Hey all,

I'm an EE student, and my girlfriend's birthday is coming up soon, and I wanted to make her something special for it. She's into photography, so I want to make her something related to photography. The only things I really could come up were a DSLR monitor (kind of pricey to make), remote shutter (rather basic), or a light meter. I'm thinking I might go with the light meter for now, but we'll see how difficult it turns out to be.

I'm going to use an ATMega328 microcontroller for the logic (Essentially an arduino), and probably some sort of voltage divider using photo resistors. I'm not sure how accurate this would be, but if anyone has any better ideas, let me know . Also, should I have the sensor exposed so that it directly sees the light, or should I put it behind a diffuser?

How do light meters work, and also, what features do they typically have? And is there anything I should take into consideration when designing this? I'm not too familiar with the equipment, but I do understand the terminology and basics.

I also don't really understand the difference between how you would measure incident vs reflected light? On a light meter, is there 2 different sensors? Or is it just a difference in how you aim it? (Aimed at subject vs aimed at sky?)

And finally, what kind of math should I use? I know formulas probably exist, but anything I should look for specifically for calculating correct iso/aperture/shutter? Do light meters display optimal settings, or do they display a measurement of the light?

Thank you all in advanced!

Depending how old the light meter is, you find 3 type of sensors: Selenium, Cadmium Sulfide and Silicon Blue.

Selenium were the first, and they don't need a battery, but age and light meters with Selenium sensors are usually out of spec today - often quite badly.

CdS were popular around the 60s, they need batteries but age less, however they take a while to adjust to lower light levels - you see that on cameras with a match-needle exposure system, the needle takes a few seconds to go down from a brighter exposure.

Silicon Blue cam out around the 70-80s and those are what is common today. Normal Silicon sensors however have a spectral response that is way different than film, so for use in exposure meters they need a blue filter (hence the name Silicon Blue) to see light in a similar spectrum as film does. If you buy a Silicon sensor check its spectral response and compare it to some film.

About how to meter different light, if you measure reflected light of an Object, you can either have the bare sensor pointing out or use a lens to measure a narrowed area (those are then called spot meter). For incident light metering, a diffuser is used, either a flat disc or a half dome depending of you want sidelight to get into the calculations or not. Many light meters that offer both have a sliding dome, others like the Minolta Autometer series have interchangeable covers with pegs that tell the meter what accessory is mounted (and adjust the meterings corresponding).

Basic features of light meters is measure the light and display the light level (Exposure Value, EV) and have a rotating scale to show the shutter speed / exposure combination. More advanced one allow you to enter film speed, shutter speed or aperture (Shutter speed is more common) and output depending on that the shutter speed or aperture. Often a bar graph is included. Often you can take more readings and then average them (with all readings displayed on the bar graph).

Here are two manuals for light meters I own, the Gossen Digisix http://www.gossen-.../ba_digisix_gb.pdf (a small model that does the basic stuff you want a light meter to be able to) and a Minolta Autometer IV http://www.butkus...._auto_meter_iv.htm (old, but great features and accuracy). Maybe you find your middle ground between them.

What I see as the biggest problem is to get an accurate exposure over a wide range - the Minolta for example has a 22 stop range with 0.1 stop precision. But good luck with your project.

I've decided I'm going to use a TSL2591 light sensor from adafruit https://www.adafruit.com/products/1980. It will be much more accurate than the homemade sensor, and can measure from 188 uLux up to 88,000 Lux.

Not sure how I will calibrate it without having a light meter first, though. I'm sure I could use the readings from my Rebel T5i? (Or just borrow one from a friend).

Thank you for all of the information, it definitely will help with this project!