SST-90 questions

[Fallingwater]

I'm planning a homemade blinder using a pair of SST-90s, but there are two questions I need to address before starting the project.

The first is: just how efficient are these? I know their maximum specs are 3.9V at 9A, which makes for a rather impressive 35.1 watts of input power; how many of these watts actually get turned into light?

Second question: I know people have been direct-driving SST90s from IMR18650 cells, with the circuit stabilizing at about 10 amps on fresh cells (I imagine this doesn't last long). But: is it possible to do the same thing with lithium-polymer cells meant for RC models?
Reason I'm asking is because I can get them a fair bit cheaper than IMR18650s, and they'd fit the form factor I'm planning better, but they are commonly specced for very high C ratings. I'm worried they might overdrive SST-90s a lot harder than IMR18650s do.

Thanks :)

 

[morelightnow]

Every bit I've read says an imr will only get you a couple amps, like 5 or less. Most people use 3-4 c cells to power these. I use 4 c's in mine but they are 5 years old and put out 8 amps to the emitter. If they were new I bet it would be over 10 amps.

IMR batteries continuously drop voltage as they drain. If polymers can hold voltage closer to 4v then they will be better than imr's, but be ready to cut the circuit quickly.

 

[badtziscool]

Every bit I've read says an imr will only get you a couple amps, like 5 or less. Most people use 3-4 c cells to power these. I use 4 c's in mine but they are 5 years old and put out 8 amps to the emitter. If they were new I bet it would be over 10 amps.

IMR batteries continuously drop voltage as they drain. If polymers can hold voltage closer to 4v then they will be better than imr's, but be ready to cut the circuit quickly.

According to AW's sale page http://www.cpfmarketplace.com/mp/showthread.php?t=191277, an IMR18650 has a max discharge rate of 10c which means for that particular battery, it's safely rated at 16 amps.

 

[bshanahan14rulz]

I can get 5 amps from a regular Li-Ion...

 

[flashfiend]

I believe the sst-90 is best direct driven on 4 NiMh batteries. The sag characteristics of this chemistry allow the LED to run at an optimal voltage and current. A single Li-Ion works well also but if you go IMR you might zap the LED if the LED you happen to get has a lower Vf than spec. LiPo batteries might be a good choice if you decide to regulate this with a driver such as the TaskLED H6Flex or DerWichtel 9A model.

 

[morelightnow]

I believe the sst-90 is best direct driven on 4 NiMh batteries. The sag characteristics of this chemistry allow the LED to run at an optimal voltage and current. A single Li-Ion works well also but if you go IMR you might zap the LED if the LED you happen to get has a lower Vf than spec. LiPo batteries might be a good choice if you decide to regulate this with a driver such as the TaskLED H6Flex or DerWichtel 9A model.

Good NiMh cells have very little voltage sag. As far as I know the only thing better is A123 cells.

It depends on which NiMh batteries you use. Many people on here have fried the emitter with 4 D size before they had time to shut the light off. My 4 C cells are old and abused and therefore don't put out the current they did when new, but still push the emitter at 8 amps direct drive.

 

[Starlight]

No one as yet has answered his question about lithium-polymer batteries. He wants to know how many amps will an SST-90 draw from a Li-Po. I can't answer about the SST-90, but I am familiar with the Li-Po's. I have one 11.1v battery that is rated at over 150 amps. I know that 150 amps will poof an SST-90, the question is how many amps can it draw, if it's not regulated.

 

[Al Combs]

+1 to flashfiend's recommendation of a Der Wichtel 9A regulator. You could put the two SST-90's in series and drive them from 3 Li-Po's off the same regulator. I don't know how an SST-90 would behave in a direct drive setup from a single Li-Po. My guess would be the same as an IMR-26650. Initially under driven but as it warms up would depend greatly on heat management. With the regulator you could safely get full power all the time.

As far as the SST-90's efficiency, a pixel interpolation from the Luminus "Relative Luminous Flux vs. Forward Current" graph on page 9 of the SST-90 doc's, gives ≈ 272% increase of output at 9 amps relative to the 3.2 amp base value. The WN bin's normal range of 1000 - 1200 lumens @ 3.2 amps equals 2720 - 3264 lumens @ 9 amps. That's a range of 77.5 up to 93.0 lumens per watt. So roughly as efficient as a P7 but MUCH brighter.

 

[CKOD]

+1 to flashfiend's recommendation of a Der Wichtel 9A regulator. You could put the two SST-90's in series and drive them from 3 Li-Po's off the same regulator. I don't know how an SST-90 would behave in a direct drive setup from a single Li-Po. My guess would be the same as an IMR-26650. Initially under driven but as it warms up would depend greatly on heat management. With the regulator you could safely get full power all the time.

As far as the SST-90's efficiency, a pixel interpolation from the Luminus "Relative Luminous Flux vs. Forward Current" graph on page 9 of the SST-90 doc's, gives ≈ 272% increase of output at 9 amps relative to the 3.2 amp base value. The WN bin's normal range of 1000 - 1200 lumens @ 3.2 amps equals 2720 - 3264 lumens @ 9 amps. That's a range of 77.5 up to 93.0 lumens per watt. So roughly as efficient as a P7 but MUCH brighter.

+1 on the Der Wichtel driver also. I'm working with a friend and building an SST-90 flashlight in his old 3D mag with a DW driver, we have it build but are waiting on batteries for it. We've tested it out with a 12v SLA and its a thing of beauty. Its going to be powered off of either 3 LiFe D cells or 2 LiFe F cells. Preferably the F cells if the regulator cuts out nicely at a voltage that isnt too high. If it does play nicely with the F cells, then it'll cut off before excessively draining the cells, and being regulated so you dont have to worry about poofing the emitter with cells fresh off the charger or too new or emitter too hot or something silly like that.

I'm building my own using a H6Flex, but thats only 6.6 amps, not running it full bore, but its only a 2x 18650 flashlight and I didnt want to make something with a stupid low run time because I had to use IMR cells

 

[Fallingwater]

Thanks for the ideas. The project is stalled, but I'm still gathering information in hope of eventually building it.

The Der Witchel driver is nice, but quite expensive. Still, I like the high efficiency figure... I'll give it a thought.

As far as the SST-90's efficiency, a pixel interpolation from the Luminus "Relative Luminous Flux vs. Forward Current" graph on page 9 of the SST-90 doc's, gives ≈ 272% increase of output at 9 amps relative to the 3.2 amp base value. The WN bin's normal range of 1000 - 1200 lumens @ 3.2 amps equals 2720 - 3264 lumens @ 9 amps. That's a range of 77.5 up to 93.0 lumens per watt. So roughly as efficient as a P7 but MUCH brighter.

Hmm. I'm sorry if this is a stupid question, but... of the 35 input watts for a SST-90, how many will have to be sunk as heat?

Another question: I have to consider the possibility that it'll prove unfeasible to efficienty sink the heat from two SST-90s in the body I have in mind, forcing me to reduce the design. Given a 9-10A current source, would it be better to use one SST-90 or two SST-50s in parallel?

 

[Al Combs]

The Der Witchel driver is nice, but quite expensive. Still, I like the high efficiency figure... I'll give it a thought.

Have you seen the DIWdiver 10 amp linear regulator? It's not as efficient as Der Wichtel's and you're pretty much limited to 4*NiMH batteries. But it is less expensive. I have an SST-90 MagLite with this regulator that works great.

Hmm. I'm sorry if this is a stupid question, but... of the 35 input watts for a SST-90, how many will have to be sunk as heat?

It's not a stupid question at all. In fact I've spent a good deal of time looking for an answer to that question in general, not about the SST-90 in particular. I'm still not satisfied with what I've been able to find. That said, I've seen several people mention that 300 lumens per watt efficacy is about the limit for a white LED. Which means an SST-90-W is wasting ≈75% of its 35 watts as heat. That sounds terrible until you stop to think a regular incandescent light bulb is about 2½% efficient.

If you're really curious about the concept beyond just figuring out how much excess heat you have to deal with in the case of your SST-90, here is a chart showing the human eye's sensitivity to various wavelengths of light termed luminous efficacy. Here is another decent link on luminous efficacy from Wiki. As I understand it and trust me when I say that's making quite an allowance, in manufacture a device called a spectrophotometer can measure the radiant flux of a particular LED at different wavelengths in milliwatts. The data sheets you download from LED manufacturers have a graph of that data called, "Relative Spectral Power Distribution". It's the one that has the sharp peak at 460 nm. Except they know the exact levels, not just the relative values. Those intensity levels have to be correlated against the luminous efficacy table to get an adjusted output level. That adjusted level is the total output in lumens. I don't know how the data is actually compiled but in principle the two graphs, luminous efficacy and spectral power distribution, are superimposed to find the area under the resulting curve. As I said, it's not a stupid question at all. I'm just not smart enough to answer it.

Another question: I have to consider the possibility that it'll prove unfeasible to efficienty sink the heat from two SST-90s in the body I have in mind, forcing me to reduce the design. Given a 9-10A current source, would it be better to use one SST-90 or two SST-50s in parallel?

The data sheet suggests the output of the SST-90 in terms of its surface area is the same as the SST-50. So two SST-50's at full power would have greater lumen output than a single SST-90. However, It's never a good idea to run 2 LED's in parallel. Since voltage is the same for elements in parallel and current floats, inevitably the LED with the lower Vf is going to draw more power than the other causing problems. In series two SST-50's would draw the same current. Also it would be easier to find a 5 amp constant current regulator than a 10 amp to wire them in parallel. If the choice was between a pair of SST-50's or a single SST-90, I would go with the single SST-90.

 

[Fallingwater]

A question on a different but related project (the main one is still in development hell - the circuitry should be OK, but the physical construction is proving troublesome as I don't have access to machining tools).

What sort of performance could I expect from direct-driving a SST-90 from four Eneloop AAs (probably soldered together to avoid the holder melting down)? I know runtimes would be pathetically low, but it would not be the main mode of operation of the light. From what I remember Eneloops are relatively tolerant of horribly high draws considering their size - SilverFox tested them at 10 amps and found a 2.2Ah AA delivered some 60% of its capacity at that high a draw before sagging into uselessness. Main worry is still to avoid poofing the emitter with fully charged cells.

 

[Lumencraft (Matt)]

Driving an SST 90 that hard with eneloops is going to very hard on the cells but if you dont mind that it can work.

The actual current is going to vary with each emitter. Typically in my experience cool SSTs have a higher VF than neutrals. If you use a cool the drive current likely be any where from about 10-16 amp out the gate.

The best way to find out what the current of your emitter is would be to heat sink the emitter and hook up the leads. Take the 4 cell pack ( after letting it rest at least a few hours off the charger) and hook it up with the DMM in the circuit.

You need either a very good DMM or one with SUPER thick home made leads to cut down on the internal resistance of the meter. With that in place what I always do when testing an unknown drive current directly is put a tall stack of magnets on the neg end of the cells the first time I touch the lead to it. Then one you see what the current is just slide the lead down the stack of magnets closer to the end of the battery ( like a variable resistor). This way if at any point the current gets too high ( above 13 amp IMO) you can just disconnect.

 

[Fallingwater]

Driving an SST 90 that hard with eneloops is going to very hard on the cells but if you dont mind that it can work.

As I said, it wouldn't be the primary use of the light, so I think the eneloops should be able to take the occasional beating without dying too quickly. I'm figuring something with 2-3A regulation or thereabouts, with a low mode (perhaps using a P7 driver), but with a switch that would hook the emitter straight to the batteries, bypassing the driver, for those moments when you want a tiny lighthouse in the palm of your hand. There would also be a tiny fan (I'm thinking pager motor), stopped (or running at very reduced power) in normal mode, but when the "OW MY EYES" switch is pushed it spins up and hopefully provides enough heat removal to use a heatsink that would otherwise be hopelessly undersized (I'm thinking northbridge heatsink from a computer motherboard). I figure this should give it a minute or two or runtime at full power before heat becomes a problem.

I'm thinking of a square-shaped light, relatively long but not too thick, pocketable if you have large pockets. I'm essentially using the idea of the current XM-L "pocket rockets", but tripling the power - hence the need for active heatsinking.

Typically in my experience cool SSTs have a higher VF than neutrals.

Great - I've always liked cool emitter more anyway.