Minimum light bulb wattage for solar simulator

[joehoy41]

TL;DR

I want to figure what wattage of bulb to use to achieve an irradiance of 1000 W/m^2.

I am trying to design a simple homemade PV solar simulator. I have picked metal-halide lamps as my light source.

If the PV panel has an area of 1m^2 and I wanted to achieve spatial uniformity across the irradiated surface with an irradiance of 1000 W/m^2, how do I go about selecting what wattage of lamp I need?

I understand that the solar irradiance is going depend on a number of factors including the number of lamps I use within a given area and the distance of the light source from the test bed. I have read research papers on others using multiple 400W metal halide lamps to achieve the 1000 W/m^2 irradiance but I don't quite understand how that works.

 

[DaveE]

TL;DR Summary: I want to figure what wattage of bulb to use to achieve an irradiance of 1000 W/m^2.

I am trying to design a simple homemade PV solar simulator. I have picked metal-halide lamps as my light source.

If the PV panel has an area of 1m^2 and I wanted to achieve spatial uniformity across the irradiated surface with an irradiance of 1000 W/m^2, how do I go about selecting what wattage of lamp I need?

I understand that the solar irradiance is going depend on a number of factors including the number of lamps I use within a given area and the distance of the light source from the test bed. I have read research papers on others using multiple 400W metal halide lamps to achieve the 1000 W/m^2 irradiance but I don't quite understand how that works.

This is complicated by the spatial distribution of the light. Watts that don't hit your DUT don't count. Lamps usually don't radiate equally in all directions, what with reflectors and such. That's why you have a power density specification. I'm afraid you're not going to get a simple answer from us. It's a relatively complex engineering problem to sort out. I think google is the place to go, plus lots of reading. You can start here, I guess.

 

[joehoy41]

I guess my real question is do I just need to make sure that 1000W of power reaches a 1m^2 surface to achieve an irradiance of 1000/m^2? Putting aside efficiency and diffusion of light, if I had 5 lamps each producing 200W laid out evenly in a 1 meter squared area, does that achieve what I need?

 

[DaveE]

I just need to make sure that 1000W of power reaches a 1m^2 surface to achieve an irradiance of 1000/m^2?

Yes, on average.

TL;DR Summary: I want to figure what wattage of bulb to use to achieve an irradiance of 1000 W/m^2.

I wanted to achieve spatial uniformity across the irradiated surface

To what accuracy? Maybe what you want is 0.1W/cm²? IDK

if I had 5 lamps each producing 200W laid out evenly in a 1 meter squared area

That depends on the lamp.
This one?

Or this one?

 

[DaveE]

Before you can choose how many watts you need, you need to figure out how many of those "watts" actually will hit your target. Lighting up the ceiling usually doesn't count. A separate question is how uniform the illumination will be.

 

[berkeman]

TL;DR Summary: I want to figure what wattage of bulb to use to achieve an irradiance of 1000 W/m^2.

I am trying to design a simple homemade PV solar simulator. I have picked metal-halide lamps as my light source.

If the PV panel has an area of 1m^2

Can you say more about why you want to do this? How much money are you planning on spending on this setup?

 

[bob012345]

I guess my real question is do I just need to make sure that 1000W of power reaches a 1m^2 surface to achieve an irradiance of 1000/m^2? Putting aside efficiency and diffusion of light, if I had 5 lamps each producing 200W laid out evenly in a 1 meter squared area, does that achieve what I need?

I think that to get 1000W/m^2 irradiance over 1 square meter will take several KW of wall power. I would suggest using a much smaller panel or even a single solar cell.

 

[tech99]

To produce 1000 W/m2 from a source which radiates equally in all directions you need first to decide the distance of the lamp from the panel. The edges of the panel will be a little further away, but maybe a spacing of 1m is good enough for your purpose. At this spacing, the power of the lamp is evenly spread over the area of a sphere of radius 1 metre. This is 4 pi R^2 = 4 pi sq metres. So to achieve 1000 W/m2 you will need 4 pi times 1000 W = 12,400 W. As mentioned in post #7, maybe reduce the size of the panel and its distance from the lamp.

 

[DaveE]

To produce 1000 W/m2 from a source which radiates equally in all directions you need first to decide the distance of the lamp from the panel. The edges of the panel will be a little further away, but maybe a spacing of 1m is good enough for your purpose. At this spacing, the power of the lamp is evenly spread over the area of a sphere of radius 1 metre. This is 4 pi R^2 = 4 pi sq metres. So to achieve 1000 W/m2 you will need 4 pi times 1000 W = 12,400 W. As mentioned in post #7, maybe reduce the size of the panel and its distance from the lamp.

Yes. There is a conflict here with the "uniform illumination" requirement though. In this example, a crude analysis based on uniform radiation in all directions would give an intensity at the corners of 67% wrt the center. So more distance is better, but is very costly in terms of power. For example the corner illumination will increase to 89% at 2 m, but you need 4x the power to do that. That's why you often see a matrix of separate sources, I think. Reflectors are also a good idea.

So, no doubt, there is some engineering required here. But this will go nowhere without some thought about specifications.

 

[jrmichler]

If the PV panel has an area of 1m^2 and I wanted to achieve spatial uniformity across the irradiated surface with an irradiance of 1000 W/m^2, how do I go about selecting what wattage of lamp I need?

Some searching using search term solar insolation in lux found this site: https://ieee-dataport.org/open-acce...of 1 Sun,m2) equals approximately 120,000 Lux. The key sentence in the abstract is: Solar Irradiance of 1 Sun (1,000 W/m2) equals approximately 120,000 Lux.

Solar irradiance of 1000 $W/m^2$ is the generally accepted value at the Earth's surface. Therefore, the desired illumination at the PV panel is 120,000 lux. Since one lux equals one lumen per square meter, you need 120,000 lumens delivered to the PV panel. If the lights have perfect (100% reflection) reflectors, and the region between the lights has perfect reflectors, and the region between the lights and the panel has perfect reflectors, then you would need lights to deliver 120,000 lumens. In practice, that will never happen.

A quick search using search terms lighting design calculations found this site: https://www.electricaltechnology.org/2017/03/lighting-design-calculation-in-building.html. It has a good (but simplistic) discussion of how room shape and reflectance affects the amount of light that gets from the lights to the surface (PV panel in your case). You can use it to estimate the percentage of light that your bulbs will deliver to your PV panel. Choosing a random number from thin air, assume that you do an excellent job of building reflectors and test box, and get 75% of the total light on the PV panel. Then you need total bulbs delivering $120,000 / 0.75 = 160,000$ lumens. Choose a size of bulb, and divide 160,000 lumens by the lumens per bulb to get the number of bulbs needed. Then mount those bulbs in one square meter, connect to power, build the necessary reflectors, and start it up. Keep in mind that the total heat produced is equal to the total power consumed by those bulbs.

 

[Drakkith]

If the PV panel has an area of 1m^2 and I wanted to achieve spatial uniformity across the irradiated surface with an irradiance of 1000 W/m^2, how do I go about selecting what wattage of lamp I need?

Is there any reason you need such a large area for a home experiment? A smaller panel would be easier to work with and require MUCH less power. A 1 kW+ setup is pretty power hungry and generates a significant amount of heat. Also note that only about half of the Sun's radiation hitting the Earth's surface is in the form of visible light that solar panels can use.

I have picked metal-halide lamps as my light source.

Metal-halide lamps are about 25% efficient. If you're trying to get 1000 W/m² of visible light, then you'll need around 4 kW of power if you're using a 1 m² solar panel, which is a substantial amount of power (about 33 amps at the 120v of common household outlets here in the US). Cut that to 2 kW if you want approximately the power of the visible part of the Sun's spectrum.

Keep in mind that the total heat produced is equal to the total power consumed by those bulbs.

I second this. 1 kW is just a bit under the amount of power consumed by a small stovetop electric heating coil, and 2 kW is a bit under what a large coil consumes. Whatever the OP is doing, they need to be careful and make sure they don't cause an electrical problem or start a fire.

 

[sophiecentaur]

There's an a alternative to using a Solar Simulator and that's to use the Sun and a calibrated meter. There would be obvious disadvantages but at least you would know what the PVs under test are getting at any one time. A lot cheaper on fuel and capital costs and it may be worth considering.