One test of one sample, of an unspecified bin, from one AliExpress seller doesn’t seem to justify the conclusions being reached here.
The datasheet shows a range between low and high bins of ~1.6x.
I’d at least want to see how the SST-20s that Kaidomain is selling stack up.
Also, FWIW, there are copper 20mm non-DTP MCPCBs for 3535 LEDs out there. I got one when I sampled the MCPCB offerings of a similar but different AliExpress LED seller. The 5050 & 2x2 array of 3535 emitters I received from them are DTP.
Knowing it is from AliExpress I suspect (or hope) it is a non-DTP MCPCB that caused the poor performance. It really seems like we should expect analogous performance to the SST 40.
For what it’s worth I will test the ones I get from KD when they arrive.
Made the changes listed above, threw in 2 new fully charged sammy 30q’s. hit 1300 lumens on startup(forgot to turn back down the vf and amp settings from yesterday’s test) Held 1300 lm. for 20-25 seconds,Then………no light. Ended up Burning out my lil dc-dc buck converter. But the LED lives!!! Confident to say 1300 lumens is a very realistic number.
Took the same LED. Sliced the dome and put it in a u-torch UT02 with mtn fet driver, 20 awg wires, and a 26650 Efest purple fully charged, After 10 seconds 1248 LUMENs with a beat up, aliexpress 1 dollar LED that i randomly selected from the original 10.
Heres the best part. The beam is tight. Real tight.
Shadow JM26 xpg2 S42B 870 lumen, has a clean dedome, Spent a ton of time playing with the focus to get 290k lux. But have to double check that. Might be 270k.
Vs.
The UT02 has not been focused, Reflector is sitting on the mcpcb. sst-20 Sliced dome. 1248 lm. No polish. Huge corona!! And Just like the sst-40 it produces some heat!
Side notes:
Shadow JM26 has a larger reflecter (guessing 4mm) then the UT-02.
Light output was lowered to the point were it complimented beam patterns and not max output.
For max output the UT-02 will significantly over-power the Shadow JM26
The sst-20 UT02’s beam pattern and corona kindof remindes me of a Acebeam K70. So I went back home and grabbed mine.
For the wall shots i have the output lowered to compare beam patterns. The sky shots are on max output.
Same 1248 lumen sst-20 UT-02. Nothing changed except a freshly charged battery
Vs.
Acebeam K70 CW factory light. 2210 Lumens. Mine seems to be missing a couple hundred lumens
Sides notes: Wall is 40 yards away. Also, In person the hotspot of the sst20 is a tad bit smaller then they appear in the pics. Pics are still a fair representation of what’s going on.
So you moved it to a DTP MCPCB? That must be the reason for the increased output, right? Can you measure the throw of the sliced SST40 in UT02?
I think that large corona is from the lit up areas to the side of the die. I noticed this with sliced SST40s, especially if you leave the silicone on the side areas. Some light from the die reflects off the silicone/air interface near the edge and bounces around in the side areas. If you remove the silicone to the sides I’ve noticed it reduces this effect. Not sure if it would also significantly increase throw (instead of bouncing around to the side areas the light might instead reflect back to the die, increasing luminance). Not that a big corona is a bad thing, but if it increases throw significantly it might be worth it to remove the side silicone.
Correct, I reflowed onto a 28mm maxtoch, as far as that being the reason for the higher output, I really cant say for certain. It could be the cooler conditions. Or the much shorter LED wires. Or a large DTP mcpcb, Your guess is as good or even better then mine.
I actually had a sliced SST-40 and later a dedomed SST-40 in the UT02 (wasted a dozen SST-40’s until I figured how to properly dedome them). In the dedomed form It did just shy of 300k, somthing like 297k at 1850+ lumens. When it was sliced it did much much worse, I cant remember exactly, but if I was to guess, mmmmm 200k-220k
BTW, any california peeps looking to dedome SST-40’s I got the magic potion. I say california because we have to use the watered down crap called (low vac products) that are generally cat piss compared to the rest of the USA.
And you are absolutly correct about the large corona resulting in that extra side material from a sliced dome. My eyes get easilly fatigued, so trimming that side dome material on such a tiny LED is just outside my range of craftsmanship. I just started dedomong the sst-20. And daym. It’s way different then dedoming a SST-40!
Also played with centering and spacing the sst-20. The beam pattern looks much nicer now. Its the only flashlight I have that has a negative tappering beam. Looks phenomenal.
As far as lux goes, Was hoping for upper 2XX but it only came in at 216k. Alil shocking, since the beam looks so daym good.
And for the guys that asked about the dedome potion, it’s from a company called (Crown and its their Pro Thinner) the only place I can find it is from Sherman William’s retail stores.
15 mins boiling for xm-l2 and xp-g2. 30-35 mins boiling for sst-40s. it does not work on the sst-20. Or at least 50 mins of boiling doesn’t work. I’ll try one more time and boil it for 1 hour and make little cuts on the sides to help aid the seperation.
Got three today, put one in my B158B, direct drive and sliced.
I don’t have any numbers (well, none that could be trusted) but it outthrow the dedomed XP-L that was in there before.
Did a led swap just now. Took a brand new sofirn SF36 NW xpl V6. My measurement came in at 960 LUMENS Sofirn advertises 1070 lumens.
Swapped In a sst-20 using the SF36 original 10mm copper mcpcb and also by passing both springs. Nothing else was done. Using the same battery, after 15 seconds 1008 lumens.
Removed the tail, Using a short 10awg wire and unit-T clamp DDM I’m measuring 4.06 amps at tail.
Pros’s:
Much tighter and simply more fun beam.
Should now have better battery life.
Higher lux and lumens.
Con’s:
Lost the xp-l V6’s awesome neutral 4500-5000k tint.
Sst-20 tint reminds me of a 1992 honda civic with illegal HID headlights.(7000k-8000k)
How did you direct drived this SST20?
I got 3 of them. Putted two in my Fet TA driver Convoy L2 and when I switched it in turbo….instant death. I got three so I only left one. I think I will test it at work with a bench power supply how much it can take without dying. but this is sad news for me. I think it is not better than XP-G2.
