[GB ended,discussion only] CRI > 80+ NICHIA 2000K-6500K [E21A/219B/219C/319A/144A/757GT-F1(Optisolis)]

You mean the difference between 119C and 119D? I haven’t seen that comparison in the posts above (or 119C being mentioned at all).

I thought the question was what’s new in 119D/219D over 119C/219C. But yes, 119 package is better thermally compared to 219 package on two pads board like my VR16S1. Larger thermal pad area
And probably somewhat higher current handling too

- Clemence

We’ve had a bit of confusion. :slight_smile:

But yes, your find is very interesting.

I could be wrong. If the configuration is like this: die - glass - phosphors embedded silicone, then its a total disaster. Dedoming would remove most all of the phosphor and leave us a big royal blue LED!

- Clemence

I’m surprised that 119 without dedicated thermal pad is better than 219…

You can use the webplotdigitizer to check the total surface area between 119 and 219 package. But high performance non DTP MCPCB or non isolated MCPCB must be used.

- Clemence

Another improvement in 119/219D is its thermally more stable. From –30C to 25C the output remains almost the same. Less output reduction in higher temperature too.

- Clemence

Happy surgery! :slight_smile:

UPDATE 180702:

Full range CCT (2700K, 3000K, 3500K, 4000K, 5000K, 6500K) Optisolis added to the store. All 3 steps MacAdam.

The VR21P4 (Optisolis/757 series only) and VR21SP4 (E21A only) already shipped to Indonesia today.

- Clemence

New Products:

VR21P4 (parallel only, 757/Optisolis only)

- The closest performance to DTP copper MCPCB.

- 1 Oz copper

- Diameter = 21mm

- Thickness = 1,5mm

- Compatible with Carclo 106xx

- Almost invisible (from afar and when lit) yellow marking for enhanced aesthetic

  • Special solder footprint dimensions for precise LED positioning. Very easy to solder with DIY techniques. No more overflow solder - bumps which can impede optic placement.

NOTE:
Ignore those erroneous guide boxes, they won’t affect the board/soldering performance. The LEDs self align perfectly.
This board is sure overkill for 757 LED family, the board thermal resistance is much lower than the LED thermal resistance itself. But with some scrapings here and there this board can be used for other high power (2 pads) LEDs as well.

NF2L757GT-F1 sm273 P9 Rfh00


.
.
.
.
VR21SP4 (larger version of VR16SP4)
!P7080621_DxO by Clemence Ronald, on Flickr!

New features:

- The closest performance to DTP copper MCPCB

- 1 Oz. copper

- Tighter LEDs spacings at 0,3mm

- Smooth backside

- 2,5mm holes can be tapped for M3x0,5 screws for alternate mounting options with thin heat sink plate.

- Diameter = 21mm

- Thickness = 2mm

  • Almost invisible (from afar and when lit) yellow marking for enhanced aesthetic

- Clemence

Nice! I like the good circuitry clearance around the mounting holes!

So you can use ultra low profile (thin) flat heat screw up to 5mm head diameter. Easier to countersunk too, without having to break through the super hard ceramic.

  • Clemence

Do you recommend any driver or host for the VR21P4 and VR21SP4 to be used with? I’m thinking using the D4 as host for the VR21P4 but the driver will kill it in an instant. Will the VR21SP4 fit in the Convoy S2+?

For a Convoy S2, I think, a VR16SP4 should be fine instead of VR21SP4 (didn’t try OP reflector of the XM L2 version of Convoy S2, a TIR lens will work there for sure, I think).

For a host, which is compatible with Carclo 1062X series optics (thus, VR21P4), I received a recommendation, which mentioned, that kiriba has a spacer for Convoy M1 host. Would be happier, if there would be a host with out-of-the-box support for 1062X optics and any MCPCB, which fits under it, though.

Uhmm, I’m a noob when it comes to flashlight vocabulary. What I know is driving these Optisolis beyong 500mA/ LED is not a good idea if you care about lumen maintenance. You may laugh, but I only have one Convoy 18350 and that was a Djozz-lumen calibrated flashlight that I don’t want to open.

- Clemence

Clemence, what do you mean by lumen maintenance?

Cook the LEDs too hot, and they lose efficiency over time. They “wear out”.

Correct, Djozz reported a loss of 0,5% initial output after 40 continuous minutes at 500mA.

- Clemence