Virence VS35SP36 MCPCB - cost effective (High CRI) General Lighting solution

Any suggestions for a casing that is weatherproof and can be used for 110V outdoor residential security/area lighting?

Built in dusk to dawn or motion sensor is a plus.

Thanks!

The possibilities are endless. First you can start with the driver. I like Inventronics, Meanwell, and Osram for their sophisticated options, control, and wide range of choices. But there’s a very good off the shelf ready to use (or assembled) with very small form factor very useful for DIY projects:
Nano & Micro driver
There are 100V, 120V, and 230V variants

For IP 6x rated casings you can use thick linear extruded aluminum heatsink and glass or PMMA cover bolted (with sealant) on it.

Below are my old 2016 design for a hotel’s garden lights. The design was 2 watts each (20pcs in total), centrali dimmable, photo sensor activated. All LED modules and the internal wiring permanently potted in clear optical grade silicone to withstand UV radiation. Plastic or PMMA will yellow and degrades in less than 2 years if exposed to sunlight 12 hours a day.


[Clemence]

Clemence…I’m into 3030 medium power leds recently. I need something cheap, powerful, 3V with ok light quality (say CRI90). I’ve seen that Optisolis overdrives nicely but suffers from high Vf and I’d like something more powerful…do you have experience with overdriving other 757 LEDs?

I will soon bring these mid power LED to the store. They’re currently (Nichia’s) most powerful and efficient mid power class LED. Unfortunately both are dual dies, not so good for flashlighting application:
3V class: NF2W757G-V3F1 up to R9050
6V class: NF2W757GR-V3 up to R9050

I will test the 6V GR-V3 soon. But at the moment only have 3000K R80, 3000K R9050, and 4000K R9050.

For single dies, but not quite as powerful and efficient as both above and should be flashlight friendly due to the perfectly centered die.
NFSW757H up to R9050.

EDIT: forgot to mention that they’re DIRT CHEAP! Estimated to be about $0,15/each retail (could be cheaper).
.

[Clemence]

Dual die is not a problem for me. I want flood and I want a few of them anyway. 8 per cell. Will happily wait for the results. :slight_smile:
These are not the most efficient around, at least on paper (beaten by Osram, Samsung and Everlight) and I have a hunch they may not be the most powerful either but being tested for overdriving and easily available would be a major value factor. :slight_smile:

You know what I meant. “They are the most powerful and efficient mid power LED…….(in Nichia’s 757 mid class family)”

Ah, I see, thanks for the clarification. I thought that you meant something more general.

Almost any good high power LED such as 219F, 319B, XPG3, or XPL2 will certainly beat mid power LED efficacy.
In real world application the cost of high power LED just can’t compete with mid power LED when it comes to sustained long run total system efficiency. Unless you are willing to invest at least 10x the investment cost.

Checked the Esttool and that’s not the case. This makes it even better! So the real drawback is only luminaire size and optic compatibility. This fact, makes me want to make a home lighting with 150lm/watt operating efficiency!
Check this:

[Clemence]

Medium power LEDs go to very high efficacies like 228 lm/W of CRI80 from Everlight 62-227ET/KK7D (even more when the LED is underdriven). I’m yet to see this beaten by high power ones.
According to Led4Power:

Here we have 199 CRI80 lm/W at 65 mA for NF2W757GT-V3F1. Not bad bat not special. In CRI90 the same model does 165 lm/W, compared to ~200 lm/W of Samsung and Everlight.

But in what I do now I’m interested mostly in CRI90 lm/$ vs. lm/mm² sweet spot. Efficiency matters but is secondary. Overdriven Medium power clearly dominates lm/$ while overdriven CSP high power clearly wins lm/mm². Interestingly, CSP medium power should be the sweet spot but I failed to find any really cheap yet good top-emitting CSP LEDs (top emission is important for dense packing).

So far I’m with medium power because while I’m not really happy with lm/mm², it’s good enough - and price is much lower.

Optisolis is very good and if I use it to estimate NF2W757GT-V3F1 performance I really like the results. Others could be better but reasonably-predicted beats wildly-predicted so this is my top pick so far. And learning that I can expect to see it tested was really nice. :slight_smile:

One idea that I have for a future project is 200 lm/W CRI90 lantern. 1/4-1/3 of LT1 size, same runtime. Different beam for different uses so not a direct competition though. 200 lm/W only on low, stretching that to 200 lm/W high would be in order of $100 per light.
Medium power LEDs seem like the only option to deliver this kind of efficacy.

Don’t forget to add 85% optical efficiency and 95% electrical efficiency in the equation. I always use 85% OE and 95% EE for high end design and 85% OE and 85% EE for normal design. 150lm/watt system efficiency is realistic with 170lm/watt LED level efficiency.

And with such low power density you can put your usual thermal transfer anxiety to sleep. :stuck_out_tongue:

[Clemence]

Clemence, I for one would love to see what you can do for home lighting.

I didn’t forget to take optics and driver into account. 200 lm/W end to end is at the edge of feasibility, depending mostly on how exactly do these LED behave when underdriven. I also tend to use 95% buck driver efficiency but for this light I assumed more efficient optics.
:wink:

And thermals are actually an attention-worthy deal. Mere 5% efficiency drop due to heating up is a big deal.

More on the 200 lm/W…I did the calculations yesterday and several days ago, now I write from memory so I may be off.
Samsung and Everlight do 230 lm/W at 10 mA.
With very low current you need much less diffusion. And the lantern doesn’t harve to be completely glare-free; low-glare may be OK.
So…
230 lm/W base, efficient buck driver, AR coated lightly patterned glass cover. Luminit claims their diffusers to be 95-98% transmissive depending on angle, so probably more like 96% overall. I’ve seen other makers claiming 94% and used that. So 205 lm/W. Now…

  • thermals reduce it
  • light at very wide angle which doesn’t hit the glass is a further tiny reduction
  • some of the light reflected off diffuser and which didn’t hit the glass will be recycled by white head interior or by phosphor

So…as I said, 200 lm/W is at the edge of feasibility, may or may not be doable.
Note that about 20 LEDs give nice about 100 lm when driven at 10 mA each. And 20 MP LEDs doesn’t cost arm and leg.

200lm/watt is definitely attainable these days. My last citylight project requirement was 150lm/watt with 3000K 80 CRI at system level.
Although we got beaten by Hori LED for the price, that was quite a learning process. And that has made me respect mid power LEDs more than before. I used to skip all low and mid power LED every time I browse manufacturer’s website, not anymore.

[Clemence]

Agro and clemence, I like your way of thinking. :slight_smile:
I also appreciate mid power LEDs more and more.

I thought you signed the contract with the City. So you didn’t get it after all?

Unfortunately no. :weary:

Oh no, after all that effort, maybe few hundred hours??, and you got nothing from the City?

Only small orders for mining lamps and handshakes :stuck_out_tongue: . Nothing is wasted, I learnt many things in the process.

[Clemence]

Knowledge and experience is priceless! :beer:

Well now back to work on the 4x4 and/or 3x3 E21A mcpcb and RRT01 mod :smiley: I think the mcpcb would be popular since it would be a direct swap into so many XHP70.2 lights in existence.