Osram has announced a 1x1mm LED that delivers 237 lm. So a 3x3 array, about the size of XP emitter would do 2133 lm minus close-spacing losses. That’s without overdriving. That tiny package would be hard to solder, but nevertheless the LED might be interesting option for flooders.
Half year after the press release there is not datasheet yet.
On a separate note a matrix of 4 Pures is about the size of a single E21A. This realization led me to comparing a Pure quad against a single E21A. At CRI80 as that’s what I started with. Unless otherwise noted all the numbers come from datasheets.
Pure is rated to sustain 700 mA a piece or 2.8A per quad. That’s twice as much as E21A
Pure has much lower Vf. 2.8V at 1.4A/quad as compared to about 3.15V per emitter at the same current
note that TA has tested E21A to have Vf lower by about 0.1V. And likely higher temperature which lowers Vf - so the numbers match.
note that at the max current of 2.8A per Pure quad Vf is still merely 2.89V.
E21 has much lower thermal resistance, 0.3K/W is incredible and honestly I find it hard to believe
At 1.4A/85°C the top bin of E21A (which should be orderable) makes about 450 lm. The top bin of Pure (which may or may not be orderable) makes 470 lm. Taking into account Vf, Pures do appear more efficient….but the top bin is unlikely to be available.
Other notes:
Being 4 times smaller enables rounder LES. F.w. quad E21A has a LES diameter of about 6.1 mm. With a LES of ~6.5mm one can have an array of 21 Pures.
With more smaller emitters the beam is going to be more uniform
Being much smaller means much harder soldering, especially of large closely spaced arrays