DALI-2 LED Driver

DALI vs DALI-2: What the Certification Actually Means

The single most common sourcing mistake for DALI drivers is confusing “DALI compatible” with “DALI-2 certified.” Chinese driver manufacturers frequently mark products as “DALI compatible” without having completed any DiiA (Digital Illumination Interface Alliance) certification process. The distinction matters in practice.

DALI (original, pre-2014). The original IEC 62386 specification defined the electrical bus, command set, and device types, but did not require formal third-party interoperability testing. Manufacturers self-declared compliance. The result: dozens of incompatible implementations, each claiming “DALI support,” that failed to interoperate reliably across brands.

DALI-2 (DiiA certification, 2014–present). DALI-2 requires:

1. Testing at a DiiA-accredited laboratory against IEC 62386 Part 101 (system layer) and the applicable device-type part — Part 202 for control gear (LED drivers), Part 207 for LED color control, Part 209 for color temperature control.
2. Interoperability testing with at least five reference controllers or devices from other DiiA-certified manufacturers.
3. Registration in the DiiA product database (searchable at dali-alliance.org by manufacturer and part number).

A driver that has not completed steps 1–3 cannot legally carry the DALI-2 logo. The DiiA enforces this via trademark licensing — use of the logo without registration is a trademark violation, not just a marketing claim issue.

Why “DALI compatible” fails in DALI-2 installations. DALI-2 systems — Helvar, Tridonic, Osram/Siteco, Lutron — implement extended IEC 62386-202 features: multi-master bus arbitration, device instance addressing, scene recall with fade time, and the automatic testing function for emergency lighting (Part 301). Uncertified Chinese drivers frequently fail on:

• Bus timing margins. IEC 62386 specifies forward frame timing at 2.4ms ± 10%. Some domestic drivers hold the bus too long during address assignment, causing bus collisions during commissioning.
• Query responses. QUERY ACTUAL LEVEL and QUERY MAX LEVEL commands return correct values only if the driver correctly implements the arc level to output mapping specified in Part 202. Drivers that return hardcoded or approximated values break the controller’s dimming feedback loop.
• Scene recall accuracy. Controllers write scene levels to driver EEPROM and verify readback. Drivers with EEPROM write timing issues fail commissioning in Helvar and Tridonic tools.

DiiA certification cost and who bears it. DiiA certification costs $3,000–8,000 per product, depending on the laboratory (TÜV Rheinland, Bureau Veritas, and DiiA’s own lab in Sweden are common). For OEM buyers sourcing an existing certified platform, this cost is already embedded in the product cost — you are buying access to an existing DiiA registration, not triggering a new certification run. For custom ODM programs requiring a new part number and fresh DiiA registration, budget $4,000–6,000 and 8–14 weeks for the certification process before production can begin.

Our sourcing service verifies DiiA registration status for every driver under consideration — we search the DiiA product database by manufacturer and part number before recommending a supplier.

Dimming Curve, Arc Levels, and Flicker at Low Dim

The logarithmic arc level system. IEC 62386 defines 254 usable arc levels (0 = off, 1–254 = minimum to maximum output). The standard specifies that the relationship between arc level and light output should be perceptually linear — i.e., each arc level step should produce an equal perceived brightness increment. Perceived brightness follows a logarithmic relationship with physical light output (Fechner’s law), so the required mapping is:

Light output at arc level N = (10^(log100/253))^(N-1) × minimum output

Each arc level step is a multiplier of approximately 10^(log100/253) ≈ 1.0063 — or about 0.027dB per step. At arc level 254 (maximum), the light output is 100%. At arc level 1 (minimum), the output is approximately 0.1% of maximum for drivers implementing the full extended range (Part 209), or roughly 0.4% for Part 202 standard minimum.

Why linear arc-level to output mapping causes visible steps. Some cheaper Chinese DALI drivers implement a linear mapping — arc level 127 produces 50% output, arc level 64 produces 25%, etc. At low dim levels (arc levels 1–30), each step represents a large jump in perceived brightness because the eye is operating in the high-sensitivity logarithmic region. A step from arc level 5 to arc level 6 in a linear-mapped driver produces a visible flicker that makes smooth continuous dimming impossible. Specify logarithmic dimming curve compliance (IEC 62386-202 Section 9.3) in your driver specification, and verify it with a photometer during incoming QC.

Flicker quantification. The IEEE 1789-2015 standard defines two metrics:

• Flicker Index (FI): ratio of the area above the average light output on a waveform cycle to the total area of the waveform cycle. FI = 0 is perfectly flicker-free; FI = 1 is maximum flicker (on/off square wave). IEEE 1789 recommends FI ≤ 0.01 for general illumination.
• Percent Flicker (%F): (max - min) / (max + min) × 100. More intuitive but less complete than FI for complex waveforms.

PWM dimming vs analog dimming trade-offs.

PWM dimming switches the LED load at high frequency (typically 500Hz–100kHz). Flicker at the switching frequency is invisible above about 1kHz, but the FI depends on duty cycle — at 10% dim, the driver is switching at 10% duty cycle, and FI is determined by switching frequency and waveform rise/fall time. At 50kHz PWM with fast switching, FI can be maintained ≤ 0.01 even at 1% dim. Below 20kHz, the IEEE 1789 low-risk threshold requires FI ≤ 0.008 at frequencies from 90–3000Hz — hard to meet at low dim levels.

Analog dimming (DALI current control, or 0–10V) reduces LED drive current continuously. FI is effectively 0 at all dim levels because there is no switching. Trade-off: color shift. At reduced current, LED forward voltage drops and the junction temperature decreases, shifting the correlated color temperature (CCT) of the LED. In warm white LEDs (2700–3000K), this is typically a blue shift of 100–200K between 100% and 10% load — perceptible in environments where CCT consistency matters (museum lighting, retail).

High-end DALI-2 drivers from Tridonic and Helvar implement hybrid dimming: analog current reduction down to approximately 10% output, then PWM at high frequency for the remaining 10%–1% range, maintaining FI ≤ 0.01 throughout. Chinese driver manufacturers implementing this hybrid approach include Inventronics and Eaglerise — confirm the dimming method in the driver datasheet before sampling.

For flicker measurement during incoming inspection, measure FI and %F at 1%, 10%, and 50% dim using a photometer with waveform capture capability (Admesy Brontes, Instrument Systems CAS 140D, or equivalent). Do not rely on driver manufacturer’s claimed flicker specifications without independent measurement.

Emergency Lighting: DALI Part 301 and Battery Backup Integration

DALI device part differentiation. IEC 62386 defines multiple device parts relevant to emergency lighting:

• Part 202: Standard control gear (dimming, scene recall, fade). This is the baseline DALI driver specification.
• Part 301: Self-contained emergency control gear. Adds battery management, emergency mode, rest mode, inhibit mode, and automatic duration/function test commands.
• Part 201: Input devices (sensors, push-buttons). Not a driver — covers the controller side.

A DALI driver carrying only Part 202 certification cannot participate in DALI-based emergency lighting control. Emergency lighting applications require Part 301 certified control gear, or a separate emergency conversion kit (ECK) paired with a standard Part 202 driver.

Battery backup within the driver. DALI-2 Part 301 drivers integrate a battery pack — typically NiMH at 3.6V–7.2V nominal — capable of maintaining emergency lumen output for the required duration. Standard durations:

• 1 hour (Class B, common in UK/EU offices and corridors)
• 3 hours (Class C, UK BS 5266-1 requirement for higher-risk premises)

Emergency lumen output is typically 10–15% of normal rated output — sufficient to meet the maintained 1 lux at floor level (1 lux on escape routes per BS 5266-1, 0.5 lux in open areas per EN 1838).

Automatic testing via the DALI bus. IEC 62386-202 and Part 301 define an automatic testing function: the controller can initiate a function test (verify lamp and battery respond to emergency command) and a duration test (run on battery for the full required duration) via standard DALI commands. Test results — pass/fail, battery charge level, duration achieved — are logged in the driver and retrievable via QUERY commands. This eliminates manual press-and-hold testing and provides an auditable electronic log for EN 50172 and BS 5266-1 compliance.

Documentation for BS 5266-1 / EN 50172. These standards require:

1. A log of all function tests (monthly) and duration tests (annually for 3-hour systems, 6-monthly for 1-hour systems).
2. Evidence that the emergency luminaire maintained the required lumen output for the full test duration.
3. Certificate of compliance from the luminaire manufacturer (which references the driver’s Part 301 certification).

When sourcing Part 301 drivers from Chinese manufacturers, request the DiiA certification certificate (Part 301 specifically, not just Part 202), the test report from the accredited lab, and a sample of the EN 50172 log report generated by the DALI controller integration. Suppliers unable to provide all three documents are not supplying a compliant Part 301 product.

Chinese Supplier Landscape and Incoming QC

Reference tier: European certified suppliers. Tridonic (Austrian, Zumtobel Group) and Helvar (Finnish) are the benchmark DALI-2 suppliers for commercial lighting integrators. Both maintain comprehensive DiiA product databases, publish full IEC 62386 command implementation tables, and offer direct technical support for commissioning issues. Lead times from European distribution are 4–12 weeks; pricing is $45–180 for 30–100W equivalents.

Chinese manufacturers with DiiA-registered products. The DiiA product database includes several established Chinese manufacturers:

• Inventronics (Hangzhou): Broad DALI-2 portfolio, Part 202 and Part 207 certified. Strong technical documentation. OEM programs available at 500+ unit MOQ.
• Eaglerise (Zhongshan): DALI-2 certified range, 10W–200W. Known for competitive pricing in the 20–50W range. OEM label programs at 200+ units.
• Moons’ Industries (Shanghai): DALI-2 certified drivers with strong focus on tunable white (Part 209). Engineering support available in English.
• OHM-E (Shenzhen): Smaller catalogue, but DiiA registered. Often available at lower MOQs (100 units).

Before shortlisting any Chinese supplier, search the DiiA product database (dali-alliance.org/dali-2-products/) by manufacturer name and the specific part number offered. A manufacturer with DALI-2 registered products in one wattage range does not automatically have certification for the product offered to you — verify the exact part number.

Incoming QC protocol for DALI-2 drivers. Our inspection service runs the following checks on sampled units from each production batch:

Electrical verification:

• Power factor measurement at 25% and 100% load, both at rated input voltage. PF ≥ 0.95 at 100% is the stated spec; verify PF ≥ 0.90 at 25% load.
• Efficiency measurement at 25%, 50%, 75%, and 100% load using a power analyzer (Hioki PW3390 or equivalent). Flag units below 87% efficiency at any test point.
• Inrush current measurement at cold start (25°C ambient, rated input voltage). Inrush should be ≤ 25A peak for <50W drivers; ≤ 40A peak for 100–150W. Excessive inrush trips MCBs in multi-driver lighting circuits.
• Output ripple current measurement at full load. IEC 62386 does not specify ripple limit, but LED manufacturers typically require ripple ≤ 10% of DC output current to avoid accelerated LED degradation.

DALI bus communication verification:

• Address assignment: broadcast INITIALISE, then RANDOMISE, then assign address 0. Verify driver responds to address 0 QUERY STATUS with correct group and scene defaults.
• DAPC (Direct Arc Power Control) sweep: send DAPC commands from arc level 1 to 254 in 10-step increments. Verify light output increases monotonically and approximately logarithmically. Any output decrease or plateau between steps indicates a firmware defect.
• QUERY ACTUAL LEVEL: after sending DAPC(127), send QUERY ACTUAL LEVEL. Response must equal 127 ± 1.
• QUERY MAX LEVEL and QUERY MIN LEVEL: verify these match the driver’s specified maximum and minimum arc levels. Hardcoded wrong values (a common defect in uncertified drivers) break controller auto-commissioning.
• Scene write and recall: write scene 0 = arc level 100, scene 1 = arc level 200. Power-cycle the driver. Recall scene 0 and scene 1, verify output levels are maintained after power cycle (EEPROM write validation).

Flicker measurement:

• Measure FI and %F at arc levels 1, 5, 10, 25, 50, 127, and 254 (approximately 0.1%, 0.4%, 1%, 4%, 10%, 50%, 100% output).
• Flag any unit with FI > 0.01 at any dim level.

For commercial lighting projects, we recommend a 10-unit sample inspection from the first production batch, with 100% DAPC sweep testing and a 5-unit flicker measurement subset. Contact us via the sourcing inquiry form to discuss inspection scope for your specific project.