Can a Servo Feeder Hold Length on Oiled Strip?
A 2.3 mm HRPO coil arrived with a heavier mill oil than the last one, and feed length drifted 0.3 mm over four hours. Nothing on the machine had changed. The friction between strip and roll had.
The plant had run the same part for two years on dry-ish CR stock. The switch to oiled hot-rolled pickled coil was a purchasing decision, and nobody connected it to the scrap bin filling up six weeks later.
This is the sequence that found the cause, and the numbers that made it repeatable.
The Drift That Started With a New Coil Supplier
The complaint was a part 0.3 mm short at the last station, appearing after the third hour of a run and disappearing after a coil change.
Operators had already adjusted the feed length by hand, which is the usual first response and the reason the fault stayed invisible for a month.
What made it worth investigating was the pattern. The drift always moved in one direction within a coil and reset on the next. A mechanical wear problem does not reset.
That pattern points at something that varies with the coil: surface condition, coil temperature, or the oil film thickness on the strip.
The old stock was CR4 with a light protective oil, roughly 1 µm film. The new coil carried 3 to 4 µm of drawing lubricant, applied at the mill for a downstream deep-draw customer.
How an Oil Film Changes the Friction Budget
A servo feeder moves strip by friction alone. The feed rolls press on the strip, and the only thing stopping the strip from sliding backwards during acceleration is the friction coefficient between roll and steel.
Dry-ish CR sits near 0.15. A 3 µm oil film drops that to 0.08 on a polished roll, and lower still if the oil is warm and low viscosity.
The force needed does not change. Moving 310 mm of 420 mm wide strip at 180 strokes a minute takes the same acceleration whether the strip is oiled or not. Only the available grip changes.
So the normal force has to nearly double. Holding the same friction force at 0.08 instead of 0.15 requires 1.9 times the roll pressure. A machine set up on dry stock has none of that margin.
That explains the drift direction too. As the roll surface picks up oil, friction falls further through the run, so the strip slips a little more each hour until the operator re-trims the length.
Temperature closes the loop. Oil viscosity at 20°C is roughly double what it is at 35°C. A shop that warms up through the morning changes its own friction coefficient without touching the machine.
Three Measurements That Named the Cause
Three readings separated an oil problem from a worn roll, and each one took under ten minutes.
| Measurement | What it told us | Where the number lies |
|---|---|---|
| Strip slip mark length after a stop-start cycle | Grip was marginal, not lost: 1.2 mm of slip per acceleration | Reads zero if the test runs at low speed, where acceleration force is small |
| Feed length at 40 SPM against 180 SPM | Error grew with speed, which rules out encoder and mechanical backlash | Hides the fault if the line never runs below 150 SPM in production |
| Roll surface wipe test with white paper | Visible oil transfer after 200 strokes; a dry roll leaves nothing | Says nothing about film thickness, only that oil is present |
The speed comparison is the one that matters most. An encoder fault produces the same error at every speed. A friction fault scales with acceleration, so it grows as the line speeds up.
Backlash in the drive train behaves like an encoder fault and would have been the next suspect. The speed test removed it before anyone stripped a coupling.
What Fixed It and What It Cost in Roll Life
Four changes went in over one weekend. Two of them cost nothing and did most of the work.
| Change | What it recovers | What it takes back |
|---|---|---|
| Felt wiper on the entry side of the feed rolls | Removes most surface oil before the grip point | Felt pads changed every two weeks; a clogged pad pushes oil back onto the roll |
| Roll pressure raised to 6.5 bar | Restores grip margin at 0.08 friction | Roll coating life down about 30 percent, and marking risk on soft strip |
| Coated roll surface, 60 HRC, fine shot blast | Holds friction with an oil film present | Roughly $1,800 a pair, and a four-day lead time |
| Feed length compensation by speed band | Covers the residual error at high SPM | Hides a real fault if the bands are set from one coil |
The wiper and the pressure change together recovered 0.24 mm of the 0.3 mm drift. The coated rolls closed the rest, and the speed-band compensation was never enabled.
That last point matters. Compensation tables are the easiest fix to add and the hardest to remove, and they will mask a worn roll for a year.
Limits of Roll Pressure as a Fix
Pressure is the first lever everyone reaches for, and it stops working in three situations.
Thin strip deforms before it grips. Above about 5 bar on 0.6 mm aluminium, the roll flattens the strip instead of holding it. The feed length becomes unstable in the other direction.
Coated strip marks. Galvanized, prepainted and polished stainless all take roll impressions at pressures that a bare CR coil tolerates without complaint.
The oil is a drawing lubricant, not a rust preventive. Some mill oils carry solid additives that fill the roll surface texture and drop friction regardless of pressure. A wiper helps; pressure does not.
Where all three apply, the answer is a different roll surface or a different oil specification, and both are purchasing decisions rather than maintenance ones.
Purchase Order Wording That Prevents a Repeat
The plant solved the machine problem and left the supply problem open. One line of purchase order wording would have avoided both.
- Name the oil: product, film weight in g/m², and whether it is a drawing lubricant or a rust preventive.
- State the surface the line can grip: dry, lightly oiled, or heavily oiled, with the film range you will accept.
- Require a sample coil before the first production order, and run it at the highest SPM the line will see.
The sample requirement is the cheap one. One coil of 400 kg costs less than an hour of scrap on a progressive die.
Running the Same Test on Your Own Line
The whole diagnosis needs one shift, one dial indicator and a stopwatch. No special tooling and no supplier visit.
Run the line at the production speed and record feed length every fifteen minutes for two hours. Plot the four values. A drift that grows and resets per coil is friction. A drift that grows and stays is wear.
Then run the same part at half speed. If the error halves with the speed, the grip is marginal and the oil is the first suspect.
Our own builds hold ±0.05 mm on oiled stock when the roll surface is specified for it, which is why the roll coating question comes up at the quotation stage rather than after commissioning.
The coil line guide library breaks down the full error budget from coil to die, including the share that belongs to friction.
Roll diameters and coating options are listed by frame size in our feeder data library.
Six questions come up every time this fault is discussed on a shop floor, and the answers are shorter than the arguments about them.
How much oil film is too much for a servo feeder to grip?
Above roughly 2 µm of drawing lubricant, grip margin falls below what a 180 SPM line needs at standard roll pressure. A rust-preventive film of 0.5 to 1 µm is normally not a problem.
Does raising roll pressure to 8 bar damage the strip?
On 2.3 mm CR it leaves visible impressions on the top surface. On 0.6 mm aluminium the strip thins under the roll before it grips, and feed length becomes unstable.
Why does the feed drift 0.3 mm and then recover on the next coil?
Oil builds up on the roll through a run and is wiped off during the coil change. That reset is the signature of a friction fault rather than a mechanical one.
Can a wiper felt replace a coated roll?
It recovers about 80 percent of the lost grip on a 3 µm film and costs $40 a pad. It will not recover grip on a 6 µm drawing lubricant, where the roll surface has to change.
How much roll life does the extra pressure cost?
Around 30 percent at 6.5 bar against 4.5 bar. A coated pair that lasted 14 months on dry stock runs about 10 months on oiled stock at the higher setting.
Does an NC servo feeder need a different setup for oiled coil?
It needs the same setup plus two additions: a wiper before the grip point and a pressure setting derived from the oil, not from the previous job. Everything else stays as it was.
Where to go next
Grip is one line in an error budget that starts at the coil and ends at the die, and it is rarely the largest one.
Our coil line knowledge base lays out the full budget, element by element, with the cost of reducing each one.
