution. If the tool enters smoothly at one angle but bumps and hums at another, the cause is almost always one of two things: extra stock in a local zone or a slight shift of the part in the clamp. One hard entry can still be tolerated; a second one often breaks the edge.

A good habit is simple: on the first pass, you listen and watch instead of walking away from the machine. Keep your hand near the stop, have feed correction ready, and know the tool path in advance. If the entry is harder than expected, it is better to stop the cycle immediately and correct the plan.

What to do after the first part

The first part gives you facts, not a final answer. After that, you can already see where stock is still left with margin and where the tool enters deeper than expected almost immediately. If you do not record that, the second part often goes worse than the first, even though nobody changed the settings.

First compare the real stock map with what you expected before startup. Go through the same points you measured before machining and compare the remaining material after the first roughing passes. If one sector cuts hard and another runs easily, do not change the whole strategy at once. Adjust the first passes only so the heaviest step is removed more gently.

It helps to keep a short note for the first part instead of relying on memory. It is enough to record the sector where the tool made noise or caught a shock, the entry point for the first roughing pass, the depth of the first 2 to 3 passes, and the cutting order that solved the problem in the heavy area.

Often the key is not a smaller depth by itself, but a different order of work: first remove the protruding sector, then go around the circumference more evenly. A day later, that small detail is easy to forget, even though it is exactly what saves time and tools.

For the next batch, create a simple inspection standard. It does not need to be complicated. Four to eight points around the circumference are enough, where the operator quickly checks the outer diameter, the face, or the remaining stock after the first rough touch. If the casting variation repeats, that template immediately shows whether the batch is stable.

It is convenient to put this on a single sheet for the machine: measuring points, acceptable spread, order of the first passes, and the condition under which the operator must stop and recheck the clamp. That removes a lot of guesswork from the shift.

If these parts are produced in a series, it is also useful to compare the machining plan with the machine’s capabilities. The EAST CNC blog publishes machine reviews and practical metalworking tips, and the company itself handles selection, delivery, commissioning, and service support for CNC lathes in Kazakhstan and other CIS countries.

When you already have a stock map after the first part, a corrected first pass, and a recorded cutting order, rough machining becomes much more predictable. For flanges like these, that is the main result.