STATION 01 · BLP-ULAuto
Tray unloading
Five cells per pick, appearance and terminal orientation checked by vision as each cell leaves the tray. Empty trays stacked automatically.
BLP-UL · BLP-CS — stations 01 to 03 of seventeen
Cell Testing & Sorting Station: Every Cell Graded Before a Module Exists
This cell testing & sorting station is BLP-UL and BLP-CS together: a six-axis robot lifts prismatic cells off the foam trays five at a time, and every cell is scanned, measured and graded on OCV and internal resistance into up to 10 bands. A cell outside your rule is removed and its place refilled, same band. Send the cell drawing and your grading rule; the station scope comes back within 24 hours.
QE Quality engineers, energy-storage pack plants
PE Process engineers, EV module plants
R Retrofit into a running segment
99.9% or more
Scan success, incoming defects excluded
Whether every cell entering the segment carries an identity the line can record and act on.
±0.06mm
Robot positioning repeatability
How precisely a cell is put down, so the next station takes the placed position as given.
What the pair is
Cell Testing & Sorting Station Pair
BLP-UL and BLP-CS are stations one to three of the seventeen in the module segment. The cell testing & sorting station is where an interchangeable cell becomes a member of a specific, graded set.
Pick, scan, measure, grade, refill
- The robot picks five cells at a time off the foam tray, checking appearance and terminal orientation with vision as it picks.
- Every cell is scanned for its barcode, measured for open-circuit voltage and internal resistance, and assigned to a band.
- Cells outside your rule are removed automatically; the place is refilled from cells in the same band, so what moves forward is a complete set rather than a short one.
- All three stations run fully automatic on the economy configuration and on the high-automation configuration alike.
Entry to a module lineStations 01–03 of a battery module assembly line.
Stand-alone stationThe pair on its own, feeding what you already run downstream.
Retrofit into a segmentCut into a running segment where grading is still off-line.
Within 24 hoursSend the cell drawing, how cells arrive on their trays, and your grading rule if you have one. The station scope for your cell comes back inside a working day.
Stations one to three
Three Automatic Stations, One Set
On both configurations all three are labelled AUTO, and the whole cell-handling segment is staffed by four people. Every station after them works on a set whose count and grading were settled here.
STATION 02 · BLP-CSAuto
Scan, test, grade
Barcode read, OCV and internal resistance measured on the same instrument, band assigned on both quantities together.
STATION 03 · BLP-CSAuto
Reject and refill
A cell outside the rule is removed; its place is refilled from a 60-cell in-line buffer, same band as the set.
Downstream, nobody repeats the measurement
- 04Flip and group
- 05Plasma-clean large faces
- 06–07Tape, end plates, insulation
- 08Stack and press to length
- 15Read every cell voltage back
Stacking cannot start a module and then wait for a cell. Pressing at the module stacking and pressing station works to a length that assumes the count is right.
Datasheet
Cell Testing & Sorting Station Datasheet
Twelve published figures for the cell testing & sorting station, and what each one governs. Three describe measurement, and they are the ones most often read as quality promises. They are not. The rest describe reach, precision, and what the station does with what it measures.
| Figure | What it governs |
|---|---|
| BLP-UL tray unloading robot | |
| Robot payload220 kg | Headroom at the wrist for the gripper plus five prismatic cells, so the pick rate holds as the cell gets heavier. |
| Reachat least 2650 mm | The horizontal working radius served from one base: how many tray, buffer and place points are reachable without a transfer. |
| Positioning repeatability±0.06 mm | How precisely a cell is put down, which lets the next station take the placed position as given rather than re-find it. |
| Cells per pick5 | The pick that sets the feeding rate, and with it the counting point for line takt. |
| Vision at pickappearance + terminal orientation | Every cell checked as it is lifted; a wrongly oriented cell does not enter sorting. |
| Empty tray stackingautomatic, alarm at 20 layers | What the station does between one pallet and the next, and so what a person does during a shift. |
| BLP-CS scan, test, sort | |
| Scan successat least 99.9% | Incoming defects excluded. Whether every cell entering the segment carries an identity the line can record and act on. |
| Voltage accuracy±(0.01%rdg + 3dgt) | How finely two cells are told apart on OCV, which sets the narrowest voltage band you can ask for. |
| Internal resistance accuracy±(0.5%rdg + 10dgt) | The same for resistance, which separates cells that look identical on voltage. |
| BandsOCV and resistance, up to 10 | How many groups the graded stream splits into on the two measured quantities together. |
| Rejected cellsremoved, place refilledStation 03 | Refilled from a 60-cell in-line buffer, same band. The set moves on complete. |
| Automation levelAUTOBoth configurations | AUTO on the economy configuration and on the high-automation configuration. |
Not a verdictNothing on this station issues a verdict on a cell's health. It compares, grades, and records. Pass and fail are decided at end-of-line testing.
Your sorting criteria document
Your Sorting Criteria Document
The band widths are not published here, because they are not ours to choose. Band width and NG condition are written into a sorting criteria document for your project and signed by both parties before the station ships.
It is written by the engineer who sets the parameters. Process defines, quality signs, and the station is set to it. It is signed alongside the acceptance criteria rather than settled during commissioning.
SOURCE 1Your cell supplier's gradingThe bands the supplier already sorted to, carried into the line unchanged.
SOURCE 2Your pack's balancing budgetThe spread the BMS can live with, written backwards into band width.
SOURCE 3A customer specificationA rule you already have to meet, signed here so the station is set to it.
Sorting criteria documentPROJECT · CELL TYPE · REV
- Quantities
- OCV and internal resistance, together
- Bands
- Up to 10 B1–B10
- Band width
- Written per project, not published
- NG condition
- Written per project; cell removed, place refilled same band
- Rest time
- Belongs to the incoming state, not to a queue inside the station
- Refill
- Same band, written against the cell number that filled the place
- Signed with
- The acceptance criteria, before the line ships
Process engineeringdefines the parameters
Qualitysigns the document
One thing this station does not do: rest cells
OCV is measured as the cell presents itself. Whatever rest time your cells need after the supplier's process belongs to the incoming state, not to a queue inside the station. A sorting station asked to absorb rest time silently is one running behind takt.
No written rule yet? The station runs to the resolution of its instruments and records every cell, so the rule can be written from your own first production data.
0 minrest time inside the station
1 shiftnew cell type, tooling + signed criteria
≤30 minswitch between qualified types
10 bandsmaximum, on OCV and resistance together
Accuracy, band width, scan rate
Accuracy, Band Width, Scan Rate
Instrument accuracy is written as a pair: a proportion of the reading plus a number of digits. A band narrower than the instrument can resolve is not applied, it is approximated, and boundary cells land there by noise.
±(0.01%rdg + 3dgt)
Voltage accuracy
- Means
- One part in ten thousand of the measured value, plus three counts on the display
- Settles
- The narrowest OCV band you can meaningfully ask for
A rule finer than this is flagged before quoting, not after the line is laid out.
±(0.5%rdg + 10dgt)
Internal resistance accuracy
- Means
- The same shape with looser terms: resistance is harder to measure at the current a sorting station can apply in the seconds it has
- Settles
- Which cells identical on voltage are told apart
The second axis of the grid: ten bands run on both quantities together.
99.9% or more
Scan success
- Means
- Nearly every cell carries an identity, incoming defects excluded
- Settles
- Something different: whether a cell can be recorded and acted on at all
A cell can scan perfectly and still sit outside your band.
Compares, grades, recordsPass and fail on the module and the pack are decided at end-of-line testing, against thresholds published there. Nothing here is a health verdict.
A rule finer than the instrument
Usually copied from a supplier datasheet. The answer comes before the quotation: widen the band to what the instrument resolves, or accept that boundary cells are placed by noise.
The gap a reject leaves
Reject One, Refill One
Rejecting a cell is the easy half. A module is a fixed count, thirteen in the 1P13S blueprint, eight to twenty-four across the stacking range. Pull one cell out and the set is one short, and short is not a state any station downstream can work with.
60 cells, same bandStation 03 removes the cell and refills the place from an in-line buffer of cells already scanned and measured, taking one from the same band the set is being built to.
The three alternatives are why this is a mechanism, not a detail
| What happens to the gap | What it does to the segment |
|---|---|
| Hold the set until a match arrives | Stops the segment. One reject becomes idle time at every station behind it. |
| Let the set go short | Builds a module to a count nobody approved. Pressing works to a length that assumes the count is right. |
| Fill the place by hand | Puts in a cell that never went through the instrument. Invisible here, visible three stations later on a welded module. |
| Refill from the 60-cell buffer, same band | The set moves on complete and graded, the refill written against the cell number that filled it. Nobody stops the line. |
The buffer size is what makes it work at takt: sixty graded cells in front of the removal point cover a run of rejects without the segment waiting for the next tray to supply a match.
Every set that reaches stacking has the right number of cells, every cell in it was measured on the same instrument to the same signed rule, and each refill is on record.
One cell, one number
One Cell, One Number
Up to this station a cell is interchangeable. After it, the cell is a member of a specific set, and everything downstream treats it that way.
What makes that survivable a year later: the identity was read here and the measurements written against it here, not reconstructed from a batch label.
CELL RECORD · STATION 02/03TO MES, AS THE CELL RUNS
- BARCODEread at station 02
- OCVmeasured, ±(0.01%rdg + 3dgt)
- INTERNAL RESISTANCEmeasured, ±(0.5%rdg + 10dgt)
- BANDB1 … B10, per signed rule
- REFILL0 / 1, written against the filling cell
FIVE FIELDS · NOT COMPILED AT THE END OF A SHIFT
Send the cell drawing
Send the Cell Drawing and the Rule
No written rule? Say so. Running to instrument resolution and writing the rule from your own data is a legitimate route, and it is quoted that way. Write from the contact page; the reply comes inside 24 hours.
Cell drawing or datasheet
How cells arrive on their trays
Grading rule, if written
The record station fifteen reads
The Record Station Fifteen Reads
Five fields leave this station against every cell number: the barcode, the open-circuit voltage, the internal resistance, the band, and whether it was a refill. They go to your MES as the cells run, not compiled at the end of a shift.
Written hereSTATION 02 · PER CELLOCV at grading, same instrument for every cell
≤5 mVcriterion
Read backSTATION 15 · CCS HARNESSEach cell's voltage through the finished module's own sampling lines
At station 15 the CCS harness is handshaked with the BMS and each cell's voltage is read through the module's own sampling lines, then compared with what this station wrote. The criterion is 5 mV or less, signed with the acceptance criteria before the line ships.
A module whose read-back disagrees is telling you one of three things
- 1A sampling line is on the wrong cellCaught while the module can still be taken apart.
- 2A cell was substituted between grading and stackingInvisible without a per-cell record from the front of the segment.
- 3The grading measurement has driftedSeen as a pattern across modules, not as a single miss.
It is also where a pack query starts a year later: the cells that went into it, the bands they were graded into, the refills marked.
The counting point for takt
The Counting Point for Takt
Across this site, PPM is counted at the cell testing & sorting station and means one thing: cells per minute entering the module segment. Not packs per minute, not modules per minute. The piece counted is a single cell, at BLP-UL.
The meter for the whole line
Site acceptance run, counted at BLP-UL8 hours continuous
0 h2 h4 h6 h8 h · 3,840 cells at 8 PPM
8PPMThe blueprint P8 rate: eight cells into BLP-CS every minute.
3,840cellsThrough this station in one eight-hour site acceptance run at 8 PPM.
≥99% yieldRequired for causes attributable to the equipment, human factors excluded in writing beforehand.
Quoted rate = what BLP-UL delivers into BLP-CS
- Every station after this one is sized to keep up with the rate counted here.
- The three configurations and the annual energy they work out to are on the capacity planning page, with the arithmetic.
Testable, not demonstratedA takt figure demonstrated in short bursts is not the same claim as one measured at a defined point over a shift.
Terminal orientation and empty trays
Terminal Orientation and Empty Trays
Vision at the pick checks two things on every cell: appearance, and which way the terminals face. Emptied foam trays are stacked automatically, alarm at 20 layers. Neither reads like a headline; both decide how the segment is manned.
Orientation, checked before sorting
A cell the wrong way round does not go into sorting. Orientation survives everything geometric downstream, and nothing objects until the polarity test at station 10, by which point a module has been built around it.
- Appearance and terminal orientation on every cell, at the pick
- Polarity re-checked at station 10 against the record made here
Empty trays, stacked to twenty layers
Trays are stacked as the pallet is worked through; the stack alarms at 20 layers so it is taken away between pallets, not in the middle of one. A station is manned by its least automatic moment.
- Four people cover the whole cell-handling segment on both configurations
- Automatic tray stacking is part of why that number is what it is
AutoSemi-autoManual
Labelled per station. These three are genuinely automatic: a smaller claim than an unmanned line, and a true one.
Cell envelope and changeover
Cell Envelope and Changeover
The cells this station handles are prismatic, within an envelope of 100–220 mm wide, 100–220 mm high and 20–80 mm thick. Inside that envelope, what changes with your cell is the gripper and the tray program, not the station.
≤30 minChanging to a cell type already qualified here: a gripper and tray program switch.
1 shiftBringing in a type nobody has run: gripper tooling, criteria document signed, first article qualified, then production runs against it.
What changes with your cell
- Gripper fingers, cut to the cell's width and thickness
- Tray program, written to how your cells arrive
- Sorting criteria document, signed before production runs
Production cannot run against a rule that has not been signed, and the signature is what makes the next changeover a 30-minute program call. Range and stops for the whole line are on the cell compatibility and changeover page.
Cylindrical cells are a different route, with their own sorting machine and their own downstream sequence, on the cylindrical and consumer pack line page rather than fitted onto this station.
The records you receive
Cell Testing & Sorting Station Acceptance
Factory acceptance runs in the 6,000 m² assembly workshop, on one of three bays where a line is stood up in its shipping configuration, using your cells. Site acceptance repeats it on your floor, counted at this station.
FATFactory acceptance
4 hcontinuous
20 + 5modules + packs
260cells + ~10%
Two to three working days, on a material list both parties confirm before shipping. For the 1P13S blueprint: 260 cells plus around 10% spare, through the full test set.
SATSite acceptance
8 hat target takt
3,840cells at 8 PPM
10 dretest window
Eight hours continuous at your target takt, counted here, over three to five working days. Missed items are recorded with station, measured value and criterion; only those are retested.
What you keep
- Factory and site acceptance reports, measured values beside criteria, station by station
- An annotated station table for all seventeen module stations
- The end-of-line test item list, with thresholds
- Your signed sorting criteria document
The eight steps are on the factory and site acceptance page; who does the work, on the about page.
Training, five working days, split by role
OperatorsSelect a setting, handle detained cells
Process engineersWrite the criteria, read the records
MaintenanceHold the spares and the remote diagnostic link
What to send
Four Inputs for This Station
Four things specify a cell testing & sorting station. With these four, the station scope, the gripper tooling your cell needs, and what your sorting criteria document would contain come back within 24 hours.
1
Cell drawing or datasheetWidth, height, thickness, terminal positions.
2
How cells arrive on their traysTray format, cells per tray, orientation as delivered.
3
Your grading rule, if one existsBand widths and NG condition, or a note that none is written yet.
4
Target taktCells per minute into the segment, or the annual energy you are sizing for.
See the line first
Two weeks' notice plus the cell drawing, the module and pack configuration and the target takt, and a bay is set to your configuration for the visit.
Where a visit is not possible, the same criteria run on video against the same signed document.
Your enquiry, ready to send
If your email client opened, it is already in the message. If nothing opened, use one of the two routes below.
Start with the cell
Send Your Cell and Your Tray
Four inputs to the engineering team and the station scope, the gripper tooling and the outline of your sorting criteria document come back within 24 hours. Behind it: the people and the bays on the about page.
Email
Phone / WhatsApp
Address
Wuhan, Hubei, China
Response
Within 24 hours