How to Evaluate a Custom Lithium Battery Pack Manufacturer

You send the same RFQ to three suppliers. Two of them come back inside 48 hours with a clean quote and a lead time. The third sends you a list of questions instead: what's your continuous and peak current, what does the discharge profile look like across a real duty cycle, what enclosure is this going into, which market is it shipping to, and what's the storage temperature in the worst case you can picture.

The instinct is to like the two fast ones and find the third one annoying.

That instinct will cost you. The two quick quotes are either guesses or catalog packs with your label printed on the wrapper. The supplier asking hard questions is the one actually trying to build the pack you need. After enough programs, you learn that the questions a supplier asks you reveal more about their engineering than anything on their capability deck.

This is a guide to reading those signals before you commit. Most of what separates a real manufacturer from a reseller never shows up on the quote. It shows up eighteen months later, in the field, when a pack that sailed through someone's bench test starts dropping cells in the cold.

Manufacturer or assembler? The label on the website tells you nothing

Everybody selling battery packs calls themselves a manufacturer. The word is free. What you're actually trying to find out is who owns the engineering.

There's a real spectrum here, and the difference matters:

• A reseller buys finished packs and prints your label on them. Zero design ownership.
• An assembler buys cells and a BMS, spot-welds them together to a drawing, and ships. They control the build, not the design.
• A manufacturer designs the pack: cell selection, electrical and mechanical layout, BMS configuration or firmware, thermal strategy, and the validation plan behind all of it.

You want the third one. The gap between the second and the third is where most field failures are born. An assembler will happily build exactly what you spec, including your mistakes. A manufacturer pushes back when your spec is wrong, because they understand the physics underneath it.

So how do you tell which one you're talking to? Ask a design question they can't answer from a catalog. Why this cell over that one for my discharge rate? What happens to your protection thresholds at low temperature? How did you size the nickel for my continuous current? An assembler deflects or goes quiet. A manufacturer has opinions, and they'll walk you through the reasoning.

Start with the BMS, because that's where the engineering hides

If I get to ask a prospective supplier only one technical question, it's about the battery management system. The BMS is where competent shops separate from the rest, because it's the part you can't fake with tidy welding.

A lot of suppliers treat the BMS as a part number. They buy a generic board rated for your voltage and current, bolt it on, and call it done. Sometimes that's fine. Often it isn't, and here's where it bites: a generic BMS ships with default protection thresholds that may have nothing to do with your specific cell or your application. Overvoltage and undervoltage cutoffs, charge and discharge current limits, temperature cutoffs, the recovery behavior after a fault: all of it should be matched to the cell datasheet and the duty cycle, not left on the factory settings.

Questions worth asking, in rough order of how much they reveal:

• Do you configure the protection thresholds to my cell and application, or run defaults?
• Passive or active balancing, and why did you pick it for this pack?
• How are you estimating state of charge, and how does it behave over the pack's life?
• Can you support my comms protocol (CAN, RS485, SMBus) and expose the data I actually need?
• Do you write or modify firmware, or are you limited to whatever the off-the-shelf board does?

That state-of-charge question is a good trap, in the nicest sense. If your chemistry is LiFePO4, the discharge curve is famously flat through the middle of its range, which makes voltage-based fuel gauging unreliable: the voltage barely moves while a lot of capacity drains away. A supplier who understands LFP will talk about coulomb counting, recalibration at the endpoints, and the drift you have to design around. A supplier who doesn't will tell you the gauge "reads the voltage." That answer tells you everything you need to know. If you want the deeper version of how a pack controller should be specified, see our guide to BMS design and configuration, and our breakdown of LiFePO4 vs. NMC for OEM packs for how chemistry drives these choices.

Cell sourcing and traceability: ask where the cells come from, then ask for proof

Cells are the most expensive and most safety-critical part of the pack. They're also the easiest place to quietly cut cost, which is precisely why you should push on sourcing.

What you're screening for is anonymous cells. Grade A cells from a known manufacturer come with a real datasheet, consistent capacity and internal resistance, predictable self-discharge, and a lot number you can trace. B-grade, downgraded, or salvaged cells don't. They arrive with a wide capacity spread and inconsistent internal resistance, and that inconsistency is poison in a series string: the weakest cell limits the whole pack and drives the cells further apart every cycle. You don't see it on day one. You see it as premature capacity fade and balancing problems a year in.

Ask three things, and expect clean answers:

1. Which manufacturer makes the cell, and can I see the datasheet? A real shop names the cell maker without flinching.
2. How do you handle incoming inspection? You're listening for capacity grading and internal-resistance matching before the cells ever go into a pack, not "we trust the supplier."
3. Do you keep lot traceability from cell to finished pack? This matters enormously the day there's a field issue and you need to know which units share a suspect lot.

A supplier who grades and matches cells by capacity and IR before assembly is doing real work you'll never see on the invoice. It's one of the clearest dividing lines between a manufacturer and a box-shifter. For how form factor interacts with all of this, see our guide to 18650 vs. 21700 cell selection.

What can they test in-house? ("we test everything" doesn't mean what you think)

Every supplier says they test everything. Push on what that sentence actually covers, because there's a wide canyon between an end-of-line functional check and real design validation.

Two different things get blurred together here. Production QC is the per-unit check at the end of the line: does the pack power up, hit voltage, pass a basic functional test. Necessary, but it only proves this unit was built like the last one. Design validation is proving the design itself is sound: capacity under load, cycle life, behavior across the full temperature range, insulation resistance, plus the abuse and environmental tests that show the pack survives the real world. The two are not the same. A supplier who only does the first is quietly asking you to validate their design for them, inside your product, at your expense.

A short, blunt list of what a serious shop can do in-house or closely manage:

• Capacity and internal-resistance testing on cells and finished packs
• Cycle-life testing (the slow, boring, expensive one that nobody fakes)
• Environmental testing across the operating and storage temperature range
• Insulation resistance / hipot on the finished pack
• Functional and safety testing of the BMS, including fault injection

You don't need every one of these under one roof. You do need to know which they own, which they outsource, and whether they'll hand you the actual test reports instead of a certificate that says "tested." That last part is the real test. Ask for a sample validation report. The quality of that one document tells you how seriously a supplier takes its own process.

Compliance is more than a stack of certificates

Certificates are easy to wave around and hard to read. The skill you're evaluating isn't whether a supplier owns a folder of PDFs. It's whether they know which standards apply to your product, in your market, and can support the test reports that back them.

The stable baseline, regardless of market: lithium batteries have to clear transport safety testing under UN 38.3 before they ship by air or sea, and cell and pack safety is generally addressed by standards such as IEC 62133 (with IEC 62619 for industrial cells). Add the market layer on top. For the US, your customers may expect UL listings. For the EU, CE marking with a real test basis behind it, not a self-declaration with nothing underneath. Shipping itself brings Class 9 dangerous-goods handling and a current SDS/MSDS. A supplier who can talk fluently about which of these applies to your specific pack and category is showing you something a certificate never can.

There's a moving piece here worth raising, because it matters if you're selling into Germany or anywhere in the EU and your pack crosses a size threshold. Under the EU Battery Regulation (Regulation (EU) 2023/1542), batteries above 2 kWh placed on the EU market (which sweeps in a lot of energy storage and larger industrial packs) face phased requirements, including a carbon-footprint declaration and, from 18 February 2027, a mandatory digital battery passport accessible via QR code. The legal responsibility sits with whoever places the battery on the market, usually you, but you'll lean on your supplier for the underlying data: cell composition, sourcing, carbon footprint. Those details are still being finalized through delegated acts, so confirm the current state when it matters. The point for evaluation is simpler: a manufacturer who already sees this coming, and can talk about the data they'd need to supply, is a different animal from one who's never heard of it. Our overview of battery pack certification and compliance covers the standards landscape in more detail.

Consistency is the hard part: unit 1 versus unit 10,000

Anyone can build one good pack. Your best technician hand-builds the prototype, takes their time, and it comes out beautiful. The question that actually decides a production program is whether unit 10,000 is the same as unit 1.

This is where process control earns its keep, and it's the least glamorous thing on the list to evaluate. What you're probing for:

• Documented work instructions and traceability. Every pack should trace back to its cell lot, its build date, its operator.
• Monitored process parameters. Spot welding is the obvious example: weld energy and pull strength should be controlled and checked, not eyeballed.
• In-process quality checks, not a single inspection bolted onto the very end.
• A real response to failure. Ask what happens when a field unit comes back. A serious shop runs a structured root-cause process (8D or equivalent) and can show you a closed example. A weak one shrugs and ships a replacement.

That last one is the tell I trust most. Anybody can promise quality. A supplier who can walk you through a specific field failure (what happened, how they found the root cause, what they changed so it wouldn't happen again) is showing you a quality culture that exists in practice, not on a slide. How a supplier handles its worst day is more informative than how it describes its best one.

The tell: a good manufacturer interrogates your RFQ

Back to where we started. The most reliable signal of engineering competence is how hard a supplier pushes on your requirements before quoting.

I'll be straight about this from the supplier side, because it's the kind of thing that actually builds trust: a vague RFQ gets you a padded quote. When we can't see the duty cycle, the real current draw, or the temperature environment, we price for the worst case and add margin for the risk we can't see. That isn't gouging. It's the only responsible way to quote a pack you've described in fog. The suppliers asking you questions aren't being difficult. They're trying to quote the pack you actually need instead of the one that protects them from your missing information.

A capable manufacturer will want to know, at minimum:

• Continuous and peak current, and the discharge profile across a real duty cycle
• Voltage window and capacity or runtime target
• Operating and storage temperature range, including the ugly extremes
• Charging method, source, and whether the pack might ever see a charge below ~0°C (it shouldn't, because charging a cold cell risks lithium plating, a rule of thumb worth designing around)
• Enclosure, IP rating, mechanical constraints, and connector requirements
• Comms requirements (CAN, RS485, or none)
• Target market, for the compliance path
• Volume, ramp schedule, and lifecycle target
• Shipping mode, which drives the dangerous-goods and packaging plan

If a supplier quotes a custom pack without asking most of this, you've learned what you needed to know. They're either guessing or selling you something off a shelf. Writing a tight RFQ is its own skill, and a good one filters out the weak suppliers for you automatically. Our guide to writing a battery pack RFQ walks through exactly what to include.

A field-tested evaluation checklist

Pull it together into something you can actually run a supplier through. Here's how the signals stack up side by side:

And the questions I'd put to any prospective supplier, in roughly the order I'd ask them:

1. Why this cell for my discharge rate, and can I see the datasheet?
2. Do you configure BMS protection thresholds to my cell and application?
3. How do you estimate state of charge, and how does it drift over the pack's life?
4. How do you grade and match cells on the way in?
5. Can you keep lot traceability from cell to finished pack?
6. Which validation tests do you run in-house, and can I see a sample report?
7. Which standards apply to my product in my market, and can you support the reports?
8. Walk me through a field failure: what happened, and what did you change?

If a supplier answers those eight with specifics and a point of view, you're probably talking to engineers. If they answer with brochures, keep looking.

Where PackForge fits

We built PackForge as an engineering-led pack manufacturer, not a trading desk, which is why this is a framework we're comfortable being held to. We design the pack rather than relabel it: cell selection and matching, BMS configuration and firmware, thermal strategy, and the validation behind it. We'd rather ask you ten pointed questions up front than quote you a pack that's wrong for your application. If you're an OEM in the US or Germany weighing suppliers for a custom lithium battery pack across an industrial, energy storage, mobility, or field-equipment program, that's the conversation we like having.

I'm curious what's on your own list. If you've evaluated battery pack suppliers before, what's the one question, or the one red flag, that's saved you from a bad one? Or the thing you wish you'd asked before you signed? I'd like to see what other engineers and buyers actually screen for, so drop it in the comments.