Incredible! Semi-Accurate Results from Free LPM?!?!

[Wolfman29]

Hey everyone. I may have just made an incredible discovery that will change the world of LPMs as we know it for hobbyists....

So I have been fiddling around with this solar cell, trying to make a decent (within 10-20% error, just for qualitative uses) LPM.

After a bunch of trial and error, I got some semblance of results...!

So first things first: I have this crappy solar panel from a long time ago, measuring in at something like 70cm^2.

Using that, I took out my solar panel outside (with a DMM attached) at highest light and measured the output current and voltage. Now, it's "common knowledge" that the sun outputs roughly 1000W/m^2 of land at highest light. So, using that and the voltage/amperage measurements, I worked out my solar-panel's efficiency to be roughly 7.8%: That is, if it was one square meter in size, it would be putting out roughly 78W of energy (at 70cm^2, it should be outputting 7W of energy, but it was only outputting slightly more than half a watt).

So that was the first step.

The next step was finding some curves that will show me how a typical solar panel reacts to different wavelengths of light. Here is what I used: What type of light cause a solar cell to work?

I tried both curves, and it seems like the second one is more accurate for my solar panel. I will get to that later.

So, I unfocused my lasers wide enough so that they covered the most area of the solar panel without losing too much energy by not hitting the panel and measured output *current*, because solar panels are current sources, not voltage sources.

My results for my three lasers were as follows:

200mW 532nm Rayfoss laser: 6mA
381m*A* LOC: 9mA
~900mA 445: 15mA

Now, there is a clear difference here, if only qualitative: the higher the power the laser, the more amperage it is producing. But how do we make this into a quantitative device?

Here's what I figured out: that second graph? It shows how much amps per watt an ideal solar panel would produce for each wavelength.

So, I did some estimating and got something like this:
660nm: .46A/W
445nm: .2A/W
532nm: .35A/W

Now, I tried it first with the first graph, but my results were so far off, and there was no reason the other one couldn't be more accurate for my panel, so I am trusting the other one is more accurate.

Anyway, what I did next was take into account the calculated efficiency of my solar panel.

I divided the found amperages by the efficiency percent to get "ideal" currents produced by panel.

So, here are those results:
Rayfoss 200mW 532: 77mA or so.
LOC: 115mA or so.
445: Roughly 192mA.

The next step was using these currents and the ratios found in the graphs for each wavelength to determine the wattage of my lasers. I divided this current by the decimal given in the graph, and the result would be the power in watts.

So, here are the final results!
Rayfoss 200mW 532: 219mW!
LOC (set at 380mA): 250mW!
445 (set at 900mA): 961mW!

How damn accurate do those seem?

It may just be a coincidence, but that is just far too unlikely... anyone else agree? Have I just stumbled upon a fantasticly cheap way to semi-quantitatively measure lasers?!
:drool:

 

[Flaminpyro]

The 445 sounds good...
The LOC can sounds ok to low...
maby the cells work beter for the 405 range ?

 

[Wolfman29]

Meh. I am not using a high-efficiency lens, just your normal Aixiz. And I have heard that at 420mA, 250mW is decent for an LOC? Maybe that's just wrong?

Everything seems right though :O Anyway, I figure these figures are within 10% of accurate, because of the estimations done. Once I get another laser diode, I will do some more tests.

 

[rangedunits]

Get jerry to investigate on this one.

 

[Benm]

I suppose its possible to get some results with a solar panel, although having to exand the beam to cover the entire thing is not very practical.

Another downside is that you have to convert the reading depending on each wavelength, and have to determine the conversion factor with lasers of known power for each wavelength.

The question that remains is, however, linearity: Does 200 mW of 532 actually produce 2 times the current 100 mW does? Or is this curve not linear at all? If this is somewhat linear you could use it to compare lasers to eachother, and perhaps also see what any modifications/upgrades actually do for performance.

 

[Lotus_Darkrose]

Even if you do find out that for example: 100mw is half the value of 200mw of a certain color, you still need an actual lpm to test it against.

Don't get me wrong, I find this experiment neat, and if you can even get within 10 percent of accurate results it would be pretty nifty.

 

[Bionic-Badger]

Really, any light/power sensor can be made into a useful LPM--if calibrated to a known source. The calibration is the bulk of what you're paying for when you get a LPM. Even the use of a thermopile over a photodiode is mostly just to avoid needing to apply a wavelength absorption constant. So it's great that you have a relatively accurate solar panel LPM, but the same thing can be applied to any light or power sensor.

 

[Cyparagon]

1000W/m² is a VERY rough estimate. It depends on way too many other variables to be considered a constant.

 

[HIMNL9]

Get a small piece of solar panel (say, 1x1 cm or similar), and close it with a tube (for take away side illumination), and ou can get some better results (always remember to not focus your beam to burn, otherwise you can destroy the cell)

Also, yes, they give some decent repetitive results, not as professional cells (and you have to get a response curve from manufacturer, for have a conversion chart), but still enough for some researches ..... remember that also the first "professional" photometers / exposimeters for photography was made with solar cells (in fact, some of them was just a solar cell, a microamperometer and a trimmer ), and they was precise enough, for the time

 

[mfo]

tl;dr

 

[BShanahan14rulz]

Shhh, don't teach the proles logic and math, what's wrong with you!

 

[Wolfman29]

Haha. The great thing about this is that I only need to do math... no calibration necessary, because it's just numbers and nothing that's not exact (for within 10% accuracy, anyway). Now I just need to actually get these lasers tested to see if I was in the ballpark.

EDIT:

@Ben and Lotus:
Yes, there is linearity, or at least there should be, based simply on the theory behind solar panels. For a given efficiency, a certain number of photons knocks out a single electron. Therefore, with more photons, more electrons are knocked out of place. Once I get another red diode or something, I will try it out, but I am fairly certain that this is how it would work.

 

[MarioMaster]

Good proof of concept and a sound project - as others have said calibration is mainly what you're paying for in an LPM. People have used LEDs in the past to measure lasers. It's not that hard to make a TEC based thermal laser sensor.

 

[Wolfman29]

Yeah. I am glad this sort of worked, I just need to make sure these measurements are accurate soon, and then I need to do a prediction test, where I twist a pot and not check the current, then measure the output, then estimate the amperage going through the thing. That should be sound evidence that this works. I may eventually streamline this and start producing photodiode LPMs, calibrated of course... but how would I measure wavelength? Hmm.

EDIT:

Or I could just make a basic one that has different buttons for common wavelengths and then calculate it from there.

 

[lasersbee]

Yeah. I am glad this sort of worked, I just need to make sure these measurements are accurate soon, and then I need to do a prediction test, where I twist a pot and not check the current, then measure the output, then estimate the amperage going through the thing. That should be sound evidence that this works. I may eventually streamline this and start producing photodiode LPMs, calibrated of course... but how would I measure wavelength? Hmm.

EDIT:

Or I could just make a basic one that has different buttons for common wavelengths and then calculate it from there.

The great thing about this is that I only need to do math... no calibration necessary

I gotta disagree with that.... you WILL need a way of
calibrating no matter what...

You measure wavelengths with a spectrometer...

Like MM said.... have you tried the DIY LED LPM...
LEDs are cheap....

BTW... the optical response curve that you linked to
in the instructables is generic and not necessarily the
same curve as the Solar cell that you are using...

The Link I gave you for a good quality optical sensor
has a data sheet and response curve for that exact
optical sensor (no guessing) that also has a much more
linear curve...

Four years ago we also thought that an optical sensor
was the way to go...
Good luck with your tests...:yh:


Jerry

 

[Wolfman29]

My solar cell is a generic solar cell. Like I said, it's not accurate, but it's consistent and, my guess is, within 10% of actual power. If I wanted something to give me accuracy to within a tenth of a milliwatt I would buy a Laserbee, but that's not what I am looking for - I am looking for something that can tell me if it's over or underspec, and a way to tell which of two different color lasers is more powerful.

And regarding the cheap solar cell you linked me to - yes, it's true, it is pretty cheap, but the point of this project was to make an LPM with decent accuracy for free, essentially, or with things I can get at RadioShack for cheap. If I bought that solar cell, why not I just buy the TEC and be done with it? The reason I have done this is because I don't have $50+shipping to spend on gear right now.

By the way, I appreciate all the feedback.