How to Evaluate and Select LED Chip Suppliers for Therapy Devices
We switched LED suppliers last year. The new chips cost $0.003 more per LED. On 150 LEDs per mask and 20,000 masks per year, that’s $9,000 more annually. But the new chips have ±3nm wavelength tolerance instead of ±8nm, and our customer satisfaction score improved by 12 points.
The LED chip is the heart of an LED therapy device. Selecting the right supplier isn’t just about price — it’s about wavelength accuracy, consistency, reliability, and the supplier’s ability to support your growth.
The Key LED Chip Specifications
For LED therapy, these specs matter most:
1. Peak wavelength and tolerance
– The wavelength determines the therapeutic effect
– Common therapy wavelengths: 633nm (red), 660nm (red), 830nm (NIR), 850nm (NIR)
– Tolerance: ±3nm (premium), ±5nm (standard), ±8nm (budget)
– Why tolerance matters: A 633nm LED with ±8nm tolerance could output anywhere from 625-641nm. At 625nm, the photon energy and tissue penetration are measurably different from 633nm. Tight tolerance = consistent treatment results.
2. Spectral bandwidth (FWHM)
– Full Width at Half Maximum — how “pure” the wavelength is
– Typical: 15-25nm for red LEDs, 30-40nm for NIR LEDs
– Narrower is better (more targeted wavelength), but most LED chips have similar FWHM within their type
3. Radiant power output
– Measured in mW at a specified drive current (typically 20mA or 60mA)
– Higher output = fewer LEDs needed for the same power density
– Must be measured at operating temperature (output drops 10-20% at 60°C vs. 25°C)
4. Viewing angle / beam angle
– Typical: 120° for surface-mount LEDs used in masks
– Wider angle = more uniform coverage but lower power density at a given distance
– Narrower angle = higher power density but requires more LEDs for uniform coverage
5. Forward voltage (Vf)
– Affects power consumption and battery life
– Lower Vf = more efficient = longer battery life
– Vf variation between LEDs in the same batch affects brightness uniformity
6. Color rendering and visual appearance
– For red LEDs: Deep red appearance (not orange-red)
– For NIR LEDs: Faint red glow (NIR is mostly invisible, but a slight red glow indicates operation)
– Visual consistency across the LED array matters for perceived quality
The Major LED Chip Suppliers
Tier 1 (Premium):
– Epistar (Taiwan): Widely used in LED therapy, good wavelength consistency, reasonable price
– OSRAM Opto Semiconductors (Germany): Premium quality, tight binning, higher cost
– Cree/LED Professional (US): High output, excellent reliability, limited therapy-specific offerings
Tier 2 (Standard):
– San’an Optoelectronics (China): Large Chinese LED manufacturer, competitive pricing, adequate quality
– NationStar (China): Popular in mid-range LED therapy products, decent wavelength control
– MLS/Forest Lighting (China): Budget option, wider tolerance, lower consistency
Tier 3 (Budget):
– Generic Chinese LED chips (various manufacturers): Lowest cost, widest tolerance, highest variability
Our experience: We use Epistar for red (633nm and 660nm) and OSRAM for NIR (830nm and 850nm). The combination gives us the best wavelength accuracy for red (where it matters most for skin penetration depth) and reliable NIR output.
Cost comparison (per LED, 2835 package, 633nm red, 20mA):
| Supplier | Price | Wavelength Tolerance | Radiant Power | Bin Availability |
|———-|——-|———————|—————|—————–|
| OSRAM | $0.025 | ±3nm | 8.5mW | Tight binning |
| Epistar | $0.015 | ±5nm | 7.8mW | Good binning |
| NationStar | $0.010 | ±5nm | 7.2mW | Standard binning |
| San’an | $0.008 | ±8nm | 6.8mW | Limited binning |
| Generic | $0.004 | ±10nm | 5.5-8.0mW | No binning |
On a 150-LED mask, the LED cost difference:
– OSRAM: $3.75
– Epistar: $2.25
– NationStar: $1.50
– San’an: $1.20
– Generic: $0.60
The $3.15 difference between OSRAM and generic LEDs represents 1.3% of a $149 retail price. Is ±3nm wavelength accuracy worth 1.3% of your retail price? For a therapy device, yes.
LED Binning: What It Is and Why It Matters
LED manufacturers sort (bin) their production output by wavelength and brightness:
Wavelength bins: LEDs from the same production run will have slightly different peak wavelengths. The manufacturer groups them into bins of 2.5nm or 5nm width.
Brightness bins: Similarly, LEDs vary in output. They’re grouped by radiant flux into brightness bins.
Why binning matters for therapy devices:
– If you buy un-binned LEDs, your mask will have visible brightness variation and inconsistent wavelength output
– If you buy from a specific wavelength bin, all LEDs on the mask output within a narrow range
– Tighter binning = higher cost but more consistent product
Our binning specification:
– 633nm LEDs: Wavelength bin 630-636nm (±3nm tolerance), brightness bin within 10%
– 830nm LEDs: Wavelength bin 827-833nm (±3nm tolerance), brightness bin within 15%
Cost of tight binning: Approximately 20-30% premium over standard binning. On 150 LEDs, that’s $0.45-0.68 more per mask. Worth it for a therapy device where output consistency is a selling point.
Supplier Evaluation Criteria
Our evaluation checklist:
1. Wavelength accuracy and consistency
– Request spectrometer data for their LED output
– Ask for binning options and tolerances
– Test 50 LEDs from a sample batch — measure actual wavelength vs. specification
2. Quality consistency across batches
– LED output can vary between production batches
– Request data from 3+ production batches
– If possible, visit the factory and observe their quality control process
3. Long-term reliability
– Request LM-80 test data (LED lumen maintenance over time)
– Target: L70 (70% of original output) at 50,000+ hours
– For therapy devices, the LED should maintain output within 10% of initial for the product’s expected life (2-3 years of daily use = ~2,000 hours)
4. Supply stability
– Is the LED a standard product or a custom order?
– Standard products have consistent supply; custom LEDs can have long lead times
– Ask about production capacity and lead times
– What happens if they discontinue the LED? Do they offer a replacement with equivalent specs?
5. Technical support
– Can they provide application engineering support?
– Do they have experience with LED therapy applications?
– Can they recommend drive currents, PCB layouts, and thermal management for their LEDs?
6. Business terms
– Minimum order quantity (MOQ)
– Pricing tiers by volume
– Payment terms
– Lead time from order to delivery
– Warranty and return policy for defective LEDs
The LED Supplier Audit
Before committing to a supplier, we conduct a mini-audit:
Step 1: Request samples
– Order 200 LEDs in the specified wavelength bin
– Measure every LED with our spectrometer
– Compare measured specs to datasheet specs
– Calculate actual vs. claimed wavelength tolerance and output
Step 2: Build a test panel
– Assemble a small test panel (10-20 LEDs) using the supplier’s chips
– Run for 500 hours continuous
– Measure output degradation
– Check for premature failures
Step 3: Evaluate consistency
– Order a second batch (different production run)
– Compare to the first batch
– Consistent quality between batches is critical
Step 4: Negotiate terms
– Based on evaluation results, negotiate pricing, MOQ, and delivery schedule
– Include wavelength bin specifications in the purchase order
– Specify incoming quality requirements (AQL levels)
Our test results (Epistar 633nm, 2835 package):
– Batch 1: Average peak wavelength 633.2nm, σ = 1.8nm, average output 7.9mW
– Batch 2: Average peak wavelength 633.5nm, σ = 2.1nm, average output 7.7mW
– 500-hour degradation: 2.3% output loss
– Failures: 0/200
Pass criteria: Wavelength within ±5nm of target, σ < 3nm, output within 10% of datasheet, <5% degradation at 500 hours, 0 failures.
Red Flags in LED Suppliers
Avoid suppliers who:
– Won’t provide spectrometer data or binning information
– Can’t guarantee wavelength consistency across batches
– Have no LM-80 or reliability data
– Offer prices significantly below market rate (likely rejects or re-binned LEDs)
– Can’t provide references from other LED therapy manufacturers
– Require 100% payment before shipping samples
The re-binning scam: Some suppliers buy rejects from Tier 1 manufacturers (LEDs that didn’t meet the original spec), re-bin them with their own label, and sell them as “Epistar equivalent.” These LEDs may have wider wavelength tolerance, lower output, or higher failure rates. Always buy from authorized distributors or directly from the manufacturer.
What We’ve Learned
1. Don’t shop on price alone for therapy LEDs. The $0.003/LED difference between a generic chip and a quality chip is negligible in the final product cost but significant in treatment consistency and customer satisfaction.
2. Test before you commit. Our supplier evaluation process takes 4-6 weeks. It’s worth the time. We once skipped the test and went with a cheaper supplier. The wavelength was 12nm off-spec. We caught it during incoming QC and rejected the entire batch. The delay cost us 3 weeks of production.
3. Specify your bin requirements in the PO. Don’t just order “633nm LEDs.” Order “633nm ±5nm, brightness bin 7-8, Epistar P/N XXXX.” The more specific you are, the less room there is for substitution.
4. Maintain a second source. We qualified two suppliers for each wavelength. If our primary supplier has a production issue, we have a backup. The qualification cost (~$2,000 per supplier) is cheap insurance against supply disruption.
5. Re-test periodically. We verify our LED supply with spectrometer measurements every quarter. This catches any supplier changes or quality drift before they affect the finished product.
The LED chip is the component that makes or breaks an LED therapy device. Invest in understanding, evaluating, and monitoring your LED supply. The difference between a good device and a great one starts with the chip.

Leave a Reply
Want to join the discussion?Feel free to contribute!