How to Restart a Servo Feeder After a Power Cut?
The plant loses power at 14:20 with a strip halfway through the die. By 14:40 the press is running again, and by 15:10 someone finds the first 40 parts are 0.4 mm short.
A servo feeder restart is not a reset button. It is a reference recovery, and the order of the steps decides whether the first part is good.
Stop Before You Reset: What the Drive Remembered
Every restart question starts with the encoder type, and most operators do not know which one their machine carries. Ask the maintenance file before the power fails, not after.
An absolute encoder keeps multi-turn position through a power loss, held by a small battery in the drive or the motor. An incremental encoder forgets everything the moment the DC bus collapses.
That difference decides whether the line needs a full re-reference or a position check. Treating an absolute machine like an incremental one wastes 30 minutes; the reverse scraps parts.
Look for the battery alarm in the drive's fault history first. A low-battery warning that was ignored for two months is the most common reason an absolute system comes back blind.
Write the battery date on the inside of the cabinet door. It is a two-dollar part that has cost more production than any other component on the machine.
Replace it on a five-year cycle, or sooner if the drive logs a voltage warning. Waiting for a failure means waiting for a scrapped strip.
Six Steps From Dead Power to First Good Part
The sequence below works on any NC servo feeder, whatever the control brand. Skipping step three is the mistake that produces short parts nobody can explain.
- 1. Isolate and record. Note the fault code, the press crank angle and the strip position before anyone touches a button. That record is the only evidence of what the machine was doing.
- 2. Clear the strip. Cut or withdraw the material inside the die. Never jog the feed with a part still gripped by a closed station.
- 3. Re-reference the axis. Home to the fixed stop or reference block, then confirm the displayed position against a tape measure on the strip.
- 4. Restore the pneumatics. Check roll pressure and pilot release pressure at the gauge, not at the regulator knob. Air has bled out and the knob position proves nothing.
- 5. Dry cycle. Run 20 strokes with no strip and the die open. Watch for servo following error and any mechanical noise that was not there before.
- 6. Marked first piece. Feed a strip with a chalk mark at the reference, run one stroke, and measure the mark to the die centre. That single measurement replaces an hour of guessing.
The Encoder Battery That Dies Mid-Move
A battery that fails while the axis is stationary loses position cleanly. A battery that fails mid-move can corrupt the multi-turn count, and the drive will happily report a position that never existed.
The symptom is a feed length that is consistently wrong by a round number: 300 mm commanded, 296.4 mm delivered, every stroke. That is a lost revolution, not a mechanical slip.
Recovery means re-establishing the reference and re-checking the electronic gear ratio against the roll diameter. Both numbers are in the commissioning record if anyone kept one.
If the record is missing, derive the ratio from a measured feed. Command 300 mm, measure the strip, and adjust the ratio by the difference.
Two iterations usually land inside 0.1 mm, close enough to run the job and finish the calibration properly.
Then run ten strokes and measure again.
A ratio that is right on the first stroke and wrong on the tenth points to slip, not to calibration.
Replace the battery with the drive powered, or accept a full re-reference afterwards. Most drives warn about this in the manual and nobody reads that page until it matters.
The Strip That Was Inside the Die
If power failed during a stroke, the strip in the die is the real risk. A crank that stopped 40 degrees before bottom dead centre is holding material under load.
Do not release the press until you know the crank angle. Rotating backwards with a part gripped can tear the strip and mark the die, and on a 200 SPM press the damage is measured in toolroom hours.
Cut the strip clear, remove the part, then check the die for sheared pilots and pulled slugs. A 0.1 mm burr left in a station will show up as a height error on the next 500 parts.
Check the feed rolls as well.
A stop under load can leave a flat spot on a polyurethane roll, and that flat spot marks every part from then on.
Roll a strip by hand through the open rolls and feel for a periodic bump. It takes a minute and catches a fault that otherwise costs a roll set.
Settings to Record So the Next Restart Takes 20 Minutes
| Setting | Why it decides the restart | Where it costs you if lost |
|---|---|---|
| Feed length and gear ratio | Defines millimetres per motor revolution | Every part short or long until recalibrated |
| Roll pressure (bar) | Sets grip on the strip surface | Slip at speed, invisible at low speed |
| Pilot release timing | Coordinates feed with pilot entry | Bent pilots on the first stroke |
| Acceleration profile | Limits following error at 200 SPM | ±0.05 mm becomes ±0.15 mm at speed |
Photograph the parameter screens and keep the images in the line file. A phone photo has recovered more production than most service contracts.
Record the roll diameters as well. Feed length is calculated from roll circumference, so a roll dressed down 0.5 mm changes the delivered length by about 1.6 mm per metre of feed.
Keep that record in the line file, not on a phone that leaves with the operator. The next restart may happen on another shift with different people.
Date the record and note who measured it. A calibration sheet without a date is a rumour.
Keep a second copy with the spare drive.
If the drive is replaced in a breakdown, the parameters go back in from the sheet instead of from memory.
Restore the parameters before re-referencing, not after. A drive that homes with the wrong gear ratio will home to the wrong place.
Limits of a Clean Restart
A clean restart tells you the axis moves and the press cycles. It does not tell you the line is accurate, and it certainly does not tell you the die survived.
Feed accuracy has to be measured on the strip over 50 strokes at production speed, not on the display at 30 SPM. Errors that appear only at 200 SPM will pass every static check.
The checklist also cannot recover material. A strip damaged by a mid-stroke stop is scrap, and no amount of re-referencing changes that.
Finally, a restart does not diagnose the original fault. If the power cut was caused by a drive trip, the same trip returns next week unless someone reads the fault log while it is still fresh.
Where the Feed Error Actually Comes From
Most restarts end with the axis correct and the parts still out of tolerance. At that point the feeder is innocent and the error lives somewhere else in the line.
Strip tension, loop height, roll condition and pilot release timing each contribute a share. Chasing them one at a time without a budget wastes days.
The error budget that separates those contributions, and the acceptance test that proves each one, is laid out in our coil line knowledge base.
For the timing side specifically, this guide to pilot release timing covers the check that catches a restart error before the first bent pilot.
The questions below arrive with almost every restart call we take.
Model capacities and control options for the feeder range are listed by tonnage and feed length.
How long should a restart take on a 300 mm feed?
Twenty minutes with a written procedure and a marked strip. Without the reference record, plan on 90 minutes and expect the first 20 parts to be scrap.
Does an absolute encoder remove the need to re-reference?
It removes the need only if the battery held. Check the position against a tape measure before trusting the display, because a corrupted multi-turn count reads plausibly.
Why are parts 0.4 mm short after a restart?
That is a lost motor revolution or a gear-ratio mismatch, not mechanical slip. Re-check the electronic gear ratio against the roll diameter before touching the mechanics.
Can I jog the feeder with the die closed?
No. Clear the strip first. Jogging with a part gripped shears pilots and marks the die, and a 0.1 mm burr will affect the next 500 parts.
How many strokes does a proper restart verification need?
Fifty strokes at production speed with the strip measured. Ten strokes at 30 SPM proves the axis moves, not that it holds ±0.05 mm at 200 SPM.
Where to go next
A restart is a small version of the accuracy question that runs the whole line. The budget that assigns error to each station, and the test that proves it, is in our coil line knowledge base.
The Coil Line Guide Library
