Shocker? Olight with SST 90...what to expect from these?

[wapkil]

we get it, your smater than us, while all your technical lingo is appreciated, itd be nice to know what you think the lumens will be on this light.

Uhm, well, I didn't want to appear smart. There is nothing "smart" in what I wrote. The symbols and the nomenclature probably may look complicated but they are not. It is also quite possible that what I wrote doesn't make much sense and someone more knowledgeable will come here to correct me

I wrote about the thermal resistance - a parameter that tells you what's the temperature difference between two things if one watt of power is generated and and transfered as heat outside. If for example it is 10 deg.C/W and 5W is dissipated, the difference will be 10[deg. C/W]*5[W]=50 deg. C. The junction to ambient thermal resistance basically tells you how hotter the insides of the LED will be than the flashlight surroundings. The LED has to be kept below some temperature for the light to operate correctly so by looking at it one can try to see how much power the LED can consume and how bright it can be.

I assumed that all the energy ends up as heat. I also completely ignored what happens to the heat when it reaches the surroundings. If the light is actively cooled it is not a problem, if it is kept in the hand some amount of heat can be dissipated but the ambient temperature will probably be ~40deg.C - 50deg.C, if it is left alone it can easily overheat. My calculations were here just to continue the heatsinking discussion started by MrGman so I thought we can start with it.

What is complicated is the actual value of the junction to ambient thermal resistance. I don't know that but I though MrGman may have some estimations after all his IS measurements, so I asked.

i dont have a clue what their talking about either. All i care about is HOW MANY LUMENS is it!

I'm just playing with numbers here so what I think may end up being way off from what is real. There are many people more qualified to answer this question. If what I wrote in my previous post is close to the truth, the light would have around 1000 LED lumens and thus around 550-700 OTF lumens, depending how well the LED is cooled. It is, btw, the light from Badbeams3's game, if OTF lumens for MC-E are assumed, although it will probably run shorter than 1 hour (but the MC-E version also will)

EDIT: To clarify, I think that if the MC-E light in the game does not overheat and can sustain 500lm, the SST-90 should be able to put out 700lm. Previously I also calculated LED lumens, not OTF, so it may end up a little better.

 

[Lightmax]

Seriously, who needs more than 1000 lumens?

Are you kidding ? You asked this type of question on this forum...........?






Lightmax !

 

[TAIGERSX]

Are you kidding ? You asked this type of question on this forum...........?






Lightmax !

He probably meant "who needs more than 1000 lumen (to go to the bathroom at night)" .

 

[MrGman]

Since I am never given the thermal resistance of junction to ambient air data its hard to calculate. For some lights I have measured actual power consumption in watts and published the lumens output and how it decays in 3 to 5 minutes time.

Nice work Wapkil, but its going to be lost here.

This SST die isn't really any brighter per watt than an MC-E from what I see, maybe slightly more. Its only real difference is that its a bigger die to spread the heat out over a larger silicon chip with no empty + space in the middle, AND very importantly is mounted on a better heat spreader to help lower the thermal resistance from the die to the heatsink and therefore the outside world so that it can be driven at higher power levels and survive if and its a big if, the final product has enough thermal mass to keep it cool.

This thing was designed to be a room light or a specialty light like an operating room light were electrical power source is a lot more than a couple of 18650 batteries and heatsinking may be a great big aluminum dish.

In a flashlight in may actually be no better than the MC-E. watt for watt to answer the original question. In a big heavy flashlight with better thermal mass it may be about 10 to 20 percent better in lumen output than the MC-E.

So if the MC-E light is putting out 1000 lumens at the LED directly, this might also put out only 1000 lumens or maybe 1200 lumens. The actual out the front lumens will probably be 20 to 30 percent less as is usual for a flashlight through a reflector and glass window or an optic and a glass window. So out the front lumens may be 840 maybe 900. We don't have one to test so its still all guestimation.

So far the triple P7 flashlights running off of 2 or 3 lithium ion C or D cells seems to be the best way to get a real honest greater than 1500 lumens out the front flashlight even after its warmed up. Bigchelis has one.
Its not a real thrower its a medium range thrower but with that kind of power it lights up the night, and more specifically my entire back yard as has been shown on my video.

 

[wapkil]

This SST die isn't really any brighter per watt than an MC-E from what I see, maybe slightly more.

[...]

So if the MC-E light is putting out 1000 lumens at the LED directly, this might also put out only 1000 lumens or maybe 1200 lumens.

I'm not sure why you say that efficacy is not really better. Of course it is not something groundbreaking but even if, as you wrote, it puts out 20% more lumens per watt, the difference is similar to that between Q3 and R2 bins in Crees. I thought it should be closer to 25% for 13W (~1000lm mark) and 40%-50% for 6W (~500lm mark). At 13W also the difference in the thermal resistance makes the LED ~30 deg. C cooler. I think it can bring something around additional 10 percentage points.

Maybe I got fooled by the parameters in the specification, but it looks like an visible improvement. Not from the brightness point of view, even at 50% more it wouldn't be much brighter to the human eye, but if it offers similar modes with longer runtime, it would look like an important difference to me. Granted, it is not 300lm/W but nevertheless seems nice.

 

[neoseikan]

Hi, MrGman. Do you have SST-50 or 90 right now?
If so, Can I send you new reflectors for test?

Since I am never given the thermal resistance of junction to ambient air data its hard to calculate. For some lights I have measured actual power consumption in watts and published the lumens output and how it decays in 3 to 5 minutes time.

Nice work Wapkil, but its going to be lost here.

This SST die isn't really any brighter per watt than an MC-E from what I see, maybe slightly more. Its only real difference is that its a bigger die to spread the heat out over a larger silicon chip with no empty + space in the middle, AND very importantly is mounted on a better heat spreader to help lower the thermal resistance from the die to the heatsink and therefore the outside world so that it can be driven at higher power levels and survive if and its a big if, the final product has enough thermal mass to keep it cool.

This thing was designed to be a room light or a specialty light like an operating room light were electrical power source is a lot more than a couple of 18650 batteries and heatsinking may be a great big aluminum dish.

In a flashlight in may actually be no better than the MC-E. watt for watt to answer the original question. In a big heavy flashlight with better thermal mass it may be about 10 to 20 percent better in lumen output than the MC-E.

So if the MC-E light is putting out 1000 lumens at the LED directly, this might also put out only 1000 lumens or maybe 1200 lumens. The actual out the front lumens will probably be 20 to 30 percent less as is usual for a flashlight through a reflector and glass window or an optic and a glass window. So out the front lumens may be 840 maybe 900. We don't have one to test so its still all guestimation.

So far the triple P7 flashlights running off of 2 or 3 lithium ion C or D cells seems to be the best way to get a real honest greater than 1500 lumens out the front flashlight even after its warmed up. Bigchelis has one.
Its not a real thrower its a medium range thrower but with that kind of power it lights up the night, and more specifically my entire back yard as has been shown on my video.

 

[easilyled]

It seems to me that the SST-90 has three advantages over the MC-E and SSC-P7.

1) Superior thermal management
2) Homogenous single-die therefore no artifacts like donut/cross in the beam
3) Greater efficiency (even if not by a huge amount)

Put all these 3 factors together and the SST-90 is clearly a superior product to either of the quad-dies.

I think that Olight should be commended for so quickly recognizing the benefits of using the Luminus SST-50/90 leds and would imagine that they probably had the power to negotiate the best bins available given the scale of their manufacturing goals.

If they have managed to design their lights with the appropriate heatsinking and current regulation to take advantage of the Luminus leds, then the resulting products should be very exciting and certainly the most cutting-edge of any mass-produced lights so far.

 

[tab665]

i guess we will have to wait on more photos and specs of this light before we can even take a stab at guessing the lumens. but basically its going to be SLIGHTLY more efficient and SLIGHTLY brighter than MC-Es. and if thats the case it might not be worth the high sticker price im sure is coming soon. then again, its just a guessing game at this point.

 

[neoseikan]

After all. The most important point to me, is a higher current. I really like it. If it can use recent reflector.

 

[LuxFAN]

I hear several comments on these posts about the SST-50 and SST-90 about how higher power means shorter runtime, higher lumens, less efficeint but also at lower power it has slightly more lumen output than MC-E and P7, more efficient, longer runtime. I think the overall idea to satisfy us critics is that the SST devices give us a large range of flexibility. If you care about a flashlight running for 50 hours then drive the SST-50 device as 1.5A or less and get 400 Lumen. I personally only use a light for short periods of time before recharging so i will go for the higher current and higher lumen output. I don't care that my batteries will only last 5 hours.

 

[recDNA]

I hear several comments on these posts about the SST-50 and SST-90 about how higher power means shorter runtime, higher lumens, less efficeint but also at lower power it has slightly more lumen output than MC-E and P7, more efficient, longer runtime. I think the overall idea to satisfy us critics is that the SST devices give us a large range of flexibility. If you care about a flashlight running for 50 hours then drive the SST-50 device as 1.5A or less and get 400 Lumen. I personally only use a light for short periods of time before recharging so i will go for the higher current and higher lumen output. I don't care that my batteries will only last 5 hours.

My concern is that the flashlight is adequately heat sinked to allow the highest power setting to run until the batteries are depleted. I'm tired of warnings that "turbo" setting may only safely used for X minutes.

 

[Badbeams3]

My concern is that the flashlight is adequately heat sinked to allow the highest power setting to run until the batteries are depleted. I'm tired of warnings that "turbo" setting may only safely used for X minutes.

I think flashlights should be more informative/user friendly. They should have a small speaker built in. "Warning thermal drop down will occur 1 minute"..."Warning.....15 seconds".

"Thermal mass cooling has been achieved...full power will be restored in 10 seconds"..."Recommend a lower setting if extended use is required"

"Batt is at 50%"..."Batt is at 25%"..."Warning low batt shut down will occur in 1 minute"..

"I love you...goodbye" :tinfoil:

 

[MrGman]

I'm not sure why you say that efficacy is not really better. Of course it is not something groundbreaking but even if, as you wrote, it puts out 20% more lumens per watt, the difference is similar to that between Q3 and R2 bins in Crees. I thought it should be closer to 25% for 13W (~1000lm mark) and 40%-50% for 6W (~500lm mark). At 13W also the difference in the thermal resistance makes the LED ~30 deg. C cooler. I think it can bring something around additional 10 percentage points.

Maybe I got fooled by the parameters in the specification, but it looks like an visible improvement. Not from the brightness point of view, even at 50% more it wouldn't be much brighter to the human eye, but if it offers similar modes with longer runtime, it would look like an important difference to me. Granted, it is not 300lm/W but nevertheless seems nice.

From what I have read off the data sheet the efficiency drops (as expected) to much lower numbers at the high current draw this thing is rated to handle. Its over 100 lumens per watt at 350mA. Its only 69 lumens per watt at 2250 lumens at 3.6V and 9 amps (32.4 watts) (right off the Luminous datasheet). And that is assuming that the flux bin output really matches the data sheet.

69 lumens per watt is not impressive. doing it all in one very large die chip with a great heat sink is nice but again, it just a bigger die not a more efficient technology. If they could pump 4.5 amps through it and get over 130 lumens per watt, I would be impressed, because that would be right around 2000 lumens for just under 16 watts of power. This chip isn't claiming it can do that.

the MC-e's because they aren't driven to those current levels actually hold higher efficiency than this I believe at their rated output levels.

So this isn't some hot new high efficiency LED that will make handheld lights better than they were before (other than removing the + in the beam). If you want 1600 lumens out the front from somthing consuming 32 plus watts that's fine but I wouldn't get excited about that myself. This thing was originally designed to be a back light for high power projectors or a room light for a house.

So in summary I don't see the efficacy as better at its intended power levels only if its substantially underdriven from what its rated for. At the 3 amp range I believe its on par or slightly better than the MC-e.

Regardless, I will certainly wait for lots of real reviews before I even get excited about something like this in a flashlight.

 

[wapkil]

From what I have read off the data sheet the efficiency drops (as expected) to much lower numbers at the high current draw this thing is rated to handle. Its over 100 lumens per watt at 350mA. Its only 69 lumens per watt at 2250 lumens at 3.6V and 9 amps (32.4 watts) (right off the Luminous datasheet). And that is assuming that the flux bin output really matches the data sheet.

[...]

the MC-e's because they aren't driven to those current levels actually hold higher efficiency than this I believe at their rated output levels.

[...]

So in summary I don't see the efficacy as better at its intended power levels only if its substantially underdriven from what its rated for. At the 3 amp range I believe its on par or slightly better than the MC-e.

Ehm, 100lm/W is not for 350mA, it's for 350mA/mm^2. This thing has 9mm^2

At 3.2A (i.e. 1000lm) SST-90 is still ~98lm/W. An MC-E at 1000lm would have efficacy of ~76lm/W so SST-90 is ~30% better. For 500lm I previously calculated that it should be ~50% better.

I agree that this is not any breakthrough. It may have larger surface. BTW, anybody knows what is an MC-E emitting area? Still it looks much better than MC-Es. At MC-E's power levels the SST-90 efficacy is substantially higher, it has much lower thermal resistance and it can accept two times more power than MC-Es, if needed.

EDIT: Ok, it's time to go to sleep, previously for a minute I've put here an incorrect correction to the math above Sorry for the confusion

 

[MrGman]

Ehm, 100lm/W is not for 350mA, it's for 350mA/mm^2. This thing has 9mm^2

At 3.2A (i.e. 1000lm) SST-90 is still ~98lm/W. An MC-E at 1000lm would have efficacy of ~76lm/W so SST-90 is ~30% better. For 500lm I previously calculated that it should be ~50% better.

I agree that this is not any breakthrough. It may have larger surface. BTW, anybody knows what is an MC-E emitting area? Still it looks much better than MC-Es. At MC-E's power levels the SST-90 efficacy is substantially higher, it has much lower thermal resistance and it can accept two times more power than MC-Es, if needed.

EDIT: Ok, it's time to go to sleep, previously for a minute I've put here an incorrect correction to the math above Sorry for the confusion

yes its per mm squared but the size of this die isn't changing and at 9 amps and the stated voltage its still only around a real 69 lumens per watt, total output for total wattage. From what I have heard of some parties doing testing their not getting more lumens out of this than driving MC-E's at the same total current/power level. Using single die to single die in the 500 lumen range that 50% efficiency improvement isn't there, and using multiple MC-E's versus one SST-90 at 9 amps, they aren't getting any improvement there either. Not in a hand held flashlight, mostly due to a lack of total heatsink mass, but that is my point.

I am not saying there is anything wrong with this chip, other than I don't see it as a great high power LED flamethrower flashlight just because it can be driven at 9 amps. At 30 watts of real consumed power (round number here), I want something that can put out a real 3000 out the front lumens and hold it, not 2250 at the LED and maybe 1600 out the front and in 2 minutes its either too hot to hold or its ramped way back.

Your math is good, not questioning that, just saying this thing just doesn't impress me as having great potential for a flashlight. The triple P7 that put out 1700 lumens and would hold 1500 lumens that bigchelis had, so far was pretty impressive. It didn't draw 32.4 watts to do it. If I remember what we saw at the time it was somewhere around 24 watts of real consumed power but I may be a little off.

If some manufacturer comes out with something that is 1800 out the front lumens and holds it for less than 24 watts of power than I would say its worth buying. Also not interested in the real cool white color temp either I want something closer to the warmer whites and those are much lower in total lumen output.

Anyway its been a good discussion and I have been learning more about high power LED chips thanks to your attention to detail, G.

 

[MattK]

Gee, I wish we had a forum specifically to discuss the merits of certain LED's.

 

[SilentK]

I think flashlights should be more informative/user friendly. They should have a small speaker built in. "Warning thermal drop down will occur 1 minute"..."Warning.....15 seconds".

"Thermal mass cooling has been achieved...full power will be restored in 10 seconds"..."Recommend a lower setting if extended use is required"

"Batt is at 50%"..."Batt is at 25%"..."Warning low batt shut down will occur in 1 minute"..

"I love you...goodbye" :tinfoil:

Hehe. I wish my light would do that. But i somehow doubt that a company such as olight would not have the sense to heatsink this thing the right way. Besides, from the photo it looks like this is a brute of a light. :devil:

 

[wapkil]

Gee, I wish we had a forum specifically to discuss the merits of certain LED's.

Yeah, it seems it drifted somehow from the topic of this particular Olight light. It's largely my fault but I'm glad I could have this discussion with MrGman. I hope others could also benefit from it. I still have doubts why it seems that in practice these LEDs don't perform as good as they should. I'm also not sure if I'm all the time really comparing apples to apples (e.g. I just noticed that the thermal resistance is given junction to case for SST and junction to solder point for Cree) but I agree this is a discussion for another section.

So back to the topic. I believe we already know practically everything that's necessary to discuss this light. No need for more information. The most important part is well known - someone already told us that it will be over twenty lumens :nana:

 

[LuxFAN]

Ehm, 100lm/W is not for 350mA, it's for 350mA/mm^2. This thing has 9mm^2

At 3.2A (i.e. 1000lm) SST-90 is still ~98lm/W. An MC-E at 1000lm would have efficacy of ~76lm/W so SST-90 is ~30% better. For 500lm I previously calculated that it should be ~50% better.

I agree that this is not any breakthrough. It may have larger surface. BTW, anybody knows what is an MC-E emitting area? Still it looks much better than MC-Es. At MC-E's power levels the SST-90 efficacy is substantially higher, it has much lower thermal resistance and it can accept two times more power than MC-Es, if needed.

EDIT: Ok, it's time to go to sleep, previously for a minute I've put here an incorrect correction to the math above Sorry for the confusion

The MC-E has 4mm^2, the SST-50 has 5mm^2 and the SST-90 has 9mm^2. According to the MC-E datasheet it is limited to a maximum current of 2.8A if you run the single die in parallel or 700mA if you run them in series. The efficiency of the SST-50 is greater at 2.8A than the MC-E because you are underdriving the device (the current density will be lower). Underdriving an LED also reduces the Vf.

 

[MattK]

The most important part is well known - someone already told us that it will be over twenty lumens :nana:

Okay over 500 perhaps - but still underdriven relatively speaking.