I’ll definitely consider various hosts for future rounds.
Also, I finally got the emitters in, so I should have all the parts I need now, assuming I don’t run out of wire.
thanks.
any news on this
Reflowed my first board yesterday. Got resupply of wire delivered today. Next build session will be sunday.
Someone mentioned a single emitter option, but these can barely handle an amp. I’d have to go down to 2-3 7135 chips.
As for the desk lamp idea, Clemence sells a board for 30+ optisolis emitters. Would need to find suitable driver and housing.
I’ve made my first triple. There’s tint shift with clear optic, green at the edges. Looks somewhat better than lh351d. With frosted optic, the green is mixed into the whole beam, and it looks worse. Experimenting with minus green now.
thanks.
good choice, buck driver !
PM sent.
Just wanted to pipe in… My work schedule is a little intense every other week, so I’m not always checking on the thread. I’m still interested.
Hey all, sorry for the lack of updates. Progress has been slow, but steady. I have to assemble the 3500k, 4000k, warm mix and cool mix, before making the promised color comparison.
Some findings: angular tint/temp shift decreases with color temperature. The 6500k version can use the clear optic without problems. Using the frosted optics mixes the green from the edge in everywhere, which is why I avoid it when possible. Minus green does a good, but not perfect job of fixing it. Even on the 3000k model I’ve gone with the clear optic, reminds me a bit of using incandescent. This warm option makes for the most saturated colors I’ve seen from a flashlight, the improvement over an sst-20 mix is noticeable. I also compared Optisolis 5000k to lh351d (a pinker than average sample). There the difference is more subtle. With the frosted optic, the former is barely better, with the clear optic the difference on food and foliage is there. Bananas look particularly nice 
So I’d say the cri difference vs conventional phosphors seems to also dip with temperature. Even though I have built the 6500k version, I haven’t used it for more than 2 minutes, as the blue cast is ghastly and hard on the eyes. This comes from someone who likes 2000k and deep red at night, so others probably won’t care as much.
Malkoff Junkie will get their light soon, so you will have an opinion other than mine at some point.
No need for folks to worry about their expressed interest, I will offer the lights in the order people inquired. Schedule has been cleared for builds around late April, and I will be doing at least one a day until they are all off my bench.
My Optisolis triples all have 10511 Carclo (frosted narrow spot) optics and slightly frosted lenses: What did you mod today?
The optics by themselves weren’t enough to resolve the artifacts in the beams.
Got in a SOIC8 clip, with the goal of flashing 12 group firmware. The KD buck drivers have 3/5 group old firmware, and they apparently shut off after a timeout. Malkoff Junkie found that problematic for their use. Also got in retaining rings for the remainder so I don’t have to solder in the pill. Will post updates as they come.
Only related in LEDs but I am putting together a Emisar DT8 using Optisolis LEDs, going to drive them at 5A (625mA per LED). They’re soldered offset in the stock 3535 pcb (using the thermal pad as negative), so that the light emitting center is at the center, instead of the geometric center. Hopefully 5000k looks nice, but I also got 4000k as a backup if it ends up too green.
Going to use 10613, or whatever Hank has available as the frosted optic, probably wide spot. Hopefully the beam looks nice, though I am not entirely sure how I will electrically isolate the PCB from the thermal path, maybe kapton tape?
Were you ever able to put together the other color temps? Clemence found that mixing from cool and warm versions would cause problems because of the large Vf difference, so I am especially curious about how the mixes worked.
The 5k and 65k have the same vf and mix fine. 4k and below also mix fine. Can’t go between the two groups unfortunately. I tried the emitters in a DM1.12 for a laugh, my method to isolate the thermal pad from negative didn’t work. I’ll try some Thermo-bond or similar instead of thermal paste at some point. Then reflow and restart if the short to the body persists. It looks impressive when run off the power supply, so I hope it works out.
Considering the rotated emitters on the pcb already even out the beam, I’m not sure you need to do them off center, bu every bit helps.
Yup, I tried that hoping I could get it to be rosy (mixing 6x 5000k 2x 3000k) but the difference was too big. I’ll probably have to resort to my secondary strategy of using minus-green
Since I won’t be using these PCBs with any other LEDs, my idea was to paint the edges with nail polish (including screw holes) and cover the bottom with kapton tape, or at least all the areas that might touch the thermal pad. Another idea was to paint the shelf (very thin coat) but just enough to cut the electrical connection.
It is mostly a side effect of using the 3535 pcb, but it is a happy accident. I discovered it when laying out the footprints in CAD.
Mouser sell Bergquist thermal pads, though you’ll probably need to use several because they’re not very large.
Looks like an interesting idea, sadly shipping from Mouser is very expensive internationally. I might try it if kapton tape and MX4 doesn’t work well.
The centering method you discovered is quite a happy coincidence! Wanted to chime in to advise against the wide frosted optic, it accentuates tint shift like crazy. The medium frosted optic is much better, same flood, better throw, minimal tint shift.
1 Thank
In that case, hopefully that is what I was sent. Too late to ask for it to be changed, now that it is on a plane somewhere. Worst case I can use Narrow Spot Plain and DC fix. Thanks for letting me know
Sadly Hank only had the wide spot frosted optic, hopefully it isn’t too bad. Sourcing the square carclo quads is annoying.
Update: parts arrived, light is built, beam is fine. What did you mod today? - #12085 by YBF650