LPM Response Time - Who needs it?

[Trevor]

Hehe, watched pot boiling. That's pretty much par for the course when you're trying to datalog random events >.> .

Yep, sounds good.

edit: Reason why I suggest a more realistic test is in yours it stays between 20mW and 55mW, and all you see is an averaged output that is actually close (in terms of delta from peak) to the actual output. In real dpss units this is usually not the case. You generally get a wildly erratic stepped response in the cold phase where output jumps between discrete levels ranging from 1-100% peak output. Then as it warms it will become less erratic while gradually increasing in output again along discrete (but not linear) levels approaching the most stable high output level. It will sit there with minor fluctuations until the crystal set heats further and moves out of the optimal thermal range, and you the inverse effect of the warming phase.

If you could make a crude MSPaint drawing of the pattern you want me to test, I can do exactly what you say.

ARG seems to have a great first world problem... "my laser is too stable."

Trevor

 

[Sigurthr]

I might be able to do one better (it is hard to extrapolate numerical data from an unreferenced graph, esp. a poorly drawn one) if you'll do half the work (adding durations):

all levels (in mW) are random duration between 1 and 3sec, except where indicated with a * to denote 7-10sec duration
Cold Phase: 2, 4, 8, 6, 8, 5, 9, 7, 9, 12
Warm-Up Phase: 9, 13, 15, 16, 10, 18
"Stable" Ideal Temp: 16, 17, 18*
Over Temp Phase: 17, 18, 15, 16, 18, 14, 13, 10, 7, 6*

 

[Trevor]

Setting up to do testing now.

Trevor

 

[ARG]

Do you have mid range response LPM's to test against?
Would be interested to see that (if possible)

 

[Trevor]

Time for more data! I set up a test where I cycled through 30 predetermined power levels in 60 seconds. From the data collected, I've generated the graph below.

The true selected power value is in green.

The Ophir 20C-A follows along the true power quite nicely; it lags a little bit, but shows the actual output of the laser quite well.

The LaserBee 2.5W USB, on the other hand, doesn't quite warm up enough to hit the peaks, and doesn't quite cool off enough to show the lower readings.

This got me thinking, though. I went back and looked for the rated accuracy of all of the slow TEC LPM's that have gone through LPF. I couldn't find any, but for the sake of argument, we'll assume +/-5%.

So I generated a new graph, using black lines as the boundary for where the +/- 5% range is. Ideally, the sampled power of the LPM should fall between these lines.

This is what I have:

The Ophir 20C-A stayed within the range, deviating only briefly as the output value moved to the next true power. The response time of the Ophir 20C-A, therefore, does not hinder its ability to accurately read the power of unstable lasers.

The LaserBee 2.5W USB, however, only very briefly fell within the +/- 5% range twice, during the duration of the test. Due to its response time, the accuracy of its local maxima and minima readings was off by values sometimes exceeding 500%.

Here's a graph of the error, with +/- 5% acceptable range indicated in blue near the X-axis.

While LPM's with inexpensive, slower TEC sensors may be suitable for metering the power of stable diode lasers, the slow TEC will be wildly inaccurate if the output of the laser fluctuates at all.

EDIT: Oops, forgot to reply to ARG. No, I do not have a midrange LPM handy.

Trevor

 

[benmwv]

This is disappointing for me. I didnt realize how extreme the laserbee 2.5w usb muffled the readings. I have a 473nm labby thats jumps all over the placd 10-40mw. I wonder what the true peaks are?

Soon I should have a quick response data logging meter though

Great job giving all of these comparisons!

 

[cheech226]

Great job trevor, this is some of the data I wanted to see when I asked my question. So in the end we are talking about resolution. And in the end I am not sure just how much high resolution matters in the real world. Does it matter if my laser actually peaks at 3W if it is only for one second then drops to 2.9 that my laserbee shows after 10 seconds? I would not want to sell it as a 3W laser. As for dpss lasers, I can see how high resolution will indicate the lasers stability but all hand held dpss are pretty unstable as they have no cooling other than a heat sink. I am all for accuracy, but to what magnitude? I can not see the lasers instability with my eyes, unless it is of such magnitude that my laserbee 2 may show. What I would like to see is a high power green handheld measured by both means.
I think it is like comparing a 5 mega pixel photo to an 8 mp photo. I can still see the forest and the trees quite well in either pic.

 

[Atomicrox]

As much as I love your work I have to say I have mixed feelings about the conclusion you drew on your last post...

The accuracy value of any LPM is (obviously) only valid for a stable reading (which is why they are quoted for both accuracy and measurement time). Using the same line of thought someone could say the Ophir is wrong by infinite % on the 25Hz test just because it reads the average value of 37mW instead of the instantaneous value of 0mW during the off-time of the 25Hz test.

Just my 2 cents

 

[Trevor]

Great job trevor, this is some of the data I wanted to see when I asked my question. So in the end we are talking about resolution. And in the end I am not sure just how much high resolution matters in the real world. Does it matter if my laser actually peaks at 3W if it is only for one second then drops to 2.9 that my laserbee shows after 10 seconds? I would not want to sell it as a 3W laser. As for dpss lasers, I can see how high resolution will indicate the lasers stability but all hand held dpss are pretty unstable as they have no cooling other than a heat sink. I am all for accuracy, but to what magnitude? I can not see the lasers instability with my eyes, unless it is of such magnitude that my laserbee 2 may show. What I would like to see is a high power green handheld measured by both means.
I think it is like comparing a 5 mega pixel photo to an 8 mp photo. I can still see the forest and the trees quite well in either pic.

It really does come down to what you need.

If you need (or want to provide a buyer) a scientifically accurate power graph, you've got to have a fast sensors.

But if the buyer and the seller are both fine with a ballpark estimate of what the laser may be outputting - that's great!

I'd need to get my hands on one of the midrange LPM's to get an idea of how they perform, but they do perform much better than the slow sensors.

As much as I love your work I have to say I have mixed feelings about the conclusion you drew on your last post...

The accuracy value of any LPM is (obviously) only valid for a stable reading (which is why they are quoted for both accuracy and measurement time). Using the same line of thought someone could say the Ophir is wrong by infinite % on the 25Hz test just because it reads the average value of 37mW instead of the instantaneous value of 0mW during the off-time of the 25Hz test.

Just my 2 cents

I agree - both sensor types completely failed at 25Hz, but only one type failed a real world type test.

I also agree that, yes, the accuracy rating of slow TEC sensors LPM's is only valid for totally stable diode lasers after 30-45 seconds.

But that accuracy statement is, like you said, thrown out the window when dealing with unstable DPSS lasers. No seller of slow LPM's - MarioMaster, lasersbee, or ARG (he sold a couple ARGLPM TEC versions) has, to my knowledge, ever said that the stated accuracy is only true when a laser remains perfectly stable for the duration of the test.

I think that's an important caveat that no one (sellers included) has really talked about. Granted, it's an important caveat for maybe 10% of the lasers bought and sold on LPF, but it's still quite relevant - especially when a test on that 10% is lasers is completely and utterly invalid due to slow response time hampering effective accuracy.

I think it sucks for buyers that "+/- 5% accuracy" actually means "+/- 5% accuracy when measuring stable diode lasers for 30-45 seconds or greater, otherwise all bets are off." The Radiant Alpha ad never mentioned that, no LaserBee ad ever mentioned that, and no ARGLPM ad ever mentioned that. Perhaps it was because we collectively didn't know that the problem was this bad, but I think now that it's been tested people have the right to know how sensor response time affects effective accuracy - especially before making the purchase of an LPM or a laser.

Trevor