Build a light testing station??

[Mark_Paulus]

Hi,

I realize this is an ambitious project, and might never get off the ground, but....

I was just wondering what would be required to build a light/torch testing station?

Here's what I envision it providing:
for a given time period t, collect light/torch samples every n units (e.g., for a 2 hour period, collect samples every 5 seconds).
spit out the collected results in raw format.
create a graph of the collected units showing lumens over time.
Perhaps be able to graph 2 or 3 different sample sets to show differences.
Would be nice if it ran under linux, and interfaced with a standard serial port.

Maybe this exists, but I think it could be handy to show run time characteristics of different types/brands/styles of batteries, such as:
how do (3) CR123 vs (2) 18650 vs (6) AA vs (2x2) 17670 all compare in initial brightness, discharge curves, and useful illumination curves.

Using these charts, one could easily see how a torch using a type of battery might compare to a different torch using the same battery, or how different batteries affect the light output for the same torch.

Any ideas on which might be required for this type of setup?
I think some sort of photo-voltaic cell for the light measurement, some sort of voltage to digital value converter (an IC of some sort I would guess), a serial plug, and then some program on the linux server to actually query the serial port and take the readings.

If such an animal already exists, I would like to know what it is, and if not, any input would be appreciated.

(Yeah, a really dark room to do the tests would also be needed. Since I have a crawl space, that's not a problem.)

 

[Roy]

Is this what you're talking about?

 

[charliek]

I'm telling you Roy- that is sharp.....

Tell me a little about your setup.

 

[Mark_Paulus]

Roy,

If your x axis is hours, and your y axis is lumens, then that would be pretty much what kind of output I would be looking for. Now, if you could "overlay" different graphs, and/or add slightly better legend and graph title info, and then let me know how you captured the data?

 

[Roy]

The heart of my setup is a Radio Shack logging multimeter with a RS232 port. ScopeView software came with the DMM, wich creates the plot above. Untill just recently, I used a 1"x2" photocell. With that setup I was able to plot mA vs time.

Last month I purchased a Extech mdl 401021 Light Meter Adapter which takes the place of the photocell. I now plot Footcandles vs time. By putting the light one foot fromt the detector, I get a measurment of Fc @ 1 Foot which is equivalent to 1 Lux @ 1 meter.

With a bare photocell, you measure the output of the photocell in mA vs time. I currently measure Fc vs time with the new setup.

Have to remember that the older runtime plots can be compaired to the new plots on the time axis ONLY.

The runtime plots are stored in the Reviews Forum two ways. Plots only are in a Runtime Plots sticky, or search for the full repors and comments by searching on "Run-Time Plots - xxxxxxxx".

 

[Roy]

"Overlays" are done with a spreadsheet program from the text files created by the plotting software.

Look here for examples of my plots and Excell plots.

 

[charliek]

I have a photocell here - how do you set up?
I'd think you'd set up the cell for maximum reading with the beam, but not too near the max the cell will put out? I think this may be a great way to measure just how much brighter the light is with Lithium vs. alkaline...

 

[Roy]

Here's what I mean by photocell.

Attach the wires to your multimeter and put the multimeter on mA scales. I always kept the photocell to light distnce constant.


For a different approach look here.

 

[charliek]

Yeah, Roy- I have one of those (actually a couple)

I have one that pushes 2.2v when I put it under the brightest light I have (a 500k Photo flood)

you say milliamp scales are better than volts?

 

[Roy]

That's what I was told when I first started doing runtime plots...so I changed to mA.

 

[charliek]

I notice the reading fluctuates closer to farther.... is there an easy way to find the peak reading? Or should I just pick a standard distance and keep it consistent?

 

[Mark_Paulus]

My impression was that I would need to treat this as any other benchmarking exercise. Namely:

Keep the distance constant.
run more than 1 test, and either throw out the run that's wildly out of range, or average the runs.
Try to isolate external variables, such as external light.