LED test / short review - Luminus SBT-90.2 (6500 K, 70 CRI)

LED short test / short review EN


Luminus SBT-90.2


2025-03-10


  • Type: single die, lateral
  • Bin: —
  • Color group: — (6500 K)
  • CRI: 70

The emitter tested here was provided to me by TLF user @Flup.88. He purchased it from Convoy some time ago. Many thanks for this at this point!

Compared to the previously tested SFT-90, the design is very different. The 11x10 mm package consists of a dark green substrate and a metal frame, on which a thin glass pane is glued on. This protects the large LED chip and the 24 bonding wires from damage. The LED chip is identical to that of the SFT-90.

The cathode and anode are marked on the top of the substrate. The SBT-90.2 fits onto standard LED boards with a 9090 footprint.

Reflections from the metal frame, which are presumably caused by the glass pane, are noticeable. The LED chip is 9.6 mm^2 in size.





  • Maximum reached at 30 A, at this point 5882 lm @ 3.60 V
  • Power at maximum 107.9 W
  • Efficiency at maximum 54.5 lm/W




Data for 25 °C Tsp (at 85 °C the luminance values are around 13 % lower).


The performance is higher than that of the SFT-90. The efficiency is quite high thanks to the low Vf, almost reaching that of the XHP50.3 HD J4.

The reflections on the metal frame reduce the luminance somewhat. By removing the glass pane, 6-8 percent higher luminance could be achieved; due to a lack of further samples, I will not try this out.

The beam is good and is extremely similar to the SFT-90 tested earlier (beamshot of Emisar D1 with SBT-90.2 here). Depending on the reflector geometry, there may be slight yellow discoloration around the spot. Otherwise, the SBT-90.2 can be used in SMO and OP reflectors without any problems, provided that the reflector has a correspondingly large internal diameter.



The spectrum corresponds to that of a cool white LED with a low color rendering index. In contrast to the SFT-90, the tintshift with increasing operating current is clearly pronounced. When installed in a flashlight, a green tint at low levels can appear pure white with Turbo.

The tint of the sample tested here (measurement at 350 mA) is fine. As the operating current increases, the duv approaches the BBL more and more.


  • Ra: 70
  • R9: -32
  • CCT: 6008 K
  • duv: 0.0049

The SBT-90.2 has been available for several years. Despite its very high price, this LED has been used in many thrower flashlights. Due to the steadily increasing availability of the much cheaper SFT-90 with similar performance and a smaller footprint, it can be assumed that the SBT-90.2 will sooner or later become EOL.

Otherwise, the SBT-90.2 is a widely used LED with a good light pattern and high performance combined with high luminance - albeit with a corresponding current draw of over 20 amps.


Thank you for reading this test. :slight_smile:

Greetings, Dominik


v1.0.0 - SR
15 Thanks

Thank you Dominik

1 Thank

Great test! Thank you for the hard work. Interesting results and it shows the limitations of the package for output vs current input. Thia explains why we don’t see the 6000 lumens and 5500 Lumens too much in handheld lights.

This plot is super interesting: even though the SBT90.2 outperforms the SFT90 on the upper end, for lower power (<30W) the SFT90 is more efficacious. Since the SFT90 has the same LES (therefore same power density at same drive power) and a smaller footprint (less than half of SBT90.2!), achieving a higher efficacy on the lower end indicates that some substantial improvement has been made, perhaps in the phosphor mixture (which would also explain the reduced tint shift over power) or the underlying die.

Efficiency is pretty similar I think. Sft-90 has bit higher output, but also higher VF at same amperages.

The quoted plot has power (i.e., voltage times amperage) on the horizontal axis.

Agree that the efficacy (not efficiency!) is similar, but one should not expect similarity as the SFT90 has less than half of the footprint for heatsinking. So even the marginal increase in efficacy is indicative of a substantial improvement.

True, I missed that it was on watts. There’s small edge for SFT-90. Might be explainable by taking account the light loss SBT90 has on it’s glass “window”.

2 Thanks

I think that SFT-90 is very good choice for flashlights with 2 cells and buck driver like convoy M21G, L6, L7. With one cell and this led with relative high Vf I think that we can’t get some reasonable usable runtime with led current 15-20 A. SFT-90 is cheaper than SBT-90 so for very close performance You get flashlight for lower price or You can buy two sets of quality batteries for that price difference.