Data Collection

Here's how I do it.

It is very evident I need to vacuum up some more dog hair!

The lightbox is a large cardboard box that a small RAID array arrived in. Internally it was painted with 4 coats of white gloss paint and a baffle (also painted white) was put in the middle so that no light can go directly from the entry port to the meter port, it has to bounce off the interior. A circular hole big enough to take a Maglite head was cut in the top at one end and a bit of a plastic milk container (well washed) taped over the top to support and diffuse the light. The box is never opened so it ought to stay clean though the paint will yellow with age. Once the readings start to look odd I will start with another box and try to calibrate it against the old one.

Here it is in use:

At right angles to the light port on the opposite end of the box I cut another hole large enough to take the measuring head of the light meter. It is nothing fancy - it came from DX.

http://www.dealextreme.com/details.dx/sku.20170

It is more or less permanently attached with duct tape over the meter hole. I cut cardboard flaps to exclude as much extraneous light as possible and to keep it clean when not in use.

Here's the meter port end.

The cardboard flap is normally down.

Note the extensive use of duct tape to hold the whole thing together. The lux readings are manual but since the Windows software gives me the time, it acts as the stopwatch. I do actually use a stopwatch for the 30 second and 2 minute lux readings.

The thick black and red leads are hanging off the multimeter. The thin black and red leads are connected to the multimeter and are attached internally to a cheap solar cell from DX.

http://www.dealextreme.com/details.dx/sku.12983

It claims to be a 5V 5mA device. I hate to think how much light it needs to make 5mA - I measure in microamps. The second light meter is the one I use for throw measurements. Since the solar cell is not calibrated in any way I use it only for relative measurements and do not extrapolate lux measurements from it. With a lengthy and tedious set of measurements I might be able to calibrate it but life is far too short - the solar cell probably is non-linear in its responses.

The desktop - this lot lives under my desk.

The multimeter was the cheapest USB connectable device I could find. Maplin no longer sell it. I picked one up on clearance for $20 a couple of weeks ago - this one cost quite a bit more than that. Like all multimeters it gets sold under all sorts of brands all over the world. This one is usable under Linux with the software, whose name has completely escaped my mind. I really should dig out the Linux laptop which will have no problem running two logging meters at once. The Mac variant of the Linux software is unusable on my Mac - which is what I'm typing this on. The Windows box is primarily used for running the meter software. If I could remember what the Linux software was called I could give you other names for the same meter, but my mind has blanked. I know I've written it in here somewhere in a reply to Mr. Admin.

I then export the time and current data to a spreadsheet and convert the times to elapsed time. Excel is a real pain about times that span midnight but I have workarounds. Given that the MRV SE has now been running for 22 hours on low I will have to do more massaging of the data when it goes past midnight tonight which it seems set to do.

The lux readings are done manually by turning on the luxmeter at appropriate times. I get through a lot of 9V batteries for the meters.

http://budgetlightforum.cz.cc/node/121#comment-838

Thanks Don for the pics, interesting setup!

Don, have you tried testing your box design with a laser? I'd be interested in how well such a thing works (instead of say, a real $1k sphere, lol).

I was also reading an interesting observation of spectral calibration for these light meters, and how they're often quite poor down in the blue range. I wonder if anyone knows the response for the dx cheapy. Don, if you have blue/green/red emitters (even just 5mm's), maybe you can do a quick and dirty test.

The light meter certainly is not a lab grade device, I mostly depend on it to be consistent. I only own one laser - a small pointer which I can't find just now. I think it is in my desk at work.

I suspect the numbers from a laser would be very low - they are the ultimate throwers. While it puts essentially all of its energy in a small dot, the absolute amount of energy is not large. When I find the pointer, I'll try it to see.

The spectral response I'd have expected would be biased towards the blue end of the spectrum as blue light carries more energy that red. About 2x as much if memory serves. This is why red cars fade worse than any other colour (Red colour means it absorbs blue and green light while reflecting red).

A "cheap" device that will do the metering and spectral response is around $6000 - http://www.oceanoptics.com/products/jazlight.asp

The problem with a DIY approach is that I have no way of knowing the efficiency of a given LED - I can measure the power in, but not how much of that power gets converted to light. I'd need coloured lights of known output.

Calibration is not easy.

I suspect the numbers from a laser would be very low

I'm more concerned about how well a box shape works to dispurse light evenly. The idea is if you can test it at different angles; in theory a good sphere should produce same result no matter what.

The spectral response I'd have expected would be biased towards the blue end

What I've heard is that those meters (at least proper ones) try to match the light sensitivity of human eyes instead of being broad band. Rods are centered about 500nm or green. Supposedly that means they will measure higher than expected for warmer leds than cool ones (too much blue).

I'm not sure how easy it is to get a few leds which have data sheets. I mean, the parts themselves should be cheap, but not necessarily in quantity of 1's.

I see what you mean about dispersal. I do have one baffle in place (though I would probably add another if I were building the box again) so it ought not to vary with angle, but I've never tried it. I will do when I find the laser.

Thanks for the great write-up Don. That’s some contraption you've got set up there. I hate to say it, but if ever I go down in an aircraft, I hope you're sitting beside me so we can "MacGyver" our way out of it together. I'll even bring duct tape and cardboard boxes for you. Perhaps even a cheapie eBay AAA light or two.
I noticed that you were testing out my favorite cheapie light (the MRV). Its still one of my absolute favorites, and at any price. You have certainly earned the title of distinction of "Master Flashaholic". Oh, and that’s a good thing! Have you tried testing the same lights a few times in your light box to see what kind of consistency you would get? I would imagine that the specific battery, age, state of charge and environmental variables would have a massive impact on the repeatability of your data. Even the solar panel will become less efficient with usage. Still, it all looks like quite a lot of fun (for a gadget nut). That reminds me, I still need to sell some of the stuff I have... to much of it! Thanks for all the time you spend... and being a true-blue-guru for us. Without you, we'd all be guessing and pointing fingers at each other... "Mines way brighter than yours!!!"