Saik SA-9 upgrade

Hi Barkuti. It would definitely give you more lumens. Flood mode will be much brighter. Since the emitter is much larger, the image of the focused emitter (zoomed in?) will be much larger, but less intense. So it will probably not throw light as far, but will light up a bigger area.

The SA-9 has a real weak link for removing heat from the pill and getting it to the flashlight body. So it's a good decision on your part to not upgrade the driver for increased the current.

Best of luck and welcome to the forum.

According to the drawings in the datasheets, to me it looks like the XM-L2 emitter is a tad smaller, doesn’t it?

Yeah, the pill in this flashlight screws and unscrews inside the head moved up and down by the focusing ring, if you understand what I say… 8) So it’s a floating thing, it isn’t firmly attached to anything (unless you focus at max, so it lands over the narrowing of the head base). The result is that the heat transfer is less than ideal. I once extracted the 16 mm round led PCB and found no thermal paste between it and the pill (!) :zipper_mouth_face: , so I added some.

Cheers

CREE XLamp XM-L2 XML2 XML Cool White LED - 16mm Board U2 Bin 1A Tint 1052 lm 5000-8300K
Cool white, 65 CRI and 412 lumen@1A

CREE XLamp XML2 XM-L2 T5 10w Neutral White Color 4000k-4500K With 16mm Star Base
Neutral white, 75 CRI and 357 lumen@1A

I lean towards the neutral white, which one would you choose?

Cheers

Yes, XM-L(2) physical size is smaller than XR-E, but the emitting surface itself is much larger. Therefore it won’t throw as well in zoomie flashlights as smaller XP-G(2), XP-E etc. Dedoming will help with the throw but tint will change to more yellow.

You should consider getting something like UF-T20. Seems to be currently best in class if upgraded with dedomed XP-G2 and 2+A driver if throw is important.

So you refer to the die size of the chip when you speak about the emitter size, the emitting surface area; I understand.
So the hotspot will look much bigger but less bright… I think I can live with that. The XM-L2 will run cooler, though.

Cheers :slight_smile:

Illustrating picture about footprint vs emitting surface (yellow area) of XR-E and XM-L (same size as XM-L2)

You will not see light output difference between cold and neutral white unless you measure the lumens. Get the more desirable tint.

Yep. I was thinking the same.
I asked the seller if the PCB of the neutral white one was round (as seen on photo) or star (what the description says).
An error here?

Cheers

Can’t tell for sure but it should be 16mm and round, because picture and headline matches, but the description may be copied from other auction.

Have you looked at Fasttech? They have quite many XM-L2 tints available in various board sizes. For example 16mm neutral white 4C tint or 16mm neutral white 3B tint

The seller answered: the photo is right, the PCB is round (good to know!) :slight_smile:
Wow! Worldwide free shipping at Fasttech? And they accept PayPal… I’ll take a look…

Cheers

I wouldn’t. So much goes into optics adjustment, which, if you swap in another LED, will not look right. Just get ready to see that much wider, fatter XM-L2 square that will be shooting at around 50 yards or so, almost wiping out any remaining throw. Beam won’t be nice and tight like an XR-E—or my personal choice, the XP-E, which seems to throw even better and is more efficient. The lumens will be there, but they won’t be put to too much of any good use unless you’re adjusted to flood. And even there, many valuable lumens are being utterly lost. I’m not a big fan of the neutral or warmer tints, but then again, you may try it and actually like it. In which case, more power to ya.

Thanks for the advice Rusty Joe, you changed my mind.
I’ve taken a look at the XP-G(2), is its emitter footprint similar to the XR-E one?

A bit of math here:

Cree XPGBWT XP-G2 R5-1A 472-Lumen 6500-7000K LED Emitter with 16mm Base The datasheet specifies 458 lumens@1.5A-85º for R5 bin… ??? If these 472 lumens are to be believed, output would be 348 lumens@1A-85º (338 if not)

Cree XP-G R5-4C 325LM 4300-4500K White LED Emitter Another R5 bin… 325 lumen??? WTF? That’s at 1A, I suppose. Adjusting datasheet values, an R5 does 405 lumens@1.5A-85º, 304 lumens@1A-85º

XR-E Q5 is 205 lumens@1A-85º, which means these XP-G(2) upgrades give 50-70% more light… OK, enough of numbers. The real question here is… Will it look good?

Cheers

Your numbers look good. I’m getting 334 lumens from an XP-G2 R5 on a fresh cell in a 501B, but it diminishes somewhat quickly, holding at around 288. With great heatsinking, you can sustain 300+ lumens from XP-G R5, and more from XP-G2 R5 rather easily. I’ve tried R5 zoomies stock and not been very impressed, though. They still don’t give nearly the same throw as XR-E or XP-E. De-doming them, on the other hand, well, that’s when things really change for the better!

Well, I’ve been reading testing two of KD’s new Nichia 219b leds (4500K’92CRI’ and 5500K) EDIT: not convinced of 92CRI, see #23 more or less.

Now I want one of those Nichia CRI92 NVSL219B-E R85 B220 LED 14.5 x 1mm Star
I love the perfect CRI of incans, and it is something I miss of leds.
BTW, they sell two variants of PCB (1mm and 1.5mm of thickness). I wonder why, anyone knows?

Cheers

Never mind what I’ve said. If the driver in the flashlight “direct drives” in high mode (see Is this safe? ), a lower Vf LED (XR-E drops 3.7V@1A) would probably get fried with too much current. So a new driver would be needed, and this flashlight has a 20 mm PCB in its hole when most drivers are 17mm in diameter… So s**t finding a proper driver… LOL

Cheers

OK, I’ve measured Vf with a freshly charged battery, on high, 3.3 Volts. I may not know how the driver works precisely, but it’s feeding the led well within specs. I’m gonna get the Nichia and see how it goes; after all, I can always reinstall the XR-E if it blows up :bigsmile:

Cheers

Hey again. Your stock driver is probably not going direct drive. If it did, your current would be much higher than 1 amp. Your Nichia should be fine. Plus, if you look at the charts, Djozz took that emitter up to 5 amps and didn't report that it died. Of course, he had it heat sinked to handle it.

I really like that emitter you are considering by the way. It may be my new favorite emitter. If you are fond of the ican tint, you may want to go with the older version (219A H1 B10, I think)sold by IS. It has a higher Vf and puts out less lumens though. But it has a wonderful tint and is confirmed 92cri. Personally, I would go with the one you are currently considering.

Finally, You may be able to drive at a higher current than 1 amp. It would seem that 1.5 amps may be fine. You could get a cheap light meter and see if output drops significantly within the first 5 minutes of being on high.

Speaking of heat sinking. You may want to consider going with copper emitter bases (Notigon or Sinkpad). You would want an "XP" type base. It would help improve getting the heat from your emitter and to those 2 small contact points on the pill. That is because copper moves heat about 2 times better than aluminum.

I have a couple clones of the light you are talking about. I want to someday look at the light to see if contact could be improved in the non-thread areas. The trick will be not getting the pill to bind up.

Yup. Djozz overdrived the led hard for the testing, so we can have a nice idea of the chip’s performance and limitations. But running the led so hard for long periods of time on a daily basis definitively can take a toll at the led’s health. Too much Vf and current and you can literally witness the led slowly dimming to death. Of course cooling matters, but even correctly cooled insane amounts of current will have a detrimental effect on the led’s life.

No, I am not fond of incan’s tint, but their uniform light emission distribution. A 92 CRI may seem cool, but will still suck at its 475nm hole and above 435nm (violets look dull).

What is “IS”? 0:) lol

I am perfectly fine with 1 amp driving for this flashlight, it’s a good balance between output and range and my Ultraf* batteries are happy. :slight_smile: Besides, the pill just gets slightly warm so the led operates efficiently.

Copper emitter bases? I think I’ll skip on that one (don’t wanna mess with reflowing things). I’ll get the fat (1.5mm) round PCB emitter, put some Artic Silver Céramique between the pill and PCB and done. The new led is more efficient, should heat up slightly less if anything (less thermal resistance). Haven’t tested it though. The only issue I may have is the PCB is just 14.5mm whereas the pill hole is 16mm (XR-E led on 16mm round PCB), but the plastic ring which surrounds the dome should keep the led centered (or so I hope).

Cheers

Sounds good. Let us know how it works out for you. I think you'll be happy with that emitter.

IS = Illumination Supply. Here is a link to the Nichia 219AT. Definitely another emitter you should check out someday. Based on what you have said, I think you will like it too. It's neutral, not Ican like (worried I mislead you on it's tint), but it is more towards ican than the emitter you are going with.

Kaidomain also has the NVSL219A-H1 B11 B11 bin, 287.5 lumens@1A-25º
Throws a little less lumens than the NVSL219B-E R85 B220 B220 bin, 310.5 lumens@1A-25º; and its Vf is higher, which could be a problem for my driver.
The lumen difference is insignificant, I know, but I can’t cope (or I shouldn’t) with the higher Vf risk.
Hence, my bet is for the NVSL219B-E.
BTW, compared to Cree leds an XR-E, these Nichias’ luminous flux is less impacted by temperature, which is good.

Cheers

isn’t that one of those twisty zoom lights.?
you need two hands to operate that light.
i would invest your money in a better light and better
batteries before swapping your led.
get a slide zoom from fasttech.
a sk-98 with a solid pill and a XM-L
and then a good battery that won’t burn your house down.