By Alden • Battery Engineer, Manufacturing & Quality Control • Himax Electronics • July 2026
Category: Li-ion Battery / Robot Power Pack / 18650 Cell / OEM Manufacturing / Quality Control
Most battery-related problems I encounter on the production floor don’t start with a bad cell. However, they start with a wrong cell. The wrong chemistry for the application, the wrong current rating for the load profile, the wrong configuration for the space available. By the time a battery pack comes back as a warranty claim or a field failure, the root cause is usually traceable to a cell selection decision made early in the design process — one that looked reasonable on paper but didn’t account for how the product actually operates.
Robot applications are where I see this most clearly. Robots — whether they’re autonomous mobile platforms, collaborative arms, service bots, or industrial guided vehicles — put batteries through a specific kind of stress that most lab discharge curves don’t capture well: irregular, high-variance load cycles. A motor starts, stops, reverses, idles. The battery sees a completely different demand profile every two minutes. And the pack has to handle it reliably across hundreds of cycles without heat buildup, capacity drop, or protection circuit nuisance trips.
I want to talk about why the Samsung INR18650-35E in a 1S2P configuration at 3.6V 7Ah — our model 36-7BP — handles this well, and what my team validates at the manufacturing stage to make sure the spec sheet numbers reflect what actually ships.

Himax 36-7BP — Samsung INR18650-35E 1S2P, 3.6V 7Ah, 25.2Wh, with GHR-4V-S connector and AWG26 output wire
The Cell Behind the Pack: Samsung INR18650-35E
Before I get into the pack-level specs, the cell selection is worth spending a moment on, because it explains a lot of the downstream performance characteristics.
The INR18650-35E is a Samsung SDI cell with a 3400mAh nominal capacity and an internal impedance of ≤35mΩ at the cell level. In the world of 18650 cells, that impedance figure is what I’d call competitive — it’s low enough that two cells in parallel maintain a combined pack impedance of ≤70mΩ, which is the spec we publish. For robot applications, internal impedance matters directly: lower impedance means less voltage sag under load peaks, which means motors get a more stable supply voltage during acceleration events.
The cell’s physical dimensions are max 18.55×65.25mm — the standard 18650 cylindrical format. The 1S2P configuration places two of these cells in parallel, which gives the pack its 6800mAh nominal (7Ah rated) capacity at 3.6V, with a minimum guaranteed capacity of 6600mAh. The two cells share current load, which also means each cell is operating at a lower C-rate than it would in a single-cell configuration — and lower C-rate cycling is one of the most reliable ways to extend calendar life in lithium-ion.
Pack Specs That Matter for Robot Designers
Energy Density: 25.2Wh in a 101-Gram Package
The 36-7BP packs 25.2Wh into approximately 101 grams — a gravimetric energy density of roughly 250Wh/kg at the pack level. For a robot where every gram of battery weight subtracts from payload capacity or extends runtime, that ratio matters. The pack dimensions are 36 × 18.2 × 67.5mm (±0.5mm), which fits the standard 18650 2-cell side-by-side footprint that most robot chassis are already engineered around.
Charge and Discharge: The Numbers and the Reality
Standard charge is 4.2V CC/CV at 1.4A for 6 hours. Max charge current is 3A. On the discharge side, standard is 1.4A with a cutoff at 2.65V, and max continuous discharge is 3A.
The 2.65V cutoff is a key spec for robot system designers. Most robot controllers set their low-voltage cutoff somewhere between 2.8V and 3.0V for conservatism, which is fine — the BMS and the controller can have complementary protection thresholds. What I want to flag is that discharging to 2.65V is the condition under which our 6800mAh nominal capacity is rated. If your system cuts off at 3.0V, your realized capacity will be somewhat lower than 6800mAh. This is a normal aspect of lithium-ion pack integration that sometimes catches product teams by surprise when their runtime doesn’t quite match the datasheet.

Cycle Life: 500 Cycles at ≥80% Capacity
After 500 standard charge-discharge cycles at 20±5°C, the pack retains at least 80% of its original capacity. For a robot used in a commercial environment — one charge cycle per operational day, five days per week — 500 cycles is approximately two years of service before the battery falls below the 80% threshold. At that point the robot still runs; it just runs for shorter periods between charges.
What I’ve observed in our aging test data is that packs using the Samsung 35E cell tend to follow a fairly predictable degradation curve — gradual, linear capacity loss rather than the cliff-edge dropout you see with lower-quality cells. That predictability matters for robot operators who want advance warning of declining battery performance rather than a sudden operational failure.
Full Pack Specification Reference
| Parameter | Value | Notes |
| Cell Model | Samsung INR18650-35E | Premium cylindrical Li-ion |
| Pack Configuration | 1S2P | 2 cells in parallel |
| Cell Nominal Capacity | 3400mAh | 0.2C, cutoff 2.65V |
| Pack Nominal Capacity | 6800mAh (7Ah rated) | Guaranteed min: 6600mAh |
| Nominal Voltage | 3.6V | — |
| Energy | 25.2Wh | — |
| Charge Voltage | 4.2V | CC/CV method |
| Discharge Cut-off Voltage | 2.65V | — |
| Standard Charge Current | 1.4A | 6-hour charge |
| Max. Charge Current | 3A | — |
| Standard Discharge Current | 1.4A | — |
| Max. Cont. Discharge | 3A | — |
| Cycle Life | 500 cycles | ≥80% capacity retention |
| Cell Impedance | ≤35mΩ | 1kHz AC method |
| Pack Impedance | ≤70mΩ | Incl. protection circuit |
| Charge Temp. Range | 0°C – 45°C | — |
| Discharge Temp. Range | -20°C – 60°C | — |
| Storage Temperature | -10°C – 60°C | — |
| Pack Dimensions | 36 × 18.2 × 67.5mm (±0.5mm) | L × W × H |
| Weight | ~101g | — |
| Output Wire | 1571 AWG26, 80±5mm | — |
| Output Connector | GHR-4V-S | JST GH series compatible |
| Standards | GB/T18287-2013, UL1642, CE61960 | — |
| Warranty | 1 year from shipment date | — |
Temperature Range: What -20°C to 60°C Actually Covers
First of all, the discharge temperature range is -20°C to 60°C. For robot applications this is relevant in two directions.
On the cold end: robots deployed in warehouse environments, cold storage logistics, or outdoor winter scenarios in northern climates need a battery that still delivers adequate current when the ambient is well below zero. Our electrical performance data shows temperature characteristic 2 — a -10°C soak for 3 hours followed by 1.4A discharge — yields ≥40% capacity retention. That’s a conservative threshold; in practice, you’ll typically see 50–60% retention at -10°C, with the floor at -20°C being lower. Robot designers operating in cold environments should plan their minimum-runtime requirements around the low end of this range.
On the hot end: robot motor drivers, motor housings, and dense electronics generate significant internal heat. An enclosure that runs at 35°C ambient at the battery location can spike to 50–55°C during sustained operation. The 60°C discharge ceiling provides reasonable headroom for that scenario. The temperature characteristic 1 test (40°C soak, 3 hours, 1.4A discharge) shows ≥97% capacity retention — meaning heat within the normal operating range doesn’t meaningfully degrade available capacity.
What We Actually Check Before a Pack Ships
This is the section I find most useful to be specific about, because “quality control” appears on every vendor’s website and means something different at every factory. Here’s what our manufacturing and QC process covers for the 36-7BP:
- To begin with, Incoming cell inspection
Every batch of Samsung 35E cells is verified against voltage, capacity, and impedance before assembly. We match cells for the 1S2P configuration — pairing cells with similar open-circuit voltage (≤10mV spread) and similar impedance (≤10mΩ spread). Poor cell matching in a parallel configuration leads to unequal current sharing, which means one cell ages faster than the other and the pack’s effective capacity degrades faster than the cycle-life spec suggests. Tight matching is not optional; it’s what makes the 500-cycle spec achievable in practice.
- Aging test (formation cycling)
After assembly, each pack goes through a formation cycling process — typically 2–3 charge/discharge cycles under controlled conditions before QC measurement. Formation allows the SEI layer (solid electrolyte interphase) on the cell’s anode to stabilize, which has a direct effect on the cell’s long-term capacity retention. Packs that skip formation and ship immediately off the assembly line tend to show higher first-cycle capacity loss than their rated spec.
- Open-circuit voltage measurement
After formation and before shipment, we verify that each pack’s open-circuit voltage is ≥4.1V within 24 hours of the last standard charge. This is the electrical performance specification from section 7.5 of the spec sheet. Packs that don’t meet this threshold don’t ship.
- Capacity verification
Each pack is discharge-tested at 1.4A to 2.65V after standard charge at 20±5°C and a 1-hour rest. The measured capacity must be ≥95% of rated capacity. A pack that measures 6600mAh when it should be 6800mAh doesn’t represent a defect within the guaranteed minimum — but a pack that measures below 6600mAh does, and it gets pulled.
- Furthermore, packs ship at 10–30% state of charge
Packs ship at 10–30% state of charge, at a voltage of 3.5–3.7V. This is the transport condition specified by lithium-ion shipping regulations. It’s also the storage condition least likely to cause calendar aging — a pack stored long-term at full charge degrades faster than one stored at partial charge. OEM customers who hold inventory should check the voltage on receipt, top up if below 3.5V, and not leave packs fully charged in storage for extended periods.
- AQL inspection
Outgoing quality control is conducted under AQL 0.65% normal inspection standards. This is the acceptance quality limit used in consumer electronics manufacturing — it means that in a large batch, the accepted defect rate is below 0.65%. For OEM customers building robot products with strict reliability requirements, it’s a meaningful baseline.

Robot Applications: What Draws Engineers to This Configuration
The 36-7BP comes up in robot design conversations for a fairly specific set of reasons:
- 6V nominal fits single-cell Li-ion architecture: many compact robot controller boards are designed for a 3V–4.2V input range that runs directly from a single Li-ion cell, without a step-up converter. This eliminates a BOM component and reduces conversion losses.
- 7Ah is the sweet spot for 1–3 hour runtime at moderate loads: a robot drawing 2–3A average current (common in mobile platforms with drive motors and sensors) gets approximately 2–3 hours of runtime from a 7Ah pack — enough for a meaningful work shift without being oversized.
- 101-gram weight keeps payload margins comfortable: robot designers working with payload constraints — especially in collaborative or mobile applications — can factor in 101g knowing the battery provides 25.2Wh. The energy-per-gram ratio is difficult to beat at this capacity class in a standard 18650 form factor.
- GHR-4V-S connector is widely compatible: the JST GH series connector is one of the more common choices in compact robotics and drone electronics. Many robot controller boards have a GH-compatible mating connector already specified, which simplifies mechanical integration.
- -20°C discharge capability covers most deployment environments. Consequently, it is a reliable choice for diverse robotic applications. logistics robots, outdoor inspection robots, and warehouse automation equipment operating in cold-chain or northern-climate environments can rely on this pack to deliver current in conditions that would shut down less capable cells.
Safety Testing: What the Pack Has Passed
The 36-7BP is built to GB/T18287-2013, UL1642, and CE61960 standards. Moreover, the safety test suite validates the pack against the conditions that represent real-world fault scenarios: The safety test suite validates the pack against the conditions that represent real-world fault scenarios:
- Overcharge: 3× max charge rate at 4.2V constant voltage for 7 hours — no explosion, no fire.
- Over-discharge: full charge, standby 1 hour, then 1C discharge for 2.5 hours — no explosion, no fire.
- Short circuit: external short via 50mΩ load at ambient temperature until voltage drops below 0.1V — surface temperature stays below 150°C, no explosion, no fire.
- Heating: 5±2°C/min ramp to 130°C, held 30 minutes — no explosion, no fire.
- Crush: 2MPa hydraulic press at 13kN — no explosion, no fire.
- Drop: 1 meter onto concrete, two axes — no explosion, no fire, no smoke.
- Vibration: 6mm amplitude, 10–55Hz swept at 1Hz/min, 30 minutes per XYZ axis — no leakage, no fire, no explosion.
For robot applications that export to North American or European markets, UL1642 and CE61960 certification is typically a prerequisite for import compliance. Having a certified cell from a qualified manufacturer — Samsung SDI in this case — in a pack that also carries these certifications simplifies the product compliance documentation substantially.
A Note on the 1S4P Option (3.6V 14Ah)
For robots with higher runtime requirements, the same INR18650-35E cell is available in a 1S4P configuration at 3.6V 14Ah — double the parallel cells, double the capacity, same nominal voltage. The 14Ah option makes sense for robots with higher average current draw or longer operating cycles between charges. Both configurations share the same cell, the same connector family, and the same quality process, which simplifies supply chain management for product lines that need multiple battery options.
If your application sits at the boundary — where 7Ah gives you slightly less runtime than you need but 14Ah adds more weight than you want — it’s worth discussing the actual load profile with our engineering team. There are usually design-level options that can close that gap.
Working With Himax on Robot Battery Integration
The 36-7BP is a standard configuration that ships quickly from inventory. But standard configurations are a starting point, not a ceiling. The connector type, wire length, physical enclosure or shrink-wrap finish, and labeling can all be customized for OEM integration without changing the core cell and electrical specification.
For product teams early in the design phase, the most useful thing we can do is review your load profile data — average current, peak current, duty cycle, ambient temperature range, expected daily cycle count, and runtime requirement — and tell you whether this configuration fits, needs adjustment, or whether a different pack architecture serves you better. We’d rather have that conversation at the design stage than at the warranty return stage.
The full specification sheet for the 36-7BP is available on the Himax Li-ion battery page. For robot-specific configurations, the robot battery pack section covers our range of configurations validated for mobile robot and autonomous platform applications.
Finally, if you’re ready to discuss your specific application or want to request sample packs for evaluation, the fastest path is directly through the Himax contact page. Our engineering and manufacturing teams handle technical inquiries, so you’ll get a response grounded in how the pack is actually built — not just what’s on the datasheet.
For an overview of Himax’s broader battery manufacturing capabilities and certifications, the Himax Electronics home page is a good starting point.
| About the Author
Alden is a Battery Engineer in Manufacturing & Quality Control at Himax Electronics. With hands-on experience across battery pack production lines, aging test protocols, and outgoing quality inspection, he oversees the processes that keep defect rates low and performance consistent across OEM production runs. He works directly with product teams on integration requirements and production scaling. |


































