C&I storage manufacturer
Specifying coverage on a new line against a customer audit that wants one criterion per station and a record per pack.
BLP-EOL · end-of-line testing, supplied twice
Battery Pack EOL Testing Equipment, Seven Gates
BLP-EOL is the end-of-line tester on our module and pack lines, fitted twice on a full line: once on the finished module at station 16, once on the closed pack at station 14. Between line-in and pack line-off a cell passes seven test gates, each with a written criterion and a filed result. This page covers the four measurements the tester takes, where the seven gates sit, what each judges, what happens to a part that fails one, and the records that come back to you. Send your cell specification and pack configuration; a draft criteria table for all seven gates comes back within 24 hours.
- C&I storage manufacturers
- Grid-scale integrators
- EV pack plants
DC 1000V
Insulation resistance
Applied between the live circuit and the enclosure and ground path. Passes at 500 MΩ or more, read as a number rather than a lamp.
DC 2800V
Hipot leakage
Passes at a leakage current of 10 mA or less. The same instrument thresholds at module station 16 and at pack station 14.
- ×2Tester positions on one line
- 5 mVCell readback window at gate 3
The tester and what comes with it
The BLP-EOL Tester: Four Measurements at Two Positions
BLP-EOL takes four measurements at each of its two positions: insulation resistance, hipot leakage, total voltage and resistance, and continuity. Insulation is applied at DC 1000 V and passes at 500 MΩ or more. Hipot is applied at DC 2800 V and passes at a leakage current of 10 mA or less. Those thresholds belong to the instrument and are the same at module station 16 and at pack station 14. What changes between the two positions is the circuit under test, which is why battery pack EOL testing on these lines comes with two criteria tables rather than one.
Four things come with the tester rather than being claimed around it
- A written criterion at every one of the seven gatesNot only at the two end-of-line positions. Each gate is judged against a number that is signed before dispatch.
- Every result filed against a cell, module or pack numberReported to your MES over PROFINET, so a verdict is a record rather than a lamp that went green.
- A failed part held under its number, with value and criterion beside itThe release decision belongs to a named engineer, not to a button on the station.
- Equipment-caused yield signed at 99% or better at site acceptanceHuman factors excluded between both parties in writing, so the figure means what it says.
One instrument, two circuits
The tester does not change between station 16 and station 14. The object under its probes does, and that is what the rest of this page is about.
Built to ISO 9001, CE and RoHS; powered in its ordered configuration before it ships. The company behind the tester is on the about page.
- 26engineers8 of them in field installation and commissioning
- 6,000m²Hall with three pre-assembly bays
- 40+segmentsDelivered into 12 countries
- 48hWear parts dispatched, remote diagnostics connected
Who arrives, with which questions
Three Buyers, Six Questions
The engineer reading this is specifying or rebuilding test coverage at one of three kinds of plant, and arrives with six questions. Each has a section below.
Grid-scale integrator
Rebuilding coverage on a running line, where the tester sits between stations that exist and reports into an MES that exists.
EV pack plant
Adding a second end-of-line position, because the pack ships with copper bars, connectors and a harness the module test never saw.
Four measurements, one automatic sequence
Four Measurements at One Station
The four run as one automatic sequence with the part on its fixture, on both the BLP-P8S and the BLP-P15 tier. Two of them ask different questions about the same insulation barrier, which is why both run rather than one standing in for the other.
| Measurement | Applied | Passes at | What it sees first |
|---|---|---|---|
| Insulation resistance | DC 1000 V≥500 MΩ | Resistance between the live circuit and the enclosure and ground path | A pinched harness, a displaced insulation sheet, an end plate creeping into a busbar, moisture or swarf left in a joint |
| Hipot leakage | DC 2800 V≤10 mA | Leakage current with the barrier stressed well above working voltage | A barrier that is comfortable at working voltage and marginal above it |
| Total voltage and resistance | String measured | Within what the configuration and the sorting record predict | A reversed cell, a cell from the wrong grading band, a cold weld carrying more resistance than it should |
| Continuity | Every intended path | Present and carrying current, from terminals through busbars and copper bars to the connectors | An interlock or low-voltage path that does not behave as drawn |
Insulation is read as a number rather than as a lamp, which is what makes it usable as evidence later. A cold weld moves the string figure before it moves anything else.
Seven gates between line-in and pack line-off
Seven Battery Pack EOL Testing Gates by Station
A cell that enters at module station 01 and leaves inside a closed pack at pack station 15 passes seven gates. Two are end-of-line tests. The other five are electrical and pneumatic checks placed where a fault is still cheap to correct, and each has a number attached to it.
| Gate | Station | What it measures | Criterion | Where the result goes |
|---|---|---|---|---|
| Gate 1 | Module 02BLP-CS | Cell scan, open-circuit voltage and resistance, grading into bands | Scan success 99.9% or better with incoming defects excluded; voltage read to ±(0.01% of reading + 3 digits), resistance to ±(0.5% of reading + 10 digits) | Filed against the cell number, carried forward to gate 3 |
| Gate 2 | Module 10 | Polarity of every cell against the orientation record written at tray unloading; insulation after banding and before welding | DC 500 V, 100 MΩ or more; polarity checked cell by cell | Filed by module number to your MES |
| Gate 3 | Module 15 | CCS-to-BMS communication handshake, per-cell voltage readback | Readback within 5 mV of that cell's sorting record | Filed by module number to your MES |
| Gate 4 | Module 16BLP-EOL | Insulation, hipot, total voltage and resistance, continuity on the finished module | Insulation DC 1000 V at 500 MΩ or more; hipot DC 2800 V at 10 mA or less | Full item list filed by module number |
| Gate 5 | Pack 02 | Cooling plate leak test, before the plate is bonded to the enclosure floor | 200 kPa dry air, 60 s hold, pressure drop 1 kPa or less | Filed against the enclosure |
| Gate 6 | Pack 13 | Whole-pack leak test, after lid, tightening and front panel | 5 kPa, 60 s hold, drop 0.3 kPa or less, against an IP67 target | Filed by pack number |
| Gate 7 | Pack 14BLP-EOL | Insulation, hipot, total voltage and resistance, continuity on the closed pack | The same instrument thresholds, measured across the whole pack circuit | Filed by pack number, into the acceptance document |
Seven automatic positions on both tiers
All seven are automatic positions on both tiers. Gate 3 sits at the manual inspection and communication station, where the readback runs automatically while a person completes the visual check. Where these gates fall in the full sequences is on the module assembly line page, which owns the 17 module stations, and the pack assembly line page, which owns the 15 pack stations. Gate 1 has its own page, the cell testing and sorting station.
Seven written criteria, not two
Every gate has a number and a filing destination. The two end-of-line positions are the ones most specifications write down; the five checks between them are where a fault is still a rework item rather than a scrapped pack.
- Gates 1 to 4 run on the module segment: one scan, one polarity and insulation check, one readback, one end-of-line test.
- Gates 5 to 7 run on the pack segment: two leak tests forty times apart in pressure, then the pack end-of-line test.
- Gate 1 rejects and refills automatically; gates 2 to 7 hold the part under its number.
- Lid tightening at pack station 11 is a release condition for gates 6 and 7, not a gate of its own.
Same tester, two objects
Module EOL and Pack EOL
Both positions use the same tester and the same thresholds, and they test two different objects. That difference is why the second one earns its station time.
Gate 4 · module station 16Object: a finished module
The circuit under test is that module and nothing else
- a13 cells in series on the blueprint configuration
- bBusbars welded
- cEnd plates and banding on
- dCCS harness fitted
- eThe whole thing on its fixture
Same instrument
DC 1000 V≥500 MΩ
DC 2800 V≤10 mA
Gate 7 · pack station 14Object: a closed pack
Six classes of part are inside the circuit that did not exist at gate 4
- 1Series copper bars tying four blueprint modules into a 1P52S pack
- 2The BMS, mounted and wired
- 3The service disconnect, in the string
- 4High and low voltage connectors, fitted
- 5The harness, routed against the enclosure
- 6The lid, on and tightened: the enclosure is now part of the ground path being measured against
Why hipot at DC 2800 V runs a second time
Every one of those six is a new joint, a new clearance and a new route to the enclosure wall. The second run is the first time the copper bars, the disconnect, the connectors and the harness are stressed at all, and the first time the barrier measured is the barrier the pack ships with. The string voltage is higher at gate 7 as well, since the pack is the modules in series.
Two end-of-line criteria tables, both written out
One for the module position and one for the pack position, each with its item list, its thresholds and its station number.
≤60 VPacks at or below
On packs at 60 V or below the DC 2800 V hipot is dropped and a DC 500 V insulation test runs in its place. Consumer and light mobility packs sit in a different voltage class as an industry classification: a power tool pack is typically 36 V from 10 cells in series, against the 400 to 800 V a storage string works at. The withstand test is swapped because a withstand test set for a storage string stresses a barrier that will never see that voltage. Those lines are on the cylindrical and consumer pack line page.
Gate 3 · module station 15
Cell Readback at Station 15
Every cell voltage is read back through the sampling channels and compared against the record gate 1 wrote for that same cell. Two wires swapped show as two cells that no longer match.
A sampling wire on the wrong terminal
Shows as one cell whose readback does not match its record.
A cell substituted after sorting, or a wrong tray
A module built from a tray other than the one the record names shows as a spread of cells that do not match.
A channel open or crossed inside the CCS
Returns nothing, or returns a neighbour's value.
The line checks itself against its own first measurement
At module station 15 the CCS sampling harness is connected and a communication handshake is run with the BMS, so the sampling channels are proven to be talking before the module reaches end-of-line test. In the same pass, every cell voltage is read back through those channels and compared against the sorting record gate 1 wrote for that same cell. Three separate faults fall out of one measurement.
5mV
The comparison passes within 5 mV of that cell's sorting record. Tighter than any fault the harness can produce, wider than the instrument's own reading.
- Cell numbers are captured at gate 1 and travel with the sorting record.
- The fixture carries an RFID tag holding the module number.
- Both report to your MES over PROFINET, so the record follows the part.
Found at station 15, not under a closed lid
A mis-wired sampling harness is found here rather than during commissioning, when it sits under a closed lid behind a tightened front panel.
Three things written as one, settled separately
Instrument Accuracy, Threshold and Verdict
Test specifications tend to write these three as though they were one figure. In battery pack EOL testing they have three different owners, and they are settled separately here.
Belongs to the instrument
Accuracy
- Gate 1 reads voltage to ±(0.01% of reading + 3 digits) and resistance to ±(0.5% of reading + 10 digits).
- A fixed property of the equipment. It does not move between projects.
- It sets the narrowest grading band you can usefully ask for.
Belongs to the project
Threshold
- The value a part is judged against: insulation at 500 MΩ, hipot leakage at 10 mA, a plate leak drop of 1 kPa, a pack leak drop of 0.3 kPa, a readback window of 5 mV.
- These are the criteria we bring, confirmed station by station in writing before dispatch.
- Leak criteria can be adjusted before dispatch against your enclosure and project, in writing, because the volume being pressurised is yours.
Belongs to the recipe
Verdict
- The rule applied to the measured value: which criteria revision is loaded, whether the reading is taken at the end of the hold or as a trend.
- What the station does with a part that lands on the line.
- Stored in the recipe with the thresholds, under your part number.
- 500MΩInsulation, DC 1000 V
- 10mAHipot leakage, DC 2800 V
- 1kPaPlate leak drop, 200 kPa
- 0.3kPaPack leak drop, 5 kPa
- 5mVReadback window
Operators select a recipe and cannot edit one
Verdict rules are stored in the recipe with the thresholds, under your part number. The process engineer who wrote the parameters changes it, and quality countersigns before the change goes live.
A criteria revision is therefore a signed event, not a setting somebody remembers changing.
Gates 5 and 6
Two Leak Gates, Two Pressures
The two leak tests are forty times apart in pressure because they test different things: a closed metal circuit that will carry coolant under pump pressure for the life of the pack, and a large sealed volume closed by a gasketed lid.
Gate 5 · pack station 02Cooling plate circuit
- 200 kPaDry air
- 60 sHold
- ≤1 kPaDrop to pass
Before the plate is bonded to the enclosure floor
The plate is tested before bonding, because after bonding a leaking plate is not a rework item. The result is filed against the enclosure the plate goes into.
Gate 6 · pack station 13Whole pack, IP67 target
- 5 kPaEnclosure
- 60 sHold
- ≤0.3 kPaDrop to pass
After the lid, the tightening and the front panel
The pack is tested after the lid and the front panel, because the seal being judged is the seal the pack ships with. The result is filed by pack number.
×40
200 kPa against 5 kPa. The plate is judged on a 1 kPa drop, the enclosure on 0.3 kPa, and neither figure stands in for the other. Both are adjustable before dispatch against your enclosure, in writing.
- Plate first, before bonding. A leaking plate found after bonding is a scrapped enclosure floor, not a rework.
- Pack last, after lid and panel. The gasket under test is the one that ships.
Lid tightening at pack station 11 is a release condition, not a test
Twin-spindle servo tightening runs at 8 to 25 N·m against the recipe stored for your enclosure, and torque and angle are filed by pack number. A pack whose tightening record is incomplete or out of window does not move to the leak station, because a low-torque fastener produces a leak result that says nothing about the gasket.
- Tightening evidence
- Leak result
- End-of-line test
Five things start the criteria table
Your Cell, Configuration, Enclosure and Takt Are Enough to Draft All Seven Gates
Within 24 hours you get back a draft criteria table for all seven gates written against your configuration, the item list for both end-of-line positions, and a note on any gate where your enclosure or voltage class changes it.
Cell specification with voltage and resistance data
Module and pack configuration
Enclosure and cooling plate drawings with sealed volumes
Target takt
The criteria table you use today, if you have one
A failed part is held, not removed
Held Parts, Values and Owners
- Gate 1: rejected and refilled automaticallyA failing cell is rejected and the empty position in the group is refilled automatically, so the station keeps its takt and the module downstream is still built from a complete, in-band set.
- Gates 2 to 7: held under its numberThe part stops under its module or pack number, with the measured value and the criterion recorded beside each other and reported to your MES.
The station has no pass button
Operators are trained to select recipes, release parts the station has passed and route held parts; they do not carry the parameters and cannot edit a threshold to make a part fit.
Operator
Selects the recipe, releases what the station passed, routes what it held.Process engineer
Set the parameter and wrote the criterion. The release decision on a held part is theirs.Quality
Countersigned the criterion, and countersigns any change to it.
That follows from how criteria are written: the person judging a held part is the person who set the number it was judged against.
Held parts · end of shiftexample · from your MES
M-2417IR 412 MΩcrit ≥500 MΩHELD
P-0088drop 0.42 kPacrit ≤0.3 kPaHELD
M-2431cell 07 +7 mVcrit ≤5 mVHELD
M-2436cell 09 reversedcrit cell by cellHELD
What that is worth shows up at the end of a shift: a list of held parts with values on it, against part numbers, in your MES. The pattern behind them is then visible: one station, one fixture, one tray of cells, one enclosure batch.
- ConcessionRecorded as one decision, under the original part number.
- ReworkRecorded as a second, separate decision.
- RetestRecorded as a third. Never folded into the other two.
Tagged to a number, not to a time stamp
Battery Pack EOL Testing Records from Each Gate
Cell numbers come from the scan at gate 1, module numbers travel on the fixture RFID, pack numbers carry through the pack segment. A cell is traced from its sorting record, through the readback at gate 3 and the module test at gate 4, out through the pack test at gate 7.
Process records filed the same way, read alongside
- Pressing force curve and post-press length, by module number
- Welding recipe and its identifier, by module number
- Gluing pressure curve and dispensed volume, by enclosure
- Tightening torque and angle, by pack number
A held part at gate 7 is investigated with all of that in front of you rather than with a single verdict line.
Documents you receive
- Factory acceptance report, measured value and criterion side by side for every station
- Site acceptance report, the same layout on your floor
- Station table, marked up for your tier
- End-of-line item list, for both tester positions
- Recipe list, thresholds and verdict rules under your part numbers
The last two matter after handover
The end-of-line item list and the recipe list are what your quality team maintains the line against once our engineers have left.
Defined, not described
Yield in the Acceptance Criteria
- FPY = (input − NG) ÷ inputFirst-pass yield: input quantity minus NG parts, divided by input quantity.
- Utilisation = TA × PA × FPYTime availability × performance availability × first-pass yield. Both definitions go into the acceptance criteria, so the signed figure has arithmetic behind it.
- 8hContinuous run at target takt
- ≥99%Equipment-caused yield, signed
- 20 + 5Modules and packs through the full item list at FAT
Human factors are excluded in writing. A cell batch outside band or an operator error is a real loss no insulation tester can fix, so both are separated when the criteria are written, not argued on the day. Annual output from the same figure is on the capacity planning page.
First of the eight acceptance steps
Criteria Signed Before Dispatch
The criteria for all seven gates are written station by station and signed before anything ships, as the first of the eight acceptance steps. They are written by the process engineer who set the parameters and countersigned by quality. The eight steps and the documents that come out are on the FAT and SAT acceptance page.
Factory acceptanceIn our hall, on your materials
- 2–3Working days
- 20Modules, full item list
- 5Packs, full item list
In a pre-assembly bay on your cells, busbars, end plates, enclosures and cooling plates, so the end-of-line criteria are exercised on your own materials before the line leaves our hall.
Site acceptanceOn your floor, after commissioning
- 3–5Working days
- 8 hContinuous run
- ≥1000 NWelded sample pull
The 8-hour continuous run at target takt, welded sample pull tests at 1000 N or more, and every end-of-line item measured against the signed criteria.
Open items, line by line
- Recorded with the station, the measured value and the criterion beside each other.
- The engineer who owns that criterion closes it.
- Only the open items are retested, not the whole run, scheduled inside 10 working days.
- If the same item fails twice in succession, both parties settle it in writing against the criteria.
Training over 5 working days, three roles apart
- OperatorSelects recipes and routes held parts.
- Process engineerWrites the recipes, reads the curves.
- MaintenanceSpares and remote diagnostics access.
Behind the signed table
26 engineers8 in field installation and commissioning
6,000 m² hallThree pre-assembly bays
40+ line segmentsDelivered into 12 countries
48 h wear partsRemote diagnostics connected · ISO 9001 / CE / RoHS
Two ways to start
Talk to a Test Engineer
Adding end-of-line testing to a line you already run
Send the station sequence you have now and the point where the tester should sit. The reply covers what it needs from the stations before it:
- Part numbering, so a result has a number to be filed against.
- Fixture location, so the probes land on the same posts every cycle.
- The interface to your MES, so the record leaves the station.
See the tests run before you commit
Give two weeks' notice with your drawings, configuration and target takt, and a pre-assembly bay is set up in your configuration for the visit. Where travel is not possible, the same criteria table is run as a video acceptance.
- Two weeks' notice with drawings, configuration and takt.
- Your configuration in one of the three bays.
- Video acceptance on the same criteria table where travel is not possible.
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.
Send the pack configuration
Battery Pack EOL Testing Enquiry
Send five things: your cell specification with its voltage and resistance data, your module and pack configuration, your enclosure and cooling plate drawings with their sealed volumes, your target takt, and the criteria table you use today if you have one. The reply comes within 24 hours. You can also start from the contact page.
- A draft criteria table for all seven gates, written against your configuration
- The item list for both end-of-line positions
- A note on any gate your enclosure or voltage class changes
- Station numbers for both tester positions in your sequence
- What changes if your pack voltage class changes later
Email
Phone / WhatsApp
Address
Wuhan, Hubei, China
Response
Within 24 hours