LED test / review - Curiosity: Unknown 7070 LED with chip from a Luminus CBT-90/SBT-90 1st gen (5700 K 90 CRI)

Flashlight Modding and DIY Parts LED Testing
1

LED test / review EN


Unknown 7070 (CBT-90 chip, 5700 K 90 CRI)



2025-04-02


  • Type: lateral, monolithic
  • Bin: —
  • Color group: — (5700 K)
  • CRI: 90
  • Rated voltage: 3 V
  • Max. Forward current: — mA
  • Max. Peak current: — mA
  • Viewing angle: — °
  • Thermal resistance: — K/W
  • Max. Temperature Tj: max. —°C

The emitter tested here was ordered from Goofish in China several months ago. There is no further information on a data sheet or manufacturer/part number. It should be noted that buying from platforms for the Chinese domestic market such as Goofish is not trivial, and forwarding/dropshipping services such as Superbuy do not always work. I was supported in this by a Chinese user. Thank you very much at this point.

I was interested in this LED because of the large LED chip and orange phosphor. This was already clearly visible in the article photos and basically indicates a phosphor mixture with a high red content, which is typical for LEDs with high color rendering.

The LED chip sits on a dark gray substrate. It can be assumed that this is the same chip as a CBT-90 or SBT-90 1st gen, the number of bonding wires and the characteristic striped vias pattern are also identical.

The chip is surrounded by a gold-colored frame in which a thin glass pane sits.

The footprint is unusual and does not conform to any (common) standard. Although the package itself is 7.00 x 7.00 mm in size, the footprint itself does not resemble the 7070 at all. The anode/cathode connections are much larger, while the thermal pad is much smaller, which could affect heat dissipation.

Due to this strange footprint, 7070 boards cannot be used; a 5050 board was modified for the test. However, reflowing to this modified 5050 board went without any problems.

The illuminated area is 9.9 mm2. Please note that some light is reflected by the metal frame.







  • Maximum reached at 30 A, at this point 2285 lm @ 3.85 V
  • Power at maximum 115.5 W
  • Efficiency at maximum 19.8 lm/W




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


As no information on the maximum current is available, only the maximum possible current and the characteristic curve are used for the evaluation.

The LED tested here is the most inefficient one I have ever tested to date. Although the LED achieves an impressive 30 A and 116 watts, which are not bad values for such a small thermal pad, 19.8 lm/W is extremely low and is well within the range of standard halogen lamps. (For comparison: an overcurrent 12 V 20 W halogen lamp achieves over 35 lm/W at 23 V!)

Of course, all modern emitters - even those with high color rendering - run circles around the LED tested here. A modern SBT-90.2, for example, achieves a maximum luminous flux that is almost three times higher with a significantly lower Vf.

The Vf is relatively high compared to other LEDs, although this is put into perspective by the large chip and the numerous bonding wires.

Based on these measured values, I assume that this LED is very old or that the performance is severely limited by the chip. The LED chip corresponds to that of a CBT-90 and later SBT-90 (1st generation), which came onto the market in 2009. The luminous flux, which is very low by today’s standards, corresponds to that of a CBT-90 in Bin LB (approx. 1300 lm at 9 A), which speaks for the LED chip and similar phosphor mixture of a CBT-90 W57H (high CRI version).

The luminance is therefore extremely low. The LED is not really suitable for throwers.


The beam is perfectly fine. A short test with an S2+ OP reflector did not reveal any disturbing artifacts such as color fringing or different colored spill.

This LED is therefore suitable for flashlights.



The tint is good. With a CRI of 91 and R9 of 50, the spectrum is similar to that of earlier high CRI LEDs.

The LED looks like a CBT-90-W57H, but not as a chip-on-board, but as a classic emitter on its own package. This theory is supported by the precisely matching color temperature.

There is no strong tintshift with increasing current.


  • Ra: 91
  • R9: 50
  • CCT: 5712 K
  • duv: -0.0025

You can find really interesting LEDs in China, like this curiosity shown here. This LED here is extremely inefficient, especially at high operating currents, and is probably much older than the package design suggests. In general, the LED is very similar to a Luminus CBT-90, at least in terms of its electrical and optical properties, although the package and design differs massively.

For flashlights or other applications, this LED is therefore of little interest today, and is really only a curiosity.


Thank you for reading this test. :slight_smile:

Greetings, Dominik


v1.0.2
18 Thanks
2

Quite an interesting emitter.

This demonstrates the counterintuitive phenomenon where R9 scores can be lowered by too much red, rather than insufficient red. I wonder what the reason for the extra red is–it messes with the tint and lowers efficacy.

1 Thank
3

One thing that can explain why the footprint is so unusual is this carrier assembly(the combination between AIN ceramic substrate and Copper frame) is not designed for regular LED.
Instead, These type of carrier assembly is for IR Laser Emitter called VCSEL(VCSEL is a special type of IR Emitter that can produce very narrow beam like laser with only few degree of beam angle without a second optic to re-focus the beam).
The VCSEL chip is much smaller compared to regular LED chip and runs at a much lower voltage and requiring a tons of current to operate.
That’s why the electrical connection is huge and the thermal pad size reduced.

2 Thanks
4

Yes, this is a general problem with high CRI emitters. I once compared the spectra of FL5009R and this unknown LED (FL5009R 5919 K/ unknown LED 5712 K - duv also very similar):

  • The peak of the unknown LED is at 460 nm, not 450 nm
  • there is virtually no cyan hole in the unknown LED.
  • The spectral range at 660 nm is extremely strongly emphasized. It appears that a different phosphor than usual is used to emphasize the deep red area more clearly. Most high CRI LEDs use red phosphor which primarily emits at 620 nm (probably CASN:Eu2+)
    Very few manufacturers use such a deep red emitting phosphor, precisely because of the inefficiency and strong distortion of red color tones (and lower R9). Even the FFL505A 3500 K rosy does not use such a phosphor, the peak in the red range is also there at around 620 nm. It would be really interesting to know in which applications such a special phosphor mixture is necessary.

I need a better spectrometer to get more into detail with the spectral data. But they are so expensive :sob:

Interesting. Thanks for the background information.
Can you imagine why a package for VCSEL emitters was chosen? It is unlikely that such a narrow beam light source would be replaced by a normal LED in an existing application. Where could such an LED with such a footprint have been used?

1 Thank
5

That’s simply because using a existing package assembly is much cheaper than re-design a new one with standard 7070 footprint. If this LED uses a completely re-designed package assembly like the one of SFT-90(standard 7070 footprint and brass frame), It will cost a tons of money(Typically 20000CNY for the continually punch mold which make the brass frame of the LED, 2000CNY for the AR coated Quartz window frame And 3000-5000 CNY for a custom AlN ceramic substrate) to make the package assembly.
And the garbage performance of this LED chip didn’t worth that much amount of money.

4 Thanks
6

I wonder what type of phosphor was used in this unknown LED. For green-yellow, it is YAG:Ce3+, like for almost all white LEDs.
For red, it cannot be CASN:Eu2+ with λmax= 620 nm. With this LED, λmax is more likely to be 650-660 nm and the bandwidth is very high. It cannot be Sr2Si5N8:Eu2+ either, as λmax is around 610 nm.
Pretty much all high CRI LEDs (and most warm white LEDs) use CASN:Eu2+ for emissions in the red spectral range, such as the Luminus and Nichia high CRI variants.

3 Thanks
7

I have to correct this.

Most likely most high CRI emitters use SrLiAl3N4:Eu2+ (in short: SLA) phosphor, which has the emission peak at around 635-650 nm (can slightly differ due to doping concentration with Eu).

What all SLA phosphors have in common is the still quite high radiant power at 700-730 nm, which is one of the reasons for the low efficiency and the well pronounced red tint of some LEDs.

CASN:Eu2+, on the other hand, is hardly ever used because the range around 700 to 730 nm is emphasized even more clearly (up to the NIR range) and there is hardly any gain for tint, color rendering and subjective perception, at least for classic general-lighting applications.

2 Thanks
8

This is yet another level of LED knowledge depth - particular phosphor formulation for particular wavelength peaks :open_mouth: cool to know, thanks! Do You have some kind of resource where one could read up some more on that?

1 Thank
9

No, unfortunately not really…

I have spent the last few weeks and months reading and working through a lot of studies, scientific literature/papers and other sources. The topic is anything but trivial, and also requires a lot of prior knowledge of optical physics, electronics (semiconductor technology) and even some chemistry… (chemistry is what I miss the most, thanks to the almost non-existent chemistry lessons at school)

The phosphor is just one small component in the functioning of a complete LED, and all the information about it goes hand in hand to form a “big whole”.