Power Meter Calibration and Comparison

[Kenom]

I'll organize the list once we get payments. We obviously won't get payments till we get a longer list of contributors. Give it some time, this has only been 2 days.

 

[pullbangdead]

As far as the temperature effects on the output power, it's actually quite interesting. If you look at the plot of input current vs output power, with overlayed plots for different case temperatures (look in the 2 data sheets I link below), the slope of the current vs. power out line above threshold doesn't change. Really, it's just shifted left or right. In essence, yes, if you take multiple power measurements at different temperatures with the same current, you will get different powers. But the slope of the line isn't changing, so it's easier to deal with. Theoretically. I'm not quite sure how to deal with it completely yet, at least not without variable but completely controllable current. For our uses, though, I don't think +/- 2 degrees C will change the output power much. On the Sharp datasheet, for instance, it looks to me like doubling the temperature from room temp at 25C to almost-too-hot-to-hold 50C changes the power about 5% or so at 100mA of current. In reality, as long as your house is climate controlled, I'd say you're going to be within at most 2 degrees C of the same as everyone else. So yes, to get a thorough completely accurate measurement, you have to take temperature and other conditions all into account.

But really, this to me is about getting your measurements to match what others see. As long as the best meter is completely accurate and precise (accurate and repeatable results) at normal room temperature, then I'd say it would be most useful to use your own meter in the same circumstances as you always do. Aka, I would say to myself "alright, this laser at normal, room temperature on an accurate and precise meter, measures xxx mW. On my meter under the typical conditions in which I use it, it measures yyy mW. Therefore, I have a correction factor of zzz". That way, the conditions under which you measure are already accounted for, and the temperature at your house is taken into account. As long as the measurement you compare yourself to is accurate at a certain temperature, and your conditions are always the same at your house, your measurement will always relate back to the conditions used in the measurement you're comparing yourself to. Make sense?

 

[pseudonomen137]

Makes sense, but still I don't know anyone on this forum with a metering setup with less than a +/-3% margin of error. So you already have that uncertainty of the reference. Then the meters its being tested on at the end will also suffer a margin of error, and temp and other uncertainties about the reference would probably add up to a few percent themselves.

Basically, I just want to remind that if we go through with it, it doesn't mean any meter tested against the reference is foolproof or entirely accurate. It means we at least know its in the right general range - not some bogus measurement, but I wouldn't feel comfortable trying to claim much less than a +/-10% margin of error. What do you guys think?

(assuming the price drops a bit, I'm in BTW)

 

[pullbangdead]

We have to assume the "most accurate" meter is always precise and accurate.  We take it as the baseline for it being the most accurate meter (I don't know yet which meter that will be), and it's easy enough to make sure it's precise, just do multiple tests over a long time period under the same conditions each time.  Every one can and should do that to get their own precision, and that level of precision, along with the precision of the standard meter, should constitute the entire error.  With good environment control, I don't really know what the best precision possible would be.

There are meters available with better than a 3% margin of error, I would guesstimate. And has anyone done a repeatability test to see how much their measurement varies by over a longer period of time?

 

[pseudonomen137]

I get what you're saying, I'm just reminding that there will be a pretty nice margin of error if we wanna be honest about this. We keep using examples like "the reference is 100mW", but we don't know that for sure, what we may be able to know is "the reference is between 97 and 103mW" (or at least with my power metering setup, that's the best I can do - can anyone do better?). So there's your +/-3% already. From there, various condition such as temp and all will add on a little, and the meter at the end will also have some accuracy/precision issues. Even if you can eliminate most of its margin of error through precision, you still need to take such things into account when coming up with your margin of error.

There are a lot of things we're counting on being accurate and precise here that really wont be quite as good as we'd like to hope. There are several steps for error to compound, and it'd just be reassuring to know no ones gunna try abusing this test to say their meter is 100% accurate or anything.


PS: About the meter margin of error, Greg and I have FieldMax II TO meters with PM3 and/or PM3Q sensors so our setups are subject to +/-3%. I don't know what electron's exact setup is, but from what he wrote I'm guessing its also a +/-3%. I don't know any more accurate setups that have been offered for use yet  :-/

 

[pullbangdead]

I'm trying to work up the nerve to see if I can use the set-up in the lab I work in on campus. If I had to guess, I'd say it's better than 3% accurate. If I can ever get up the nerve to ask to use it, I'll certainly find out all the specifics of the set-up there. And, I would be able to use the spectrometer and other equipment with that as well. I'll even see if I can get some pictures. No promises, but I'm going to try.

I figure there's a good shot they'll let me use it, since I'm hooking the lab up with a new laser for the photoluminescence test lab (their current setup uses a HeCd at 325nm, but it can't go as deep into some of the things they want to look at, so something like 405nm is very desirable for looking at PL deeper inside samples, and wouldn't you know it, I knew of a CHEAP source of 100mW of 405nm light . So both lasers will be incorporated and either one deployable). And what I want to test will be pretty much exactly the same thing as what I'm setting up for the PL.

 

[GooeyGus]

I'm in, but do these really need to be in a fancy host? We could do this MUUUCH cheaper if it's not complete fancy build. Or, we can make the fancy ones, sell them in the end, give the builder a percentage of the sales money, and split the rest between the contributors.

I mean, just throw the diode in a module with a driver attached, or put it in a box with some batteries. Also, I just skimmed through the posts really quickly so I might have missed the reasoning on this one, but why are we making two?

 

[Kenom]

hah two allows us to see both ends of the visible spectrum. 405nm and 660nm.

 

[Kenom]

another issue is how long should the laser be kept while there are others waiting to use it?? Then once the lasers have been used and there is nobody left in the queue, where does the laser go?

 

[danq]

I'm in, but do these really need to be in a fancy host?....
...I mean, just throw the diode in a module with a driver attached, or put it in a box with some batteries.

I'm with Gooey - let's just have a very basic but solid unit.
It just needs to be a diode and a driver in an "Aixiz" module... putting it in a case or box of some kind would be good, but let's keep any such case simple.

Kenom's estimate back on page 3 of $135 seems about right.

But if the builder insists on fancying it up, well... that'll be how it goes.

Either way, I'm in!


DanQ

 

[IgorT]

There should be a predetermined number of tests that each person should make, and write down the results of, to calculate the average later.

The lasers should definitelly not be used for anything but the measurements, simply to avoid power degradation from use, dust collecting on the lens or someone trying to change the fixed focus..


And as far as i understood, when the list is over, the lasers would go to the first person again, to make sure, the power remained the same. This part is definitelly important.. Either that, or it should go to the person in the middle with the most precise meter again, and then back to the first. The best thing would be, if the most precise meter was at the beginning and at the end.


About the temperature, i am not as interested in absolute precision, as i am in relative precision - us being able to compare our results.

 

[Zom-B]

I'm not so sure if just an AixiZ is a good idea. A metal host will give s fair amount of heat sinking. Operating the laser for more than a few seconds already raises the case temperature, which is effectively the same as having a different ambient temperature and a different measure length. (For example, measuring 10 seconds in 25°C equals measuring 7 seconds in 30°C (not too scale, but just to give you an idea)).

I've also never considered the way temperature effects a laser output. Because the curve shifts left and doesn't change slope, my formula is useless. I could devise a new formula for it but it would require measurements at two different points (eg 50mA and 100mA) for the same laser. This will need a different driver or a lab supply, so it is not feasible in this case.

The best way is to use the laser at a temperature as low as possible. (The graph in the Sharp data sheet is logarithmic, so lower temperatures have less mW/°C slope.) However, it is generally accepted that everything works 'as designed' at 25°C. We should also establish this measure temperature, to say, this laser is 100mW (±...%) at 25°C. This is in my country (and probably in the gross part of the world at this time) higher than average room temperature.

Just to get back at what everyone keeps saying, there should be a sheet made with strict test instructions, with temperature, duration, diameter, focus, battery, distance from sensor, zeroing, ambient sunlight (thermal sensor), and what not. I'll volunteer for this if no one else thinks he can do it better (for example, someone when an expensive meter already got it with thorough instructions. Beside that I don't think it is a bad idea to make a sort of contract, so people will not misuse the laser for things like, hey dad look what I got! *beam* *drain*.


Btw, a question for every and all contributors, Do you have thermometer near the power meter and what is its precision? And are you able, with it, to get either the laser or room to 25°C?
Depending on if there are people who can not control temperature, or measure temperature, we'll need another way of calibrating with temperature compensation.

 

[Zom-B]

This is my first draft for the instructions. Any constructive criticism or complete revisions appreciated.

(For this to work, the laser and on/off switch are preferably not mounted on the same case, but for example, at the end of a wire coming out from the laser. (to prevent misaligning the laser while switching it on/off)

Repeat the entire procedure for every color (starting with Red)

Materialistic precautions:
1. Check if the dust cap was on when receiving the laser. If it was not on, try to guess how much dust has collected. If you are unsure if it's still safe to use, return it to the maker.
2. Check if the focus ring is still fixed. Try to rotate it with minimal force, if it rotates, return it to the maker immediately.

Environmental Precautions:
1. Close the windows and doors.
2a. Work at night, or
2b. Work during the day, but find a room with minimized influence from sunlight. (eg. close sun blocking curtains and turn on as much electric light as possible)
3. Be the only one in the room to prevent body heat from others to pollute the measurements.
4. Be in the same location every time a measure is performed (changing position might pollute the measurements)

Preparations phase:
1. Use a new, freshly charged or slightly used battery. Don't use batteries that are probably less than 75% charged (for a broad safety margin)
2. Mount the sensor and laser in a stable mount (eg tripods and/or vice). This is to minimize misalignment due to vibrations.
3. Place the laser aperture 102mm (4 inch) from the meter aperture.
4a. When necessary, turn on the air conditioner to make the room temperature 25°C / 77°F, or
4b. When necessary, heat (eg. by body heat) or cool (fridge?) the laser until it is exactly 25°C / 77°F

Measurement phase:
1. Zero the meter.
2. Check the ambient or laser temperature. Restore temperature if necessary.
3. First run. Turn the laser on for 5 seconds (this is probably the time needed for the slowest meters to stabilize)
4. Check the zeroing. If it changed, zero it and repeat from 2 otherwise record the reading.
5. Check the ambient or laser temperature. Restore temperature if necessary.
6. Second run. Turn the laser on for 30 seconds
7. Check the zeroing. If it changed, zero it and repeat from 5 otherwise record the reading.
8. Check the ambient or laser temperature. Restore temperature if necessary.
9. Third run. Turn the laser on for 5 seconds
10. Check the zeroing. If it changed, zero it and repeat from 8 otherwise record the reading.

Analyze phase:
1. If the first and third measures are not equal (within meter tolerance), repeat entire Measurement phase. (and pay extra attention to Environmental Precautions)
2. Return the dust cap.
3. If laser was red, repeat from the beginning (Environmental Precautions or Preparations phase) with the violet laser.

Postprocessing phase.
1a. If you are the person with the most accurate power meter, post your second measures to the forum and skip remaining steps, or
1b. If you are not this person, wait for that person to post the second measures to the forum, and take those measures and your own second measures to calculate the correction factors.
2. (Optional) If laser was red, use one of YOUR OWN red lasers with stable output to recalibrate your meter. Recalculate the correction factor for the blue diode with the existing data (math skills required)

[edit] changes in Analyze phase/Postprocessing phase

 

[danq]

I'm not so sure if just an AixiZ is a good idea. A metal host will give s fair amount of heat sinking.

Yes, a (cheap) metal host is a good idea... would something that fits one of Jayrob's heatsinks be inexpensive? then maybe he could supply one in leu of payment?

 

[danq]

re: physical setup - am I correct in presuming that a reading will be more accurate with the beam spread out over a larger portion of the sensor, rather than focused into the smallest point?
if so, let's be sure the fixed focus makes a decent beam that fits into the smallest meter's aperture at the prescribed distance...

 

[daguin]

re: physical setup - am I correct in presuming that a reading will be more accurate with the beam spread out over a larger portion of the sensor, rather than focused into the smallest point?
if so, let's be sure the fixed focus makes a decent beam that fits into the [highlight]LASERBEE thermopile [/highlight] at the prescribed distance...

There. I fixed it for you

Peace,
dave