LED test / review – Lumileds Luxeon MZ 5700 K min. 90 CRI (3V) – Finally! High CRI, High power, beautiful light

Flashlight Modding and DIY Parts
1

LED test / review

EN

Lumileds Luxeon MZ (3 V variant)

LMZ9-QW57



UPDATE --- 06/22/18 --- diagram Vf / light flux depending on Tj added!

UPDATE --- 06/23/18 --- raw data added! (link)



Some time ago I found a LED which have sparked my interest. The most LEDs with high color rendition (known as ‚high CRI‘) are not very powerful, have not the best efficiency and also can’t handle high currents like other LED series.

The Luxeon MZ is very interesting because of the design, the size and the very low thermal resistance which is much lower as of all other well-known High CRI LEDs and should make a very high performance possible.

I ordered some of this LEDs at Mouser. Some months ago I posted an earlier test of this emitter already. With this new test the old one is no longer vaild anymore und will be also not longer supported!



Technical data

Tj 85 °C / If 2,800 mA

Order code: LMZ9-QW57

Type: quad die (Flip Chip)
Binning: typ. 770 lm (min. 700 lm) @ 85 °C
Rated voltage: typ. 2.80 V (min. 2.63, max. 3.00 V)
Max. forward current: 4,800 mA
Max. peak forward current: 5,500 mA
Viewing angle: typ. 120° (total 140 °)
Thermal resistance: typ. 1.25 °C/W
Junction temp.: max. 150 °C

The official datasheet can be found here (offical Lumileds website, pdf)




First appearance

Unless most other LEDs, Lumileds chooses an own and very special footprint. The LED itself is 4.00 x 4.20 mm (0.157 x 0.165 in) in size. Because of the domeless design this LED is very flat.

The footprint is not the same as XM or XP. The size is between the two, which makes the using of industry standard boards more difficult. Additionally the use of lathe spun center rings is not (fully) possible, because of the irregular size.

The use of XP boards is possible but there is not much space for the thermal pad. If not reflowed correctly the thermal pad and anode/cathode are shorten. In the most cases there are several attempts required to solder the MZ correctly.

On XM boards this LED simply doesn’t fit.

The LED features a electrically insulated thermal pad. This makes using of DTP boards easy. Above the thermal pad there is some kind of matrix code, possibly to name batch or manufacturing date.

LED chip and die

The Luxeon MZ features four LED dies (like XHP50 / 70, but without dome) which makes this LED very special. In size and design it equals to a dedomed XHP50. And like this, there are also gaps between the individual dies. An ugly beam pattern can at least be expected here unless you use a diffuse or matte foil / lens.

Like other Lumileds LEDs the MZ is designed in Flip Chip technique. Big advantage is that no bonding wires are needed which covers small areas of the LES. There are no other colored surfaces around the single dies, like first generation XHP LEDs.

The total LES is in 7.92 mm² (0.01228 sq in) in size. Every die covers an area of 1.98 mm² (0.003069 sq in). I assume that the single dies are the same Lumileds use in the Luxeon Z ES.



Performance and overcurrent capabilites.

25 °C Tsp, unless noted otherwise

Raw data available here!

Within official parameters

  • At 4,800 mA (max rated current): 1,344.6 lm @ 3.00 V (peak 5,500 mA: 1,492.6 lm @ 3.04 V)
  • Power at rated maximum: 14.40 W (peak: 16.72 W)
  • Efficiency at 4,800 mA: 93.4 lm/W (peak: 89.3 lm/W)
  • At 2,800 mA (binning current, but 25 °C Tsp): 867.8 lm @ 2.90 V
  • At 2,800 mA (binning current, corrected to 85 °C Tsp acc. To Lumileds datasheet): 818.7 lm

Overcurrent:

  • Maximum at 16,800 mA, 2,909.1 lm @ 3.48 V
  • Power at maximum 58.5 W
  • Sweet spot at 9,000 mA (2,173.6 lm @ 3.20 V)
  • Power at sweet spot 28.8 W
  • Efficiency at max 49.7 lm/W, in sweet spot 75.5 lm/W

I define the sweet spot as the position in the diagram which gives a good average between light flux, current and efficiency. In most cases it lays roughly 15 to 35 percent lower than maximum possible light flux.

Interesting facts

  • The stated binning of typ. 770 lm @ 85 °C Tj is reached.
  • The Luxeon MZ is very powerful and reaches 16,800 mA before the light flux decreases. So it can also be powered higher than the XP-L2.
  • The Vf is very low. Even at 16.8 Amps it stays below 3.5 V!
  • Unless the most other LEDs on XP boards which reaches their maximum power at 45 – 50 Watts, the MZ delivers max. 59 Watts.
  • This is one of the very few (available!) LEDs which can deliver very high power and high light quality (details below)
  • In case of building a Triple light with this LED the currents to deliver over 5000 lumens could exceed 25 to 35 Amperes which makes the design of the driver very complex and requires new technologies to build such a light.
  • Because all four dies are connected in 4P (4x parallel), at 16.8 A every of them runs at 4.2 A – this is the reason why the Vf is very low. At this current such voltage of a single LED chip can be expected.
  • The MZ is also available in 6 (RW) and 12 V (SW) variant.

The manufacturer Lumileds doesn't provide any data regarding changing Vf at rising temperature Tj, so these data are very new for all of us. It is important to know that the max Tj acc. to manufacturer is 135 °C, not 150 as shown in my diagram! I cannot recommend also the operation at higher Tj than stated in the official datasheet!



Luminance

I established a new method for determining luminance, especially to ensure more realistic values for 'real-life' conditions (flashlight use). The measurements are taken with a new original Convoy C8 reflector, but with same heatsink / setup as previously used in flux measurement.

Reflectivity for reflector 85 %

Transmission UCL glass 97,8 %

Values at 25 °C Tsp, for 85 °C Tsp values are 6 to 14 percent lower, depending on LED

LEDs marked with Warning sign uses old values are still taken with previous method (determining die size) due to problematic light distribution (donut holes etc).

Because of the weird light pattern and irregular light distribution (see also next topic Light quality) I haven't updated the values with the new testing method.





Light quality and use in optics

With reflectors the light pattern isn't nice, as expected. The reason for this are the four single dies with the noticeable gap between them, a problem already known from the first Cree XHP50/70.

The light used in this pictures is a JAXMAN M8 (formerly known as JAX Mini C8), equipped with an OP reflector.

The dark hole in the spot is not only visible on white or flat surfaces, but everywhere – especially if the light is moved!

To solve this problem the use of matte optics or foils is necessary. Then the light pattern is really nice and absolutely smooth. The donut in the spot disappears completely and the light is fully mixed. The spot goes steplessly into the corona, in this case I would categorize this beam as ‚allrounder‘.

The light quality itself can be described with one word: awesome. The tint is a clear white with no yellow, blue or green in it. Between Nichia 219BT-V1 sw57 and this Luxeon MZ tested here you can basically see no difference at all. The MZ has also a high amount of red in spectrum, which makes the color rendition (especially of red and brown tones) extremely good and natural. Overall the spectrum seems to be very similar to that of the 219BT-V1 sw57. It is important that the tint could differ between different LEDs of the same charge.

With the matte foil, a 105c and 4.5 Amps (12*7135) the so modded M8 delivers over 1000 real Lumens!

Unfortunately this LED is not available in 4000, 4500 or 5000 K CCT, but only in 2700 (QW27) and 3000 K (QW30). It would be very nice if Lumileds offers this LED in a high CRI option with 4500 or 5000 K to establish a serious competitor for Nichias well-known high CRI LEDs like 219BT-V1 or 219C and Osrams CQAR.CC.



Conclusion

This LED is very interesting and now my absolute favourite LED. Nearly 3000 Lumens in High CRI in (almost) XP footprint and 3 V voltage, with high efficiency and extremely good light quality. Because of the parallel connected dies the Vf is very low, and the efficiency is (for high CRI) very high despite missing dome.

I did not like the big gaps between the single dies and the bad light pattern resulting from this, if clear optics / lenses are used – matte lenses are mandatory. Also the unique footprint is a kind of problem because accessoires for XP LEDs are not usable easy.

But – this LED is very unique and a very good choice for everyone which likes cold white light with extreme high color rendition AND high power. For those who needs more efficiency in high CRI the domed Luxeon MX in same color and tint might be an option (test coming soon), even if the problem with the quality of light is likely to continue.

Pro

  • high flux possible
  • Very low Vf
  • Relatively efficient
  • Very good light quality and color rendition
  • beautiful tint
  • XP board usable

Contra

  • Special footprint
  • No symmetric shape
  • bad beam pattern if clear optics used

Thanks a lot for reading! :)

Greetings, Dominik (aka BLF member koef3)



Mistakes, suggestions or offers / sponsoring of LEDs which I should test next are best sent via PM.

2

Very interesting, thanks.

3

Interesting LED, but it seems like its power hungry! 9A for (only) 2,173.6 lumens…
Yes, CRI makes a difference but still, 16A for just under 3000 lumens sounds like wasting power!

4

Interesting, thanks. Looks like between XP and XM in both size and capabilities. In the latter more than the former. However in most charts you only compare it with 3V XP sized LEDs. Maybe you could add a power vs flux chart which includes XHP50.2?

5

this LED in cri70 could be powerful but the donut hole :person_facepalming:

There is a new DTP board you can use for 4040 footprint
https://led4power.com/product/l4p-4040-16mm-dtp-direct-thermal-path-copper-mcpcb/

6

Nice work koef3. Thanks for posting it up. :+1:

7

Man I totally forgot about these LEDs, I played around with some of the non high CRI versions about a year ago, the main thing that came of it was a very silly light that ran 4 of them in a Shadow JM07. They are very power hungry and the light originally had one of them but direct drive it was pulling too much current so in a fit of late night histeria I decided I could fix that if I just added 3 more of them and ran the 26650 hard enough to sag down to a more reasonably current draw. The result is a pocket flood monster with 16 LED dies in one reflector.

8

On first graph, the darker blue curve is for the Luxeon MZ corrected to 85°C Tj.
On your second graph, the same darker blue curve is for the XP-L2 V5 40E.

Something seems wrong.

9

Very interesting…

10

how is it compared with xhp50 high cri ?
since its maximum lumen is about 2000 lumen, special footprint, low Vf( more loss power if use linear current regulator like amc7135)
i doubt it will beat xhp50 hi-cri

11

Near 100lm/W @1000 lumens with 4 dies - I think xp-g2 quad can do it better.

12

Yeah, I forgot to change the LED data in my template.
It’s fixed now. :slight_smile:

At these who calls this LED ‘power hungry’: Mind you, the MZ is connected 4P, the XHP35 HI 4S. In highest bin E2 the XHP35 HI needs 3,200 mAmps for 2,850 Lumens, at 49 Watts. For hypothetical 4S config the current can be roughly quadrupled, which then corresponds to 12.8 A. Maybe the Vf would be much lower, approx 3.60 V or so. But you see that the power hunger can be expected, especially in it’s high CRI variant.

Cree is ‘cheating’ also in cool white high CRI (to increase efficiency) because they have much more green in spectrum, unlike the MZ. In spectrum it is very similar to the Nichia 219BT-V1 sw57 without much green in it.

I’m still waiting for 5700 K min. 90 CRI XP-G2 (which doesn’t exist, for sure). So this might be not really comparable to this MZ, especially if very high light quality is desired…

13

Thanks for testing this interesting led. Good to read that Lumileds consistently delivers great tints.

A sort of Cree equivalent may be the XHP50.2 in 5700K and 90CRI, Kaidomain had a very small batch of those and I bought one that I may use for a build one of these days. I’m sure that the tint will not be great but the beam profile may make up for that.

14

I got this light from Köf3. Considering the diffusor it still throws rather far.
The tint is special. It’s an almost perfect “snow white” with hardly any discoloration (maybe slightly yellow and blue, certainly no green). It’s perfectly even across all parts of the beam (the diffusor helps here though). I have tried a Nichia 144A 5700K R9050 3-step before and it was much bluer (of course there are indiviual variations, these are just single samples).

One thing to note here is the extremely low Vf. It allows for perfectly flat regulation with linear drivers until the battery is basically empty. This makes it great for building cheap, but good high-cri lights.

The only real competitor to this LED is the Cree XP-L2 with 90CRI. It’s not available in 5700K and 90CRI and has a very yellow/green corona when used with a reflector.

15

Inspired by this test I wanted to check that 5700K high CRI XHP50.2 so I swapped it into a KDIY KF8 (it had a boring 5000K XHP50.2).

And I’m not disappointed (tint is not horribly green, pretty white but more blue and green than red, and wood tones look not as yellow as with 70 CRI leds of same colour temperature), but I am also not super-impressed (219B 5700K 92CRI has a bit nicer tint, and being a new Cree, it has some ok not too annoying yellowish corona around the hotspot). The OTF output is about 1600 lumen (pretty constant, the KD boost driver regulates quite well).

16

Looks nice, thanks for the tip.
Is this board also available in 20 mm size? My test setup doesn’t allow the use of these little 16mm boards.

@djozz: could you make some whitewall shots and maybe a little comparison to Nichia 219BT-V1 sw57? This LED sounds interesting but I have always doubt about light quality and beam pattern because it’s a new generation Cree emitter. I tested the XP-G3 5700 K min 90 CRI already and the tint was ugly green with the known discolored corona…

17

I'll have 16mm and 20mm DTP boards soon. When I was designing first batch of 4040 16mm pcbs, this is actaully LED I had in my mind, Luxeon V is released later,but conveniently has same footprint.

I have few of these LEDs in my parts for almost THREE years!, that is how old is this LED, and still in 2018 it is brightest 3V LED (except very large luminus LEDs who don't fit in most flashlights).

I did some tests with this least ~6months ago, and lumen/voltage curves are very similar. Cross size in reality doesn't look that bad,gap is smaller than XHP-50/70 but bigger than XHP-50.2/70.2.

With medium roughness orange peel reflectors (tried OP C8 reflector) cross/hole is not visible.

In don't think people realize how good this LED is, it's a basically dedomed XHP-50 with 3V drive and lowest Vf of any 3V LED - regulated high mode would be available for a long time with linear drivers with decent overall efficiency ~85%, this would be similar like having 6V LED+7A output boost converter, but at lower cost.

BTW 5000K 70CRI variant is available in T flux bin (full reel), which is about 30% higher that LED tested above, I calculated ~3500+lm at 14-15A is possible (I'm considering to buy 1 reel for some time).

Some triples and quads can have higher lumens/efficiency, but even with mild OP reflector, this LED should have much better throw.

18

Has anyone stuck triple or quad Carclo optics in front of some of these? I think that’s what I would be most interested in. Neven, are you planning on selling these LEDs on your MCPCBs? I’d be interested in trying out a quad with them.

19

Quad quad :smiley:

20

Now that’s interesting. It would have the surface luminosity of about XHP35 HI E2 with much higher output. And be available in 3/6V.