Why Does Feed Acceleration Limit NC Servo Feeder Speed?
Two lines, same feeder, same strip. One holds plus or minus 0.05 mm at 200 strokes a minute. The other loses 0.3 mm the moment it goes past 140.
The difference is almost never the motor. It is the acceleration the move demands, and whether the strip can survive it.
This is the arithmetic behind NC servo feeder speed limits, worked through on a real feed length, with the sensitivity table that tells you where your own line will fall over.
The wider error budget this sits inside is laid out in the coil line guide library; the machines are listed in the product range.
The Index Time Budget on a 55 mm Feed
Start with how much time the feeder actually has, because it is far less than the press cycle suggests.
At 200 strokes a minute, one cycle is 300 milliseconds. The strip has to be stationary while the die is closed and working.
On a typical progressive die that leaves roughly 55% of the cycle for the index, or about 165 milliseconds. That is the whole budget.
Inside those 165 ms the strip must accelerate from rest, travel 55 mm, and come back to rest. Nothing about the machine changes that arithmetic.
A symmetric trapezoidal move with 40 ms of acceleration and 40 ms of deceleration leaves 85 ms of cruise.
The distance equation is simple: travel equals cruise velocity multiplied by the cruise time plus the ramp time.
So 0.055 metres equals velocity multiplied by 0.125 seconds, giving a cruise velocity of 0.44 metres a second, or 440 mm per second.
Peak acceleration is velocity divided by ramp time: 0.44 divided by 0.040, which is 11 metres per second squared, a little over 1 g.
That is a comfortable number. Most NC servo feeders are designed to run at 1 to 1.5 g all day.
How Acceleration Eats the Available Time
Now push the press speed and watch what happens to that comfortable number.
At 400 strokes a minute the cycle halves to 150 ms and the index window falls to about 82 ms.
Keeping the same 55 mm feed, the ramp times have to shorten. Take 25 ms to accelerate and 25 ms to decelerate, leaving 32 ms of cruise.
Cruise velocity becomes 0.055 divided by 0.0575, which is 0.96 metres a second. Peak acceleration is 0.96 divided by 0.025, or 38 metres per second squared, close to 4 g.
That is the whole story in one comparison. Doubling the press speed multiplied the acceleration demand by roughly four, not by two.
The reason is that acceleration depends on velocity divided by ramp time, and both of those scale inversely with the available window.
This is why feed length matters as much as speed. A short feed has room to spare; a long feed at the same speed is a different machine.
| Feed length | Peak velocity at 200 SPM | Peak acceleration at 200 SPM | Where it costs you |
|---|---|---|---|
| 25 mm | 200 mm/s | About 0.5 g | Nothing much; roll inertia dominates and tuning is easy |
| 55 mm | 440 mm/s | About 1.1 g | Strip slip starts to matter on oily or coated surfaces |
| 100 mm | 800 mm/s | About 2.0 g | Peak torque climbs, and a light strip may not survive the start ramp |
| 200 mm | 1,600 mm/s | About 4.0 g | Beyond most standard roll feeds; needs a longer window or a different feeding principle |
The window fraction is the lever nobody checks. If the die can give you 65% of the cycle instead of 55%, every acceleration figure in that table drops by about a quarter.
Where Inertia Match Sets the Ceiling
Acceleration is only half the limit. The other half is how much of the motor's torque goes into accelerating the roll assembly instead of the strip.
Every servo-driven feed has a load-to-motor inertia ratio. A ratio of 3 to 1 or lower tunes easily; 5 to 1 is common; beyond 10 to 1 the response becomes hard to hold and the loop starts to hunt.
Roll diameter is the biggest single contributor. Inertia rises with the fourth power of diameter. Going from a 100 mm roll to a 130 mm roll multiplies roll inertia by roughly 2.9.
That is why a heavy-duty feeder with large rolls feels sluggish on a short feed where a lighter machine is crisp.
Strip mass adds to it. A 1,500 mm wide, 2 mm strip loop has real mass, and at 4 g the strip itself becomes part of the load.
Check the continuous torque rating, not the peak. Peak torque is available for a second or two. At 200 strokes a minute you are asking for that peak 3.3 times a second, all day.
A servo motor that is comfortable on a 25 mm feed at 300 strokes a minute may be thermally overloaded on a 100 mm feed at 200.
What S-Curve Ramping Buys You
S-curve ramping is often sold as a way to go faster. It is not, and understanding why changes how you set the machine up.
A trapezoidal profile with an instant step in acceleration is time-optimal for a given acceleration limit. Nothing moves a mass between two points faster.
Adding jerk limiting smooths the corners, which means for the same move time you need a slightly higher peak acceleration, or for the same peak acceleration a slightly longer move.
What you gain is a lower rate of change of acceleration. That matters more than it sounds.
Less strip slip. The initial grip is applied gradually, so the friction limit is approached rather than jumped over.
Less mechanical shock. Gearbox, coupling and roll bearings see a gentler load step, and the audible clack at the start of each index disappears.
Less resonance excitation. A step input excites every natural frequency in the assembly. A ramped input excites fewer.
| Ramp profile | What it gives the run | Where it costs you | Where it misleads |
|---|---|---|---|
| Trapezoidal, hard step | Shortest possible index time, easiest to model | Maximum shock and the highest risk of slip at the start ramp | Looks fastest on a scope trace while quietly wearing couplings |
| Moderate jerk limit | Most of the speed, noticeably less slip and noise | A few milliseconds more move time, and tuning takes longer | Settings that work on steel may still slip on coated strip |
| Heavy jerk limit | Smoothest engagement, best for thin or coated material | Meaningfully longer move time, so it caps the achievable strokes per minute | Can hide a genuine grip problem that needs a roll change, not a softer ramp |
The practical setting is usually a moderate jerk limit, tuned until the strip no longer slips at the start ramp on the worst material you run.
Beyond that you are buying quietness, not accuracy.
Where Acceleration Limits Do Not Apply
The arithmetic above is a ceiling, and it is not always the ceiling that bites first.
Short feeds are not acceleration-limited. Below about 30 mm, the move is dominated by settling time and the servo's own response, not by the ramp. Adding a bigger motor buys nothing.
A long window makes it moot. If the die gives you 70% of the cycle, most lines run out of press capability before they run out of feed acceleration.
Slow presses with long feeds. A 60 stroke a minute line feeding 300 mm is limited by loop stability and material handling, not by ramp time.
Coated and thin strip. Here the limit is friction, and it arrives well below the acceleration the machine could deliver. The fix is roll surface and pressure, not more torque.
Loop control, not the feeder. A feeder that hunts at speed is often fighting a loop that is too short or a pay-off tension that rises as the coil empties. The acceleration figure was never the problem.
After twelve years building these machines we still ask for the die's window fraction and the strip surface before quoting a maximum speed. A number quoted without those two is marketing.
What to Measure on the First Run
- Feed length at jog speed, then at production speed. The two numbers differ. Only the second one counts, and the gap tells you how much slip you have.
- Torque demand on the drive display. Compare continuous demand against the motor's continuous rating. Above about 70% sustained, expect thermal drift by mid-shift.
- Slip marks on the strip at the first 100 strokes. Run the worst material you have, not the best.
- Repeatability over 50 strokes. Take a sample every ten strokes and record the spread. That spread is your real accuracy, not the single best reading.
- Loop behaviour at the top of the coil and at the remnant. Pay-off tension changes with coil diameter, and a setting that holds at full coil may not hold at 30% remaining.
- Speed at which the first slip appears. Find it deliberately. Knowing your actual ceiling is worth more than believing a catalogue figure.
Record the four numbers that matter on the machine card: maximum speed before slip, feed length spread, continuous torque demand and loop depth at the remnant.
Whoever runs that line next shift will need them.
At what stroke rate does acceleration typically become the limiting factor?
For a 50 mm feed on a 55% window, expect acceleration to dominate above roughly 250 strokes a minute. Below that, loop stability and material grip usually limit you first.
How much feed length can a standard NC servo feeder handle at 200 SPM?
Up to about 150 mm on a 55% window before peak acceleration passes 3 g, which is beyond what most light strip will tolerate without slip.
Longer feeds are possible at lower speeds, or with a longer window negotiated with the die designer.
Does a heavier feeder give better accuracy at high speed?
Not automatically. Larger rolls raise inertia with the fourth power of diameter, so a heavier machine can actually respond more slowly on a short feed.
Match the roll size to the feed length rather than buying the largest frame available.
How much does a 10% longer index window reduce acceleration demand?
Roughly 17%, because acceleration scales with the inverse square of the available time. Ten percent more time is worth more than ten percent more motor.
What inertia ratio should I ask for on the drive?
Target 5 to 1 or better between load and motor inertia. Above 10 to 1, expect to spend tuning time and accept a softer response.
Can jerk limiting be added to an older feeder?
Usually yes, as a parameter change in the drive rather than a hardware change. Check that the controller exposes a jerk or S-ramp setting; many older units only offer a trapezoidal profile.
Where to go next
Acceleration is one line in a larger error budget that runs from the pay-off to the die. The full breakdown, with the contribution each source makes, is in the guide library.
the coil line guide library
