Your portable oxygen concentrator prototype works perfectly on the bench.
Compressor delivers 90%+ oxygen purity. Algorithms optimise power consumption. Industrial design looks production ready. Then you spec the аккумулятор, and everything falls apart.
Commercial off the shelf (COTS) modules are:
- 300g heavier than your weight budget allows
- Wrong voltage (24V when you need 14.8V nominal)
- 2C continuous discharge rating (you need 5C pulse for motor startup)
- Missing IEC 60601-1 certification, adding months to your timeline
We’ve delivered certified battery systems for 12+ portable medical devices. Here’s what actually works.
What Medical Device Engineers Actually Need
Off the shelf battery modules fail portable oxygen concentrators in four specific ways.
Wrong Voltage Architecture
Your motor controller needs 14.8V nominal (3S4P or 3S5P Li-ion configuration). COTS packs come in 24V or 7.4V. Redesigning your power electronics around available batteries is backwards engineering.
Inadequate Pulse Discharge Capability
Your compressor draws 8A steady state but spikes to 25A for 200ms during motor startup. Most COTS packs are rated for 2C continuous discharge. This rating is fine for steady loads, but it is fatal for motor startup transients.
When voltage sags below the controller’s minimum threshold during that 200ms pulse, your compressor stalls. Oxygen delivery stops. In a medical device, that’s a potentially fatal failure mode.
Certification Gaps That Delay Launch
You need IEC 62133-2 (cell safety), ООН38.3 (transport), and integration into your IEC 60601-1 device certification. COTS suppliers provide the first two. The third becomes your problem. Unfortunately, this is a reality that most teams discover too late.
Form Factor Constraints
Your industrial design has a 180mm × 65mm × 45mm cavity. COTS packs come in standard sizes. “Close enough” means redesigning your enclosure or accepting 15% more volume.
We design around your requirements. Not around what’s available off the shelf.
The Energy Density Reality Check
Walk into any battery supplier and they’ll quote you impressive numbers.
“Our cells deliver 270 Wh/kg.”
True. However, it is the cell level. You’re not shipping bare lithium cells in a medical device.
What Gets Added to Bare Cells
- Система управления аккумуляторами – monitors cell voltages, manages charge/discharge, handles safety cutoffs
- Protective circuitry – overcurrent protection, short circuit protection, thermal monitoring
- Structural enclosure – must survive 1 metre drop onto concrete per IEC 60601-1
- Interconnects and wiring – rated for medical device current and temperature requirements
- Thermal management – keeps cells below 45°C during continuous flow operation
- Safety disconnects – redundant protection if BMS fails
All that essential infrastructure cuts energy density nearly in half. Premium 21700 cells might hit 270 Wh/kg on the datasheet. Your finished medical battery pack delivers 150 to 180 Wh/kg in production.
This isn’t bad engineering. It’s physics and regulatory reality.
What This Means for Your Product
A realistic 160 Wh/kg pack delivering 80Wh provides:
- 4 hours of pulse dose operation
- 2 часа of continuous flow operation
- 500g battery assembly
- 2.8kg total device weight
Your marketing team wants to claim “6 hour runtime.” Your engineering team knows that requires a heavier pack that breaks your portability threshold, a less efficient compressor, or pulse dose mode only. Pick one.
We help you navigate these trade-offs with real numbers.
Why Your Compressor Motor Destroys Bad Batteries
Oxygen compressor motors are brutal on battery packs in ways that don’t show up in steady state testing.
The Startup Surge Problem
Compressor motors need high torque to overcome static friction on the first stroke. That initial surge pulls three to five times normal operating current for 100 to 300 milliseconds.
Cheap cells with high DC internal resistance crater during this surge, causing the voltage to drop sharply. The motor controller’s minimum threshold is typically 10.5V for a 14.8V nominal system. If the voltage falls below this level, the compressor stalls and oxygen delivery stops.
Temperature Makes Everything Worse
Cell internal resistance varies dramatically with temperature and state of charge:
- At 0°C: Resistance can triple compared to 25°C
- At 20% charge: Resistance roughly doubles compared to 80% charge
- Combined effect: A cold, nearly depleted pack can have 6x the resistance of a warm, fully charged one
Room temperature bench testing with fresh cells hides the failure mode that kills devices in the field. When operating in cold weather, with only a 30% charge remaining and three kilometres from home, it behaves like a completely different battery.
Cell Selection Criteria That Actually Matter
We specify cells rated for:
- 3C continuous discharge – for extended continuous flow operation
- 5C pulse capability – for motor startup without voltage sag
- Stable performance across 0°C to 40°C – full operating range
- Consistent output through degradation – still meets spec at 80% capacity after 500 cycles
We test packs under your actual load profile at temperature extremes and various states of charge. Not just room temperature continuous discharge.
The Hidden Engineering Challenge: Light AND Durable
Industrial battery packs use thick plastic and steel brackets because nobody cares about an extra 200 grams. Medical oxygen concentrators care desperately about every gram, even while facing mandatory testing that industrial packs never see:
- Drop test – 1 metre onto concrete, 6 orientations, no rupture or electrical safety failure
- Vibration test – simulated transport and daily use, all connections intact
- Thermal cycling – −40°C to +60°C, enclosure maintains structural integrity
- Crush test – simulated user sitting on device, cells undamaged
How We Engineer Light AND Safe Enclosures
Finite element analysis identifies exact stress points. We model drop impact and vibration loads in CAD, then add reinforcement only where physics demands it, rather than using uniformly thick walls that waste material. Custom polycarbonate enclosures optimised this way save 15 to 25% mass compared to standard industrial designs. On a 500g pack, that’s 75 to 125g saved, providing enough headroom to add 15Wh of capacity within the same weight budget.
Strategic material selection:
- Polycarbonate for high impact zones – 30% lighter than ABS with better impact resistance
- Aluminium heat spreaders only where thermal modelling identifies hot spots
- Polyimide films for electrical insulation – thinner and lighter than traditional polymers
The Adhesive-Only Myth
Pure adhesive bonds fail under shock loads after thermal cycling degrades bond strength. We’ve seen packs pass initial drop testing, then fail after 200 thermal cycles – 18 months into field use with 5,000 units deployed.
Our approach combines structural adhesive with mechanical retention features moulded directly into the enclosure. Adhesive handles normal loads and vibration. Mechanical features prevent catastrophic failure as adhesive ages.
Design for Field Replacement
Telling a patient to bin their £3,000 concentrator because the battery died after 18 months is both wasteful and commercially short-sighted. Your competitors offer replaceable batteries.
We design packs with documented replacement procedures, connectors rated for 50+ insertion cycles, keyed interfaces that prevent incorrect installation, and BMS diagnostics that report cell health before catastrophic failure occurs.
Serviceability has to be designed in from the first CAD model. It cannot be retrofitted.
Fuel Gauging: Why Patients Need Minutes, Not Bars
Consider what patients actually see with a well-designed BMS:
The display shows “87 minutes remaining in pulse mode.” They switch to continuous flow. It updates to “41 minutes remaining.” They know exactly what they can do with remaining charge, and they leave the house.
Now consider what they see with voltage-based gauging: three bars that mean almost nothing through the middle 60% of discharge, where the lithium voltage curve is nearly flat. A patient who thinks they have 45 minutes remaining but actually have 15 gets stranded. Your device gets blamed.
Why Voltage-Based Gauges Fail
A lithium cell’s voltage tells you almost nothing about remaining capacity between 80% and 30% state of charge. The curve is flat. Accuracy degrades to ±25-30% in real conditions. This margin of error represents the difference between making it home and not.
What Advanced BMS Systems Provide
Modern BMS firmware uses coulomb counting combined with impedance tracking, measuring actual charge flow rather than inferring capacity from voltage. The system accounts for:
- Capacity fade as the battery ages
- Power demand differences between pulse dose and continuous flow
- Temperature effects on available capacity
- Recent load history
This delivers ±5% accuracy under real operating conditions.
Integration Requirements
This requires proper communication between BMS and main system controller from the start:
- SMBus or I2C protocols for real-time data exchange
- Mode switching updates so BMS recalculates when compressor changes modes
- Temperature data sharing between system controller and BMS
- Failure status reporting before degraded cells become a clinical problem
Plan this integration from day one. It cannot be bolted on during final testing.
The Certification Path: Realistic Timelines and Costs
Someone always asks: “Can’t we modify an existing certified pack slightly?”
No. Any modification to a certified medical component triggers full recertification. You might save 6 weeks in development and lose 8 months waiting for retest results.
IEC 62133-2: Cell and Battery Safety
Tests external short circuit, thermal abuse, mechanical shock, vibration, and overcharge protection. Your cell supplier certifies individual cells. You still need separate pack-level testing because your BMS, enclosure, and interconnects change the safety profile.
Samples required: 3 to 4 packs | Cost: £1,500–£2,000 | Timeline: 3 to 4 weeks
UN38.3: Transport Safety
Eight mandatory tests simulating air transport: altitude exposure, thermal cycling, vibration, mechanical shock, short circuit, crush, overcharge, and forced discharge. Mandatory if patients travel by air, and they will.
Samples required: 3 to 4 packs | Cost: £700–£900 | Timeline: 3 to 4 weeks
IEC 60601-1: Medical Electrical Safety Integration
Your battery must integrate into complete device certification. The testing lab evaluates cell venting impact on enclosure IP rating, thermal runaway user burn hazard, electrical fault shock hazard, and BMS failure impact on oxygen delivery. The notified body reviews your FMEA for every battery failure mode.
Cost: £2,000–£3,000 for battery subsystem | Timeline: 4 to 8 weeks for battery-specific testing | Note: This integrates into your complete device certification, which typically runs 6 to 12 months
ISO 13485: Quality Management System
This is not a product test; it is a process requirement. Every component needs documented traceability: cell manufacturer qualification, incoming inspection procedures, assembly process validation, and change control.
Cost: Internal (typically 0.5 to 1.0 FTE quality engineer) | Timeline: Ongoing
Real Cost of a Cell Swap
A client substituted one cell model sharing the same manufacturer, same capacity, and an “improved” status. The notified body required full IEC 62133-2 retesting because the chemistry datasheet showed minor differences.
Result: 4 weeks of lab time. £1,800 in fees. One month delay. Freeze your cell specification before starting certification.
Total Realistic Timeline
How We Reduce Certification Risk
We’ve navigated IEC 60601-1 with four different notified bodies across EU and US markets. We know which test protocols apply to battery subsystems and which documentation gets requested during review.
We build that documentation from day one: complete FMEA for every battery failure mode, test reports showing BMS response to fault conditions, thermal runaway mitigation analysis, and software validation for BMS firmware.
This doesn’t make certification faster. It makes it predictable, avoiding the risk of discovering missing requirements when you’re 80% through review.
Supply Chain Reality: Why Cell Selection Goes Beyond Specs
Your procurement team will ask: “Why can’t we use cheaper cells?”
Here’s what they’re not seeing.
Cell Availability Over a 5+ Year Product Life
You’re designing a device with a 5 to 7 year market life. Consumer electronics cells get discontinued every 18 months as manufacturers chase smartphone design cycles.
We select cells from manufacturers with medical and industrial product lines. These suppliers offer long-term availability commitments because their automotive and medical customers demand it.
What poor cell selection actually costs: A consumer-grade 18650 used in a portable ventilator went obsolete 14 months after product launch. The manufacturer had 8,000 units in field with no compatible replacement. Emergency redesign and recertification cost £180,000 and delayed their next product by 6 months.
Lot-to-Lot Consistency
Consumer cells carry ±5% capacity variation between production lots. Medical-grade cells hold ±2% with full lot traceability. Every cell has a date code and lot number, so if a quality issue surfaces, you can trace it to a specific production batch.
This matters when you’re claiming “4 hours runtime” to clinicians who make care decisions based on that number.
Supplier Qualification Under ISO 13485
Your quality system requires supplier audits. Can your cell supplier provide:
- Conflict minerals declarations (required for EU and US medical device sales)
- RoHS and REACH compliance documentation
- Factory audit access for your quality team
- Documented change notification – 6 months advance notice before any process change
Tier 1 manufacturers, such ass Samsung SDI, LG Chem, Panasonic, Murata, they provide all of this. Alibaba vendors provide none of it.
Second Source Strategy
During COVID, battery lead times went from 8 weeks to 26 weeks. Companies with single-source designs stopped production.
We qualify second-source cells during initial design and maintain that documentation throughout the product life. When your primary supplier goes on allocation, you switch to the qualified alternate. This eliminates the need for emergency redesign or recertification scrambles.
This matters most when you’re scaling from 100 to 1,000 units per month and your primary supplier can’t keep up.
Thermal Management and Battery Life
Why Operating Temperature Defines Product Economics
Most quality lithium cells tolerate 60°C sustained operation. But cycle life degrades sharply above 45°C:
| Operating Temperature | Cycles to 80% Capacity |
| 45°C | 500 cycles |
| 35°C | 1,000+ cycles |
| 25°С | 1,500+ cycles |
For a device used daily, that’s the difference between 18-month and 48-month battery life. It is also the difference between manageable service costs and a product economics problem.
We target maximum pack temperature below 50°C during continuous operation at 25°C ambient. This provides margin for higher ambient temperatures in summer, thermal performance degradation as adhesives age, and dust accumulation on vent paths reducing airflow.
Thermal Design Approach
Cells in the pack interior run hotter than perimeter cells. Cells near BMS components receive additional heat from electronics. Thermal modelling during design identifies these hot spots and guides strategic cell placement and heat spreading.
Most portable oxygen concentrator packs use passive cooling, relying on thermal conduction to the case followed by natural convection. Active cooling adds 20 to 40g of mass, 0.5 to 1.0W of parasitic draw, another failure mode, and noise that disrupts nighttime use.
We design passive cooling that handles typical use, backed by BMS temperature monitoring that alerts patients when pack temperature exceeds safe limits, reduces compressor speed if temperature continues rising, and logs thermal data for field failure analysis.
Extending Practical Cycle Life
Design the pack larger than minimum requirements. If you need 80Wh for 4 hours runtime, design a 120Wh pack. Typical use only discharges 65% instead of 100%, extending cycle life toward 1,000 to 1,500 cycles. The weight and cost premium pays back through reduced field battery replacements.
Limit charge rates to 0.5C. Medical devices charge overnight. There’s no clinical benefit to 1-hour fast charging. Slower charging reduces lithium plating and extends calendar life significantly.
Build replacement into the product design from day one. Patients should expect battery replacement after 24 to 36 months of daily use; this is an honest approach. Forcing them to replace a £3,000 concentrator because the battery died creates competitive vulnerability and damages your brand.
What Happens When You Contact Us
We need your actual requirements, not aspirational targets.
What to Send
- Compressor electrical specs – startup current, steady-state draw by mode, voltage range, peak pulse duration
- Physical constraints – maximum dimensions, weight budget, connector location
- Operating environment – temperature range, altitude, duty cycle
- Target runtime – hours per charge in each operating mode
- Production volumes – affects tooling investment decisions
- Timeline and certification pathway – FDA 510(k), EU MDR, other markets
Feasibility assessment first. Some combinations of performance, size, and weight cannot be achieved with current lithium technology. If your requirements are achievable, we’ll map out the available trade-offs: add 50g and gain 1 hour runtime, reduce peak current to 4C and save £8 per pack, accept 3.5-hour minimum instead of 4.0-hour guaranteed. Real options with real numbers.
Preliminary design with engineering justification. Cell selection rationale, pack architecture options, BMS feature requirements, performance estimates at temperature extremes and end of life, rough cost at your production volumes. Two weeks.
Prototype and validate before committing to certification. We build prototype packs and test them under your actual load profile: motor startup transients at various states of charge, continuous flow to thermal steady state, pulse dose operation matching your clinical duty cycle, temperature extremes at 0°C and 40°C, and end-of-life performance simulation.
Spending £50,000 on prototypes that reveal design flaws is cheap. Spending £300,000 on certification testing that fails because you skipped prototype validation is expensive.
Certification support with realistic timelines. We provide test plans mapped to IEC 62133-2 and IEC 60601-1, FMEA documentation for notified body review, design history file documentation for regulatory submission, and supplier qualification documentation for your ISO 13485 audit.
What We Don’t Do
We don’t promise magical solutions. Battery energy density improves 3 to 5% per year. There are no 50% breakthroughs coming.
We don’t ignore regulatory requirements. “We’ll figure out certification later” is how projects fail at the finish line after 18 months of development.
We don’t pretend development takes 3 months when it takes 18 to 24. Medical device timelines are long. Compressing them on paper doesn’t help anyone.
Часто задаваемые вопросы
How long does custom medical battery development actually take?
18 to 24 months from initial requirements to regulatory clearance: design (3–4 months), prototyping (2–3 months), design validation (2–3 months), component certification testing (3–4 months), device integration testing (6–12 months), regulatory review (6–12 months). Phases overlap where possible. Companies promising faster timelines are skipping validation steps or don’t understand medical device requirements.
What happens if our cell supplier discontinues the specified cell?
We qualify second-source cells during initial design and maintain that documentation. When primary cells go obsolete, you switch to the qualified alternate without emergency redesign or recertification. This costs more upfront. It costs far less than the alternative.
Can patients replace batteries themselves or does it require a service centre?
Either approach works. Patient-replaceable designs need foolproof connectors and clear instructions. Service-centre replacement allows more complex integration. We’ve designed both successfully. The right choice depends on your service model and cost structure, not on what’s technically easier to build.
