Why Your Battery Warranty Claim Became Customer Abuse

A warranty claim on an industrial lithium-ion pack is usually lost at the receiving dock, months before anyone files it. The document that decides which side of the line a returned pack lands on is the incoming inspection record made at the moment of receipt: crate condition, shock and tilt indicators, photographs of terminals and connectors, per-pack open-circuit voltage, stated state of charge, storage location, ambient temperature, and the name of whoever did the check, all timestamped. With that record in hand, a supplier has to argue with your data. Without it, a supplier only has to point at the empty space, and the pack gets reclassified as customer-induced.

I have watched that reclassification happen with the pack still sitting on the table. An AMR pack came back from a refrigerated distribution center after eleven months of three-shift duty. The denial form had three checkboxes on it, the way they usually do. Box two was ticked: over-discharge in storage. Attached was a photo of a terminal with a green bloom on it and a service report that put the lowest cell group well under the low-voltage cutoff. Nobody at the supplier had touched the pack before writing that. They didn't need to. They had a number from the return bench and nothing from my side to put next to it.

What I noticed first, honestly, was the crate. The foam was gone. Not damaged, gone, like it had been repacked by someone in a hurry and the original inserts got left on a floor somewhere. I stood there holding the lid and thinking that this was the entire argument, sitting in my hands, and it was going to be worth nothing because there was no photo of the crate sealed on the day it arrived eight months earlier.

The form has three boxes, and none of them say "we think you broke it"

The word that does the work is not "abuse." It is "induced." That framing puts the burden on the physical condition of the pack, and the physical condition of a pack is something that changes continuously from the moment it leaves the supplier's loading dock. Every hour it exists, the evidence gets worse. That is the whole asymmetry, and it is structural, not malicious.

The three grounds you will see over and over are transit and packaging damage, over-discharge during storage, and charging outside the specified voltage window. There is a fourth that shows up on water-related returns, and a fifth and sixth for opened housings and altered firmware. Once you know that list, you know exactly what your receiving file has to answer.

Transit damage is the easiest one to disprove, and almost nobody does

A pack that shows up with a cracked housing or a bent connector is a claim on day one, and it is a clean claim, because nothing has happened to the pack yet except shipping. But the claim only exists if you can show the damage was there before you unloaded it.

I take four photos before the crate comes off the pallet. The seal and the banding. The label with the serial and the date code. The shock and tilt indicators, if the shipper used them, because those are the supplier's own devices and their state is their own evidence. And then the crate opened, with the original foam and inserts still in place, before I pull anything out.

That last photo is the one I've used. When the foam is missing from a returned crate and there's no photo of the crate as it arrived, I have no way to separate damage that happened in transit from damage that happened in my own rack. The supplier's report says "shipping damage." My file says nothing. Box one gets ticked.

Over-discharge is the one buyers hand them

This is the ground that takes the most claims, and it is the one where the buyer does most of the supplier's work for them.

Packs ship at a partial state of charge, and the supplier knows the exact figure because they set it. Ask for the ship record by serial number: the state of charge at release and the open-circuit voltage measured before it went in the box. That converts a bare reading on the return bench into a delta, and a delta is an argument. A pack that left at a known voltage and came back at a lower one over a known interval is a conversation about time and storage, not a conversation about what I did to it.

Which brings up the part that stings. Most over-discharge in the field does not happen in the field. It happens in your own building, in the gap between receipt and installation. The pack lands on a dock in July, gets signed for, and sits in a non-climate-controlled cage for four months because the robot it was bought for is behind schedule. Nobody measures it. Nobody logs the temperature. Then someone pulls it out, finds it flat, and files a claim that reads, from the other side of the country, like a confession.

The fix is embarrassingly cheap. A shelf voltage reading on the day it goes into storage and another at a set interval, with the storage location and the ambient range written down next to them. If a pack drifts down on your shelf, you want to know that while it is still your problem and still inside the storage interval, not after it becomes an argument.

The voltage window question is a receipt question

Charging outside the specified range is the third box, and it usually surfaces as a charger question rather than a pack question. If your charger's absorption voltage or its temperature-compensated cutoff doesn't match the pack spec, the supplier will find that out from the BMS log and the claim will be reclassified fast.

What I do at receipt is put the charger model and its charge profile into the same record as the pack serial, with a note on who verified the match and when. It takes five minutes and it kills the entire line of argument before it starts. When the profile lives in a folder nobody opens, it is not a record; it's a rumor.

A pack that arrives with an unbalanced cell group is the ugly version of this. If you have no state of charge and no storage condition recorded at receipt, the supplier can call that imbalance customer-induced and there is nothing in your file to contradict it, even if the imbalance was there in the crate. The absence is the case against you.

Water is decided by a number on the datasheet

Ingress protection is an exclusion, and the IP rating on the datasheet is a warranty term whether or not it is printed in the warranty section. If the pack is rated for splashing and your sanitation crew does a full washdown on the AGV, water ingress is outside the coverage and the supplier will say so plainly.

I have seen this argued for weeks by people who never looked up the number. My habit now is to write the IP rating into the receiving record, next to the cleaning procedure for that robot, so the mismatch is visible on the day it exists rather than on the day a pack comes back wet. If the cleaning procedure doesn't fit the rating, the fix is the procedure or the spec, not the claim.

Capacity fade is a different claim with a different file

This one causes more confusion than any denial. The defect warranty covers materials and workmanship. Capacity fade is handled under a separate capacity guarantee, with its own conditions: the discharge rate, the temperature, the cutoff, the depth of discharge, how many cycles count, and who runs the measurement.

If a pack is delivering less than it should and you file it under the defect warranty, the supplier gets to say, correctly, that this isn't a workmanship defect. To your operations manager that reads as a denial. It isn't. It's the wrong file for the right problem.

The capacity guarantee is where you need to be specific at the front end, because a fade claim lives or dies on whether the test conditions in the guarantee match how you actually run the robot. A guarantee written against a gentle reference cycle means nothing for an AMR doing opportunity charging around the clock. Get the duty profile into the guarantee language before the order, and when the fade shows up, the measurement protocol is already agreed.

The two things that void the file even when you are right

Opening the housing. And flashing firmware.

Both of those are standard exclusions, and both get triggered by people with good intentions. An engineer pops the lid to check a busbar, or flashes a BMS to test a communication handshake, and the pack is now unrecoverable from a claims standpoint even if the underlying problem was there before anyone touched it. The torque seal on the housing screws is the cheapest insurance there is: photograph it at receipt, and photograph it again before anything goes back.

Firmware is worse, because a flash usually leaves a trace that the supplier's bench reads on return. If a test genuinely requires it, get the authorization in writing first, with the serial number and the tool named, and keep that email in the same file. A written authorization turns a void into an approved action.

Some denials are correct, and you should let those go quietly

Not every reclassification is a dodge. A pack that comes back full of cleaning solution when the datasheet said splash-rated is excluded, and arguing it costs you something real: the credibility you will need on the next claim, the one that actually matters.

The file is what separates the two. When I have the crate photo, the terminal voltage at receipt, the storage log, the charger profile, and the torque seal image, I can look at a denial and know within a minute whether to push or to write it off. That is not a legal determination and it isn't meant to be. It's a read on which side of the line the evidence puts the pack.

The boundary is built in ten minutes at the dock

None of this gets fixed after a denial. A record made three weeks later is a story, and both sides know it. The narrow window where you control the facts is the time between the truck pulling away and the pack going onto a shelf, and it's short.

So the last thing I'll say is the thing I wish someone had said to me early: the supplier's claims reviewer is not looking at your pack, they are looking at a form and a photo and a set of readings. Your job is to make sure the second file on that desk is as thick as theirs. Do that at receipt, every time, even when the pack looks fine, especially when it looks fine, and the claim stays on the warranty side of the line where it belongs.