Regarding recent questions of LPM accuracy.

[Trevor]

Here's an abbreviated version of what's going on. I'm still quite scattered from the events of last week, so bear with me and I'll get to your questions as I can. Please don't hesitate to call me... this way you're guaranteed to get a reply. My number is <redacted>. I'm on the east coast, so just call at a reasonable hour. If I don't pick up, leave me a message and I'll get right back to you.

I've been working on a research project here at Virginia Tech - I met with a Coherent representative earlier this semester. They donated a 641nm 100mW-rated CUBE laser to me. It is brand new, perfectly calibrated, and is computer controllable.

A few weeks ago a user came to me when he was experiencing vastly different readings on his Kenomter USB vs. his Laserbee 2.5W - lasers were reading around 15% higher on the Kenometer USB. I sent him the Kenometer USB firmware update and gave him a new version of Luminosity. The issue persisted.

For the sake of science, I decided to test my four Ophir sensors. I set my CUBE to output 80mW. The results were as follows:

• 10.7% LOW
• 5.4% LOW
• 4.1% LOW
• 8.3% LOW

I was rather dismayed. This opened up the possibility of the LaserBee 2.5W LPM's to be incorrect. If we thought the Ophir sensors were reading high but they are actually reading low, then it means the LaserBees are probably reading very low.

This was rather distressing. Jerry calibrates his LPM's against a Newport LPM. They should be accurate and we assume them to be accurate. I went back and found a graph on PL that showed them to be accurate. If they are not accurate, it could mean that a lot of our benchmarks and fundamental assumptions regarding laser diodes (output in millwatts being less than input in milliamps, efficiency, etc).

However, I had not tested the user's Ophir sensor. So he shipped it to me. I once again set my CUBE to output 80mW. I fired up my Apex LPM with his Ophir sensor attached. It read 3.5mW with no power input. I moved the laser over to the sensor surface, expecting to see a high reading.

83.5mW. 3.5mW + 80mW = 83.5mW. That was NOT supposed to happen. This is BAD.

I shipped the Ophir sensor back to the user, intending to test his LaserBee next.

Last Thursday (after I'd been covering the shooting for the student paper), I got a call from him. He'd gone to test some of his lasers at a nearby university and had gotten numbers that were effectively the same as the LaserBee's.

This points to some pot in his sensor being weird. Perhaps it misreads only at high power.

And that is where we stand today.

I suspect that there is just something screwy with his sensor that throws it off at high powers. I certainly hope that is the case.

Any assistance in collecting more data would be greatly appreciated. Any insight as to what's going on with the rogue sensor would also be appreciated.

I have a high power 445nm laser here that I can use to do a differential test with, but I have not used it yet due to not having a perfect read on exactly what it's putting out.

So there you have it. My findings presented for your objective consideration.

Gentleman, what do you make of it?

Jerry: I did not want to take this into the public eye unless absolutely necessary. I still have not collected enough data to come to a conclusion. Only because it has been brought up do I post my limited findings. My apologies. I've gone about this completely objectively and hope to get it sorted out as soon as possible for the sake of the community.

Trevor

 

[Wolfman29]

Re: LPM Variance and Data Quality

Interesting. But doesn't the Kenometer Pro use a different base line depending on the power of the laser? From another thread, I am to understand that lasers under 1W will be read off of the processor's internal line, whereas lasers with a power greater than 1W will be read off of the external 5V rail. Could that not have something to do with it?

 

[Things]

Re: LPM Variance and Data Quality

Wolfman, I was actually the one that discovered we could switch the analog reference voltage as we wanted, to allow a lower power reading, and now that you mention it, it may be a valid point & something to test.

By default, the Arduino uses a 5V voltage reference. It compares the sensor readings against this. When the power reading is lower, it switches to the ATMega's internal 1.1V analog reference. The code MAY have changed this, as I wasn't really involved much more than the basic, bare level reading.

It may be worth testing whether this switch is working correctly on the Kenometers. When i first wrote that snippet of code, it worked fine, and appears to be working fine for everyone els, however maybe changing to a different form factor ATMega broke the switch in the code Or even just component tolerances?

It may also be worth looking at the real USB voltage the ATMega is getting. If it's getting under 5V, your readings will be over, and vice versa.

 

[rhd]

Re: LPM Variance and Data Quality

Trevor - didn't you run into a similar 5V USB sag issue recently with the Arduino nano?

 

[Trevor]

Re: LPM Variance and Data Quality

When running off USB power, the 5V reference sags to ~4.8V.

This would push readings high by a small amount, but not the 15% described by some. My findings pertain solely to Ophir sensors - not to Kenometers.

Trevor

 

[Things]

Re: LPM Variance and Data Quality

OK, if it's only the sensors doing it, then disregard that. I'd start looking at the power supply rails of the boost modules, and see if they're drifting from each other. what you're describing sounds like a definite power related issue to me. Maybe you could try powering one up from a computer power supply (Since they're super regulated and stable) and see if they still do it?

 

[DTR]

Re: What Venue II

I believe I know the member in question he sent me a few lasers of different wavelengths to test. Both my Ophir and Laserbee II came out to right about the same as his Laserbee but his Orphir was reading very high.

For reference here are the readings I took.

I believe he said he Ophir meetered the 445 @ like 2.4W.

 

[lasersbee]

Recent Questions of LPM accuracy... or not...

In the past few weeks questions have been raised as to the accuracy
of the LPMs being sold on LPF... Specifically the Kenometer USB which
uses an OPHIR Thermopile Head and the LaserBee 2.5W USb that uses
a Thermopile Sensor Head designed by J.BAUER Electronics.

The subject was originally posted in the vets section and a poll was
made and it was decided that any Info on this matter should be shared
with the entire Forum population...
Reference....

OP=Daguin (the Poll)
http://laserpointerforums.com/f61/what-venue-ii-69708.html

OP=Trevor (1st tests by Trevor)
http://laserpointerforums.com/f61/lpm-variance-data-quality-69745.html

Since it was decided to post any info on these issues in the general
Forum perhaps the one or both Threads could be moved to this section
of the Forum for non Veteran members to get up to speed...

==============================================
Our testing of the Kenometer USB with OPHIR Head in question.

It is to be noted that all the tests were performed on the same
equipment that we use to calibrate all our LaserBee products.


We received the Kenometer USB and OPHIR head that has caused
a questioning of the accuracy of LPMs sold on the Forum. The
questions were pertaining to the comparison of this LPM with a
budget LaserBee 2.5W USB.

1) The first thing we did was let the equipment including the K-USB
and OPHIR head acclimate to the testing area for at least an hour
before doing any tests.

We then installed the Luminosity software on a Laptop running
Vista 32bit.
We set a stable 808nm variable output Laser to 500mW on our
Newport LPM.

The first thing we noticed was the Zero input reading of 3.9 to 4.8mW
with nothing shining on the OPHIR Sensor....
Then we measured the same Laser on the Kenometer USB moving
back and forth between the Newport and K-USB to make absolutely
sure that the Laser was stable at 500mW and our readings were
as accurate as possible. We did the same test at 1000mW....

@500mW the K-USB measured between 528mW to 541mW.
@1000mW the K-USB measured betwen 1089mW to 1098mW.


2) The next test we did was to use our simple split 9volt battery power
supply circuit connected to a Fluke DMM.
We tested the OHPIR Head from the K-USB against a new untouched
OPHIR head from our stock.

We ran the same 500mW and 1000mW tests as above...

@500mW the K-USB measured between 532mV to 539mV.
@1000mW the K-USB measured between 1081mW to 1089mW. (~9%)

@500mW our OPHIR measured between 499mV to 503mV.
@1000mW our OPHIR measured between 998mW to 1008mW. (~1%)


3) we then checked the DC/DC supply of the K-USB and found it to be
ripple free and fairly stable at 11.99VDC to 12.08VDC on both the
Positive Rail and the negative rail...


4) We then checked the actual output voltage reading of the K-USB
OPHIR head while connected and running on the Luminosity software.
We needed to start and stop the Software a few times as we ran this
test.... We spent a day running and re-running the Luminosity software
and found that the link between the K-USB and Software drops out
once in a while. This could again be caused by our new Laptop running
Vista.

We were also able to read the data stream on a PIC that the Arduino
Nano was sending to the Luminosity software...
Everything seemed to match up... The DMM voltage... the PIC reading
and the Luminosity reading were identical.

5) We moved the head to a new position to take some different power
readings with some hand held Lasers and we see a reading of ~214mW
with no laser on the head... WTF....

The PIC and Luminosity were reading the same thing...
We closed the Luminosity software and restarted it.... Now the Luminosity
software shows ~3.3mW but the PIC/Arduino now reads ~209mW from
the Arduino in the K-USB.

Something seemed strange here... (see pics below)

We found a "Zero meter" option under the Meter/Tools menu and clicked it.
It only zeroed the Software but not the Arduino....

We shut every thing down and wanted to test the cable between the
K-USB and OPHIR head. While pulling on the connector the cable came
completely off.
We replaced the connector with what we had available in the shop and
have not been able to repeat that strange reading.

Doing a few more tests we found that if we put a 1.6K Ohm resistor
in line with the Negative rail we could repeat the error by unbalancing
the split rail supply.

I think that the problem was an intermittent connection on the Cable.

We wanted to test the Zeroing function of the Software...
(BTW Trevor... I see no reason that the Luminosity Software should
need to be Zeroed... It is the Arduino that needs to be Zeroed if required)
(The reasoning for this has been explained by Trevor to me in PM)

We shone an 8mW Laser on the Head and clicked Zero on Luminosity.
Then we removed the Laser and took a 50mW Laser to see if it would
read correctly.. It did not and as we removed the Laser from the sensor
the reading dropped to about 23mW and the software locked up.

We tried this Zeroing procedure quite a number of times and the same
locking up of the Luminosity Software happened every time we removed
a higher power Laser... This may again be due to our copy of Vista or
perhaps no-one else has yet stumbled on it.

6) One final test we did was to let the OPHIR Heads acclimate to the
test area temperature for 30 minutes with no Laser or power applied.

Then we took a DMM reading of the signal line from each OPHIR head
once powered by two freshly charged 9Volt batteries every 5 minutes
for 15 minutes....

KenoMeter OPHIR----------Our OHIR

Time-- mV ------------------------ mV

0 ---- -0.9 ----------------------- -1.5
5 ----- 2.0 ------------------------ 0.0
10 ---- 2.8 ------------------------ 0.3
15 ---- 3.0 ------------------------ 0.6

From this it seems that the OHIR head on the left was set to ZERO by
someone without letting the head come to equilibrium while powered
but when power was first applied... It also appears that the Gain had
also been changed. The head on the left increased by 3.9mV over 15
minutes while the pristine head on the right only increased by 2.1mV.

Conclusions...

1) We feel that the OPHIR Head that was sent reports high but still within
<10%.

2) We feel that the OPHIR Head that was sent does not Zero correctly
when equilibrium is reached and the Gain is no doubt off as well.

3) The excessively High Readings reported were no doubt due to the OPHIR
head being off calibration and not Zeroed as well as the Intermittent Head
Connector we found.

4) Why this OPHIR head is so far off is not clear... With all the similar
OPHIR 20-C heads that have passed through our shop I have never seen
one this far off... It almost seems that someone had tried to zero it and
missed the mark. I don't believe that OPHIR would send out a head that
is not calibrated correctly...

I contacted Trevor with my Luminous Software findings before posting
and he said he will look into the matter when he has a chance...
I'm sure he will clarify all that here.


Jerry

 

[daguin]

Re: Recent Questions of LPM accuracy... or not...

I'll look at the other threads to see what should be merged here

A WORD OF WARNING TO ALL--
Past discussions of LPM's have resulted in fighting and drama.
I WILL be watching this thread for problems.
If you see something that you think is off topic, trolling, or a personal attack, please REPORT the post.
DO NOT reply or argue back.

Peace,
dave

--EDIT-- I finished the "clean up" AFTER your last post, Jerry

 

[Trevor]

Re: Recent Questions of LPM accuracy... or not...

Regarding the issue Jerry brought up with Luminosity, that's a peculiarity of how "zeroing" works.

Because some of the Kenometer USB's liked to hover above zero, you could zero in the software and set an artificial "zero" level at whatever the reading was hovering at.

So, if you plugged in your meter and it was hovering at 10mW, you would zero in Luminosity and it would set the zero level at 10mW.

If this is done when there is laser input on the sensor, then it can confuse the software and will give you erroneous readings.

The fix for this would be to add a zero level reset like I did in LumenOS for the Kenometer Pro. I'm away from my development computer right now (Winter Break), but I will address this once I am back and upload a new version of Luminosity.

I've never heard of anyone else trying to zero with laser input on the sensor, but I will go ahead and make the fix to do due diligence.

Relevantly, the gain setting being off would make sense as to why it looked fine to me at low power, but ended up more and more off at higher powers. It would be interesting to graph the error as the power input increases.

Trevor

 

[lasersbee]

Re: Recent Questions of LPM accuracy... or not...

There are 2 reference points in the data above. One at 500mW and the
other at 1000mW and the error seems to be fairly linear...

EDIT...

For the sake of science, I decided to test my four Ophir sensors. I set my CUBE to output 80mW. The results were as follows:

• 10.7% LOW
• 5.4% LOW
• 4.1% LOW
• 8.3% LOW

I was rather dismayed. This opened up the possibility of the LaserBee 2.5W LPM's to be incorrect. If we thought the Ophir sensors were reading high but they are actually reading low, then it means the LaserBees are probably reading very low.

From where exactly/directly did the other OPHIR heads
that you tested come from...:thinking:


Jerry

 

[mark-in-dallas]

To note: DTR is correct and I was the person with the Ophir head that likes to underspec lasers, and by an average of 15%, but with one laser was was even much higher than that!

Just to set aside anyone thinking that I may have sold any underspeced lasers to anybody though, I have not used the Ophir readings to state power on any occasion that I've sold off any of my lasers.

I purchased the Kenometer USB and Ophir Head several months ago from nother LPF member, and when I got the LPM the first thing I did was to start measuing my laser collection. And what I found was that this LPM seemed to measure my lasers at a higher wattage than my LaserBee.

I had recently gotten my LaserBee USB back from Jerry after having the sensor reattached and recalibrated, and the readings I had been getting from it did seem a little on the low side, even factoring in a 15-20% effeciency loss, so I assumed it had to be the LaserBee and returned it to Jerry for recalibration.

Jerry contacted me soon after to say that they had tested and restested my LPM and it was accurate. At around the same time I had finished building a laser that I anticipated would hit around 2 watts or maybe a little higher, but to my amazement and disbelief the Kenometer measured it MUCH higher, at a peak of 2742mW and average reading over 1 minute of 2507mW. I then tested it on my Laserbeee and it hit and remained steady at 2076mW.

Next I built a 9V power source to test the Ophir head independently of the Kenom, to determine if the Kenom and Luminousity software were properly powering the Ophir and interpriting the output correctly. Upon connecting the Ophir to the power source and my DMM, the readings were consistant with what the Kenom read. This told me that the Kenom and Luminousity were reading and displying the correct results, but did not tell me whether they were accurate or not.

I had also been talking to Trevor about the issue and he offered to test the Ophir against his equipment. While he found it to be accurate against his LPM and reference laser, he was testing at a much lower power than I had been metering at, and I still wasn't certain where the problem lay.

I then sent the 2 Watt or much higher reading laser and along with a 635nm laser to DTR to measure on his LPMS, a LaserBee and another with an Ophir head. Both of his meters registed between 2114mW and 2020mW on the one that my LasberBee read at 2076mW, or within 2% of my LaserBee. The other laser I sent him was a 635nm that read 300mW on my LaserBee, but substantially higher on the Ophir. And again, DTR's readings were within just a few milliwatts of my LaserBee.

During the same time that my lasers were on their way to DTR I also found a local laser salvage dealer willing to et me meter a couple lasers on his Coherrent Technologies LabMaster. So, I picked 2 out of my collection and metered them on both the Kenomet and LaserBee, noting the readings I then headed out to Laser Surplus Sales and ran them on the Coherrent.

The readings from the Coherrent Techologies LPM were consistant with my LaserBee, so at that point I was relatively positive that my Ophir the where the problem lied, and when DTR sent me pics of the readings he took I was positive.

But just to remove any doubt whatsoever I sent it to Jerry to check on his equipmetn as well.

I believe that between everything that we've done here, we've proven that I just got a crappy Ophir head and this issue is not widespread. But, if there is anyone else out there that does have an Ophir heard that gives substantially higher readings than anticipated or than any other LPM they may have, my suggestion would be to assume the other LPM is accurate and maybe send the Ophir and/or a laser to another member you trust to compare against readings they get with their LPM.

The laser that started all of this was the one that I recently sold in this thread: http://laserpointerforums.com/f39/sold-2-watt-445nm-ultrafire-u-501b-beast-laser-69893.html. It was built with one of RHD's Mosquito driveres set at 1.74A and using a G2 lense. And just for laughs, here are the readings I got from the Ophir:

 

[lasersbee]

I believe that between everything that we've done here, we've proven that I just got a crappy Ophir head and this issue is not widespread. But, if there is anyone else out there that does have an Ophir heard that gives substantially higher readings than anticipated or than any other LPM they may have, my suggestion would be to assume the other LPM is accurate and maybe send the Ophir and/or a laser to another member you trust to compare against readings they get with their LPM.

I'm wondering if that is the case or not since Trevor tested 4 other
OPHIR Heads on his new CUBE Laser and they all measured low except
for the one OPHIR Head I had tested at 9% high and he tested the same
head at 80mW ...

Now it seems that Trevor's new CUBE Laser set to 80mW may not be
putting out exactly 80mW....:thinking:

I'd still like to know where Trevor got his 4 OPHIR Heads...


Jerry

 

[mark-in-dallas]

That is a good question. Someone clearly tampered with my Ophir head, and I do know 2 others who have Ophir's that have read similarly high to mine. If the inconsistancies with these heads weren't so close I would not think this, but the fact that each has observed readings that are 15% higher than their other LPM's suggests that they were tampered with by the same person.

It would be interesting to trace who's hands each of our Ophirs passed though and see if there was a common individual along the ownership path.

At least for myself though, I've satisfied my suspicion that my Ophir is off and somebody messed with it.

 

[Trevor]

I'm wondering if that is the case or not since Trevor tested 4 other
OPHIR Heads on his new CUBE Laser and they all measured low except
for the one OPHIR Head I had tested at 9% high and he tested the same
head at 80mW ...

Now it seems that Trevor's new CUBE Laser set to 80mW may not be
putting out exactly 80mW....:thinking:

I'd still like to know where Trevor got his 4 OPHIR Heads...


Jerry

My CUBE laser was brand new, directly from Coherent. I'm pretty sure it's the most recently calibrated instrument we have at our disposal.

It seems like the issue is more pronounced at higher powers, and your two data points support that. For now, I'm confident my CUBE readouts are accurate. I'm getting hold of a 2.5W LaserBee so that I can test it against my CUBE both to quantify its response time and compare it against my results with the Ophir sensors.

Once that comes in I'll report my results in this thread.

All four of my Ophir sensors were procured this summer from the ebay seller that was selling them.

Trevor

 

[lasersbee]

My CUBE laser was brand new, directly from Coherent. I'm pretty sure it's the most recently calibrated instrument we have at our disposal.

It seems like the issue is more pronounced at higher powers, and your two data points support that. For now, I'm confident my CUBE readouts are accurate. I'm getting hold of a 2.5W LaserBee so that I can test it against my CUBE both to quantify its response time and compare it against my results with the Ophir sensors.

Once that comes in I'll report my results in this thread.

All four of my Ophir sensors were procured this summer from the ebay seller that was selling them.

Trevor

I would also find it strange that a Brand New instrument from Coherent
would not be calibrated as you say....
I also bought a few heads from the eBay seller and the OHIR head I
used for the tests was a pristine one from that purchase.

I would assume that these surplus eBay heads are NEW and correctly
Calibrated by OPHIR.. That is if no-one plays with the 3 pots on the
amplifier PCB in the head.

I'll test all the new pristine ones I have here and see what they read at
500mW and 1000mW to keep in line with my original test criteria...
It may take a few days. I'm on 2 other projects at the moment...


Jerry