Calibration needed for LPM

[zartoop]

I have a TEC based LPM detector that I would like calibrated.

I am in Adelaide Australia. Anyone local with a calibrated LPM that could help me out please?

Any help much appreciated.

Heatsink - CPU heatsink with fan - surplus from local computer shop - $0
TEC - Custom Thermoelectric TEC - pn 03111-5L31-03CF - $20.00
Krylon UltraFlat Black Paint


Polished the mounting area for the TEC to a mirror finish
Painted one face of the TEC with black
Glued the TEC to the heatsink with Artic Silver
Painted the inside of the aluminium tube with black
Glued the aluminium tube to the heatsink with high grade cyanoacrylate
The aluminium tube has about 1mm clearance all round the TEC
Glued the lead wires to the heat sink for strain relief
Measure mV output direct from the TEC with multimeter (no amp required)

The aluminium tube prevents problems with stray air currents, direct light and also provides a great way to mount various attachments for measuring mounted and unmounted diodes.

Haven't needed to use the fan at all but only measuring up to about 350mW so far. The fan will help when measuring higher powers for longer duty cycles as the airflow will help maintain a more stable heatsink temp.

All works like a charm with high repeatability and stability but now I need calibration to correlate mV to mW.

 

[lasersbee]

I have a TEC based LPM detector that I would like calibrated.

I am in Adelaide Australia. Anyone local with a calibrated LPM that could help me out please?

Any help much appreciated.

Heatsink - CPU heatsink with fan - surplus from local computer shop - $0
TEC - Custom Thermoelectric TEC - pn 03111-5L31-03CF - $20.00
Krylon UltraFlat Black Paint


Polished the mounting area for the TEC to a mirror finish
Painted one face of the TEC with black
Glued the TEC to the heatsink with Artic Silver
Painted the inside of the aluminium tube with black
Glued the aluminium tube to the heatsink with high grade cyanoacrylate
The aluminium tube has about 1mm clearance all round the TEC
Glued the lead wires to the heat sink for strain relief
Measure mV output direct from the TEC with multimeter (no amp required)

The aluminium tube prevents problems with stray air currents, direct light and also provides a great way to mount various attachments for measuring mounted and unmounted diodes.

Haven't needed to use the fan at all but only measuring up to about 350mW so far. The fan will help when measuring higher powers for longer duty cycles as the airflow will help maintain a more stable heatsink temp.

All works like a charm with high repeatability and stability but now I need calibration to correlate mV to mW.

What you have there is only a Sensor. It is not an LPM.

You will need some kind of electronics to be able to calibrate
your TEC mounted Sensor to anything near mV to mW accuracy.

You will also need to know your coating wavelength errors over
the intended power range of the coating.
Not to mention any linearity deviations of the TEC.


Jerry

You can contact us at any time on our Website: J.BAUER Electronics

 

[zartoop]

Thanks ...

What you have there is only a Sensor. It is not an LPM.

Yes I know thanks – that’s why I wrote “I have a TEC based LPM detector “
What I have pictured is the sensor. By measuring the output (and when calibrated) it will be a LPM.

You will need some kind of electronics to be able to calibrate your TEC mounted Sensor to anything near mV to mW accuracy.

I don’t see why. The output in mV can be directly measured by multimeter to 0.1mV accuracy so no amplification required. (a ~5mW pointer outputs 5mV and I have tested diodes upto 200mW with mV in the 20+ range. So the TEC output is large) The offset required can be accomplished by reading the initial output of the sensor and simply subtracting that amount from any readings (assuming that the relative temp between the heatsink and ambient remains constant, which for all practical purposes it will. This is of course highly simplified I realise).

So mW = f(mV – n)
Where f = multiplication/calibration factor, n = offset voltage

You will also need to know your coating wavelength errors over the intended power range of the coating.

Already know that. JPL tested Krylon and it has ~97% absorbtion from UV to NIR. Calibration against a known source will supply data related to wavelength and power.

Not to mention any linearity deviations of the TEC.

The TEC is essentially linear over the small temp range being used.

If I have missed something I’d be happy to know.

So …. Would someone be kind enough to help me calibrate against a known setup please?

 

[lasersbee]

Thanks ...


Yes I know thanks – that’s why I wrote “I have a TEC based LPM detector “
What I have pictured is the sensor. By measuring the output (and when calibrated) it will be a LPM.


I don’t see why. The output in mV can be directly measured by multimeter to 0.1mV accuracy so no amplification required. (a ~5mW pointer outputs 5mV and I have tested diodes upto 200mW with mV in the 20+ range. So the TEC output is large) The offset required can be accomplished by reading the initial output of the sensor and simply subtracting that amount from any readings (assuming that the relative temp between the heatsink and ambient remains constant, which for all practical purposes it will. This is of course highly simplified I realise).

So mW = f(mV – n)
Where f = multiplication/calibration factor, n = offset voltage


Already know that. JPL tested Krylon and it has ~97% absorbtion from UV to NIR. Calibration against a known source will supply data related to wavelength and power.


The TEC is essentially linear over the small temp range being used.

If I have missed something I’d be happy to know.

So …. Would someone be kind enough to help me calibrate against a known setup please?

It seems you have it all figured out...
You also seem to be on the right track according to what you
believe.
Good luck calibrating your Krylon coated raw TEC Sensor... :beer:

BTW... I wonder why no one else has ever figured out your
simple theory/logic to make an accurate LPM using only a TEC
and DMM...:thinking:


Jerry

You can contact us at any time on our Website: J.BAUER Electronics

 

[bobhaha]

HA...

Good luck man, you're gonna need it.

 

[zartoop]

It seems you have it all figured out...
You also seem to be on the right track according to what you
believe.
Good luck calibrating your Krylon coated raw TEC Sensor...

Semantics Jerry ... of course I am not calibrating the sensor ... I am calibrating the system. And don't knock TEC sensors even if it is tongue in cheek. The Laserbee uses a coated TEC sensor. Are you saying they're no good? Krylon paint is a low cost, very effective coating - plenty of research literature on that.

BTW... I wonder why no one else has ever figured out your simple theory/logic to make an accurate LPM using only a TEC and DMM...

Not too sure on your sarcasm ... Plenty of people claim to make an accurate LPM using a TEC and DMM - your Laserbee to name one.

Laserbee LPM is a TEC based sensor connected to your meter.

The Laserbee meter basics consists of instrumentation opamp, MCU processing for zero offset, temp compensation, display, datalogging and data IO. Plus software of course for PC display etc. Your advantage is that you can calibrate Laserbees against a known standard. All credit to you for commercialising it and having happy customers.

This is a hobby and I enjoy tinkering so I decided to make my own. Don't knock it. You asked some questions/ made some comments and I answered them - like I said is there something specific I missed?

Mine is a TEC based sensor connected to my datalogging DMM. It works well in regards to stability, repeatabilty, temp compensation etc - so do you have any leads in Oz for me calibrate ...

 

[zartoop]

HA...

Good luck man, you're gonna need it.

??????

 

[bobhaha]

The problem is that in practice your sensor will not be linear over the wavelengths. If you want any sort of accuracy you need to test the coating over specific wavelength ranges and find the correction factor. Then you need to test it over different power ranges as well. Needless to say this is tedious.

You're basically asking people to do the hard work for you...

I'd recommend buying the main 3 wavelengths and testing it your self. A PHR, LCC and a 532. Put them in a large heatsink... let them warm up and settle out, then meter it with your sensor. Record the results and go from there. It'll only cost you about $30 max.

PS: having a fan on the heatsink is just stupid....

 

[zartoop]

The problem is that in practice your sensor will not be linear over the wavelengths. If you want any sort of accuracy you need to test the coating over specific wavelength ranges and find the correction factor. Then you need to test it over different power ranges as well. Needless to say this is tedious.

You're basically asking people to do the hard work for you...

I'd recommend buying the main 3 wavelengths and testing it your self. A PHR, LCC and a 532. Put them in a large heatsink... let them warm up and settle out, then meter it with your sensor. Record the results and go from there. It'll only cost you about $30 max.

PS: having a fan on the heatsink is just stupid....

Sorry if I was not clear - I am not asking anyone else to do any work – I’ve done it all and will do the rest.
As we are all learning and doing this hobby together I thought a member might provide access to either
1) A calibrated LPM to allow me calibrate mine against, or
2) Known power output lasers so i can calibrate my setup, or
3) Some [politely put] useful tips and pointers.
“Fan on heatsink stupid” – yep probably (although you probably could have been a bit more diplomatic) and haven’t used it yet – because for any reasonable power level and ambient conditions the temp rise of the heatsink will be least of the variables to correct for.
I already have bought all the wavelengths and have tested the sensor/coating and recorded it’s characteristics in relation to wavelength and relative power levels.
So all that is needed now is access to a known standard to allow conversion and corrections from mV to mW

 

[lasersbee]

Semantics Jerry ... of course I am not calibrating the sensor ... I am calibrating the system. And don't knock TEC sensors even if it is tongue in cheek. The Laserbee uses a coated TEC sensor. Are you saying they're no good? Krylon paint is a low cost, very effective coating - plenty of research literature on that.



Not too sure on your sarcasm ... Plenty of people claim to make an accurate LPM using a TEC and DMM - your Laserbee to name one.

Laserbee LPM is a TEC based sensor connected to your meter.

The Laserbee meter basics consists of instrumentation opamp, MCU processing for zero offset, temp compensation, display, datalogging and data IO. Plus software of course for PC display etc. Your advantage is that you can calibrate Laserbees against a known standard. All credit to you for commercialising it and having happy customers.

This is a hobby and I enjoy tinkering so I decided to make my own. Don't knock it. You asked some questions/ made some comments and I answered them - like I said is there something specific I missed?

Mine is a TEC based sensor connected to my datalogging DMM. It works well in regards to stability, repeatabilty, temp compensation etc - so do you have any leads in Oz for me calibrate ...

Sorry if I was not clear - I am not asking anyone else to do any work – I’ve done it all and will do the rest.
As we are all learning and doing this hobby together I thought a member might provide access to either
1) A calibrated LPM to allow me calibrate mine against, or
2) Known power output lasers so i can calibrate my setup, or
3) Some [politely put] useful tips and pointers.
“Fan on heatsink stupid” – yep probably (although you probably could have been a bit more diplomatic) and haven’t used it yet – because for any reasonable power level and ambient conditions the temp rise of the heatsink will be least of the variables to correct for.
I already have bought all the wavelengths and have tested the sensor/coating and recorded it’s characteristics in relation to wavelength and relative power levels.
So all that is needed now is access to a known standard to allow conversion and corrections from mV to mW

All you have shown is a Krylon coated TEC glued to a heatsink..
with a fan I may add...
That does NOT make a calibrated broadband LPM...

The only electronics you have mentioned is a DMM...:thinking:

It's like asking someone to balance a tire for you.. but it has
no rim just the raw rubber tire...

Since you are intent to compare our LaserBee products to your
raw TEC glued to a Heatsink... I'll enlighten you a bit.

Our LaserBee LPMs are coating corrected in the Firmware to
compensate for our proprietary coating's non-linearity (yes
we have more than one formula) running on the on-board
Microcontroller.

In case you think it is only our coating that is not linear... let
be assure you that in the past 5 years we have tested all types
of "black paint" we could get our hands on and our proprietary
coatings are the most linear we have found to date.

Our LaserBee LPMs are calibrated using that MCU Firmware and
are calibrated against a recently calibrated Newport 1825c LPM.

As I stated in my 1st post you will need some type of electronic
control to calibrate your "LPM"...

You can not use a DMM directly unless you include a manual
or electronic calculator to be used on every reading to calculate
your "actual" readings as you seem to want to do.
It is doable like that but seems you still need a calibrated LPM
to work out your DMM reading to Watts ratios.

That is what you are asking someone to do for you and as
bobhaha has stated will require a lot of repetitive work to
accomplish.

Yeah.... I get a bit sarcastic when I try to steer new members in
the right direction but still think they know it all better and want
to re-invent the wheel with old and tried ideas...


Jerry

You can contact us at any time on our Website: J.BAUER Electronics

 

[Bluefan]

Nice work Zartoop! A simple opamp amplifier circuit can make life easy by amplifying the signal to let millivolts correspond to milliwats, but if you're satisfied by just applying a conversion factor you already have a working power meter.

Assuming TEC's and coatings are fully linear is actually reasonably good. To be sure you'd have to do some measurements but you seem to have a known coating and a linearity test is VERY simple, so you're nearly there already.

I'm on the other side of the world, but should you be interested I can offer a calibration service (in a few weeks, I'm now at CERN).

And Jerry, lots of people have figured out that with a tec and a dmm you can make something that reads power. zartoop even did a good analysis and tests of his system, all he needs is comparison to an absolute value. That's where nearly all people get stranded. You make the building of power meters look more complicated than it is, only the calibration of power linearity and wavelength linearity involves more work but those are actually not that important. As example: The Ophir OEM heads use no power nonlinearity correction at all.

 

[zartoop]

Thanks Bluefan:thanks:! I appreciate your advice and great member/customer care.

I had actually missed the point of the opamp being used to apply hardware conversion to correspond mV to mW. I will knock up a instrumentation preamp to interface the sensor with my DMM. I had intended applying the conversion in Excel calcs using the data from the DMM (or my head for one offs!). In reality I'll end up doing both for a while as even a precision preamp adds yet another variable to be compensated for. I think we'll characterise the sensor raw output only to see how much variation there is across wavelengths. The power linearity within a wavelength seems to match with expected results. Wavelength errors are another matter, theoretically from 400 to 550 nM the coating reflectance varies from 3.8 to 3.6%, there are some quite large variations apparently expected from 600 to 800 nM but still 'only' 3.1 to 4% reflectance. There is a huge spike at about 580 nM but that might be some other test artifact. I can live with a 1-5% total error = it's a hobby after all.

Your website looks great and more importantly is very professional with more complete product specs than many others = that generates confidence in your products and services. Well done!

I'd be really happy to use your service if nothing else surfaces between now and when you get back from CERN. Would you please get back to me with a quote for 4 wavelengths x 4(min) power data points in each from 5mW to 1 watt? If you think there is a better data point arrangement I'd be happy to know.

It will be interesting for me to see how the sensor and coating perform relative to datasheets, published data and design calcs.

 

[bobhaha]

Sorry if I was not clear - I am not asking anyone else to do any work – I’ve done it all and will do the rest.
As we are all learning and doing this hobby together I thought a member might provide access to either
1) A calibrated LPM to allow me calibrate mine against, or
2) Known power output lasers so i can calibrate my setup, or
3) Some [politely put] useful tips and pointers.
“Fan on heatsink stupid” – yep probably (although you probably could have been a bit more diplomatic) and haven’t used it yet – because for any reasonable power level and ambient conditions the temp rise of the heatsink will be least of the variables to correct for.
I already have bought all the wavelengths and have tested the sensor/coating and recorded it’s characteristics in relation to wavelength and relative power levels.
So all that is needed now is access to a known standard to allow conversion and corrections from mV to mW

Yeah I could have been more diplomatic, but you have no idea how many times these types of questions/threads are made.

You seem really intent on getting this done... So I'll help you out. I have an old Laserbee I 1W unit I'm willing to send you. I live in Sydney so it shouldn't take too long to get to you.

If you're interested send me a PM and we'll work this out.
-Adrian

 

[zartoop]

Thanks Adrian:wave: PM sent.

 

[gillza]

PS: having a fan on the heatsink is just stupid....

Could you explain why? Im genuinely curious. Wouldn't it help dessipate heat from the heatsink and assist in "pumping" the heat from the cold side?

 

[AnthoT]

Could you explain why? Im genuinely curious. Wouldn't it help dessipate heat from the heatsink and assist in "pumping" the heat from the cold side?

The amount of heat generated on the TEC is nothing any heatsink can't handle... If anything the fan will possibly keep it too cold & throw it off. Most LPM's are calibrated at aproximately 22-24 degrees celcius & the fan will cool it too much...

That's a possibility not 100% accurate