Powering/Driving a Stadium type of light

[frippe75]

So what does the commercial unit cost?

What is this going to be used for?

200,000 lumens, 2 hours run time is never going to be cheap. That's 3000 w-h of batteries, not to mention 250,000 lumens of LEDs, but the LEDs will probably be cheaper than optics and everything else.

I expect anything commercial would be quite expensive.

How many are you planning to build?

I haven't really found anything viable. Just things in the $5-10k++ range that's non-battery operated and comes on a trailer.
When I say 200.000 lumens I'm referring to raw lumens and not outputted (losses included). I'm not sure this will be enough anyway.
But now I have stopped looking because it's quite fun building.

Use case is a sport-activity in nature far away from a power-grid. Activity area ~50x50 meters. But also at least 50m away from the mast.

Plan is to build 6 of them initially and the scale up from there.
I already have the Cree CoBs so I will probably mix and match between the reflector type and the lens type lamp.
Reflector one works more like a flood-light given the amount of spill. Also tried a Fresnel type lens from Ledil but I'm not sure I trust the 90% "efficiency rating".

 

[des-con]

That was an interesting topic! Wasn't planning on have a mcu for that reason but need to understand exactly whats non-linear. Not sure if you meant that available input voltage will vary (which is will due to battery discharge) or that current draw could vary due to temperature. vf is around 3.2V * 20 = 64 V ... Don't think I will get above 48V getting me into the realms of an high voltage application as well.

Appreciate the input!!

Let's assume that you may wish to dim the array for day/night or to compensate for aging in the LEDs and diminished brightness. You could do that without a micro, but it's easy to do with one and dimming should extend the LED lifetime. Depending on how you arrange the array there are constant current drivers available, or switching regulators that can be converted to constant current. The TPS5430, ALED6000, NCL30160 or LT3783 are some although you may prefer to design your own.

 

[JustAnOldFashionedLEDGuy]

6 units - so let's call that all said and done, probably $15,000 in DIY parts give or take a few thousand (and maybe more).

That does not sound like "hobby". That sounds like a commercial application (especially with the vague - "sport-activity" in nature).

Should be up front about this. (if you are making money). People are giving their time freely.

 

[frippe75]

6 units - so let's call that all said and done, probably $15,000 in DIY parts give or take a few thousand (and maybe more).

That does not sound like "hobby". That sounds like a commercial application (especially with the vague - "sport-activity" in nature).

Should be up front about this. (if you are making money). People are giving their time freely.

@JustAnOldFashionedLEDGuy this seems to be your biggest concern here since it's your second comment of this nature.
What happened to simply being over-enthusiastic or simply getting into a subject matter way over you head. Which is really the case here.

Started counting expenditure for this en-devour and speaking to my wife at the same time. She's saying out loud that.
"If you spent even 10% of that sum on lamps I'll file for a divorce". Still married but I could end up close to 10% if I follow thru.

Here are some renderings of the Cree version with the Ledil fresnel lens. Lens was 13$ each. I bought four of them 😲💸


Expenditure disclosure: 3D printed housing $2.50, Water cooling block $6, Cree COB ~$35, Fresnel Lens $13, O-Ring string $2.
Thermal paste (already had it), PVC tubing $10

Four of them daisy chained. Bad idea since the FD models shows water temp from inlet to outlet per lamp increases 2-3 degrees.
So probably needs to be fed in parallel.

 

[frippe75]

Let's assume that you may wish to dim the array for day/night or to compensate for aging in the LEDs and diminished brightness. You could do that without a micro, but it's easy to do with one and dimming should extend the LED lifetime. Depending on how you arrange the array there are constant current drivers available, or switching regulators that can be converted to constant current. The TPS5430, ALED6000, NCL30160 or LT3783 are some although you may prefer to design your own.

The whole mast will be portable and run time is two hours and with good thermal management I expect the LEDS to last longer than I could possibly imagine. I'll sure have a look at those ICs you're referring to! But I don't think I'll take that route. But I don't expect to be dimming anything. It will probably be full-on and not enough.

Edited:
Wow that NCL30160 looks interesting! 40V in max and up to 40V out at 1.5A.
But MAX PD of 1.11 Watt... Trying to wrap my head around that one.

Doing a string of 10 LEDS... 32 Volts at 0.7A = 22.4 W.
Fully charged battery would start at 37.8V (3 x 12.6). Probably around 92-93% efficiency?
That would be about 1.6 Watt of heat, right? So that's above spec... No matter how good heat dissipation?

It would have been nice to do 10 instead of 5 since it's about $0.7

From datasheet:

But maybe the MCU with PWM could be used to keep it just a the tipping point. But that's never a good design I guess...

 

[LEDphile]

When I say 200.000 lumens I'm referring to raw lumens and not outputted (losses included). I'm not sure this will be enough anyway.
But now I have stopped looking because it's quite fun building.

Use case is a sport-activity in nature far away from a power-grid. Activity area ~50x50 meters. But also at least 50m away from the mast.

Plan is to build 6 of them initially and the scale up from there.

Before diving further into fixture design, you need to figure out exactly what sort of light levels you need, along with what sort of beam pattern based on your expected fixture locations and what sort of size you're willing to deal with.

For recreational sports fields, a light level of approximately 200 lux is the usual target, which for a 50mx50m field is 500k lumens on the field. From 6 poles, that's about 100k lumens per pole assuming some spill around the edges. Figuring about an 80% optical efficiency puts you at around 120k source lumens per pole, and about 1kW of power to the LEDs (as previously noted).

That's putting you solidly into the realm of commercial light towers as commonly used on construction sites (e.g. https://www.towerlight.com/products/light-towers/electric-line/ )

 

[frippe75]

@LEDphile that was a great link as well! Haven't found that site. But I agree it would be best to know all parameters upfront. 500k is probably not realistic money wise.

Looking at one of their most portable mast. 98kg @ 23100 lumen.
My mast (@9m) now would be about 14kg, 4x Cree lights 1kg, cables + other things 3kg + batteries ~7kg. That's more like 25 kg and somewhat more portable with a theoretical lumen output of 60.000 lumen. But only 1.6 hours of runtime. Doubling up on lumens would give me 8kg more.
Then two masts. 240.000 lumen. But if that wouldn't work I probably wouldn't pursue further.

 

[frippe75]

So I have looked into different driver ICs and got stuck on the LM3406 for driving 5 LEDs per string initally.
Documentation from TI is much better than the LED2000 in my opinion. A few more external parts... Will do some EAGLE schematic tonight (since I have a day-job as a computer consultant )

An extra expenditure disclosure. I did order my second 12V 20Ah LiFEPO4 battery today. To get a 24V source with more amps than my benchie-power-supply (12V/3A max)...

 

[frippe75]

Well... Turns out doing a 5 LED per string gets costly. Just components stack up to $7 per string @ 100 strings total. Or ~285 lumen/dollars of components. Selected the TPS92200 from TI. Problem being max output voltage is 30V which is not enough for 10 LEDS and even though it's inexpensive ($0.7) it only accounts for about 10% of the total component cost. LED2000 goes up to 40V but not sure I can get the efficiency up enough to make the heat dissipation requirements low enough to fit the spec...

Back to NCL30160 one more time
Do these calc hold up? From the datasheet I tried to guestimate what 10 LED would end up on the chart... My vF LED is still 3.2V.
And calculated heat dissipation at worst case scenario and realized I had to lower the current from 700mA -> 650mA to get within 1.1 W max heat dissipation from the IC. And thats cutting it close?

 

[DIWdiver]

If I were doing this, I think I would aim at building a 5x5 module with one regulator to run two strings of 12 LEDs. If necessary, I would make my own regulator (I've done it many times before). I might still spend $7-10 in parts for each regulator, but only once per 24 LEDs. If you wanted to keep the voltage down, you could run 4 strings of 6 LEDs instead, but supply current would double.

You could even make the module so it could handle either voltage configuration. Optimizing cost would probably mean selecting different parts depending on the supply voltage, but that wouldn't be absolutely necessary.

A linear regulator would be cheap and easy, but would not be very efficient (75-80% overall, depending on LED voltage) and would dissipate around 15W per module. A switcher could be much more efficient (95%-ish) and run cooler, but would be more complex and expensive.

Then just make as many modules as necessary.

 

[frippe75]

I think this idea is worth giving some though!

 

[HarryN]

http://taskled.com/ might have some drivers that can work for you.

He is a forum member, not sure if he has posted lately or not .

What you need to be careful of in the driver design is that the LED string Vf needs to be 2 - 3 volts at least less than the Vbat for it to work.

Your skills might allow use a an IC running it close to the data sheet max, but if you want it to be long term reliable when others are using it in the real world, then more margin is needed.