On Sun, 19 Dec 2010 17:13:42 -0800, John Larkin
<jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

>On Sat, 18 Dec 2010 20:13:07 -0800, TheGlimmerMan
><justaglimmer@thebarattheendoftheuniverse.org> wrote:
>
>>On Sat, 18 Dec 2010 19:21:13 -0800 (PST), Nicholas Kinar
>><n.kinar@usask.ca> wrote:
>>
>>>Hello,
>>>
>>>I am using an IR thermometer (MLX90614 from Melexis) to measure the
>>>temperature of a surface. This device is in a TO-can, and I am using
>>>the standard package version (cf. first page of the datasheet, option
>>>code A).  I would like to place a lens in front of this IR thermometer
>>>so that I can measure the temperature of the surface with a spot size
>>>of 3 mm.
>>>
>>>How do I select or calculate the required
>>>
>>>(a) focal length of the lens;
>>>(b) lens diameter;
>>>(c) distance of the lens to the detector
>>>
>>>I would like the lens to be situated at a distance of 5 cm from the
>>>measurement surface.  I assume that I would require the lens to have
>>>an IR response and appropriate coating, since according to the
>>>thermometer datasheet, the spectral response of the IR thermometer is
>>>limited from 5.5 um to 15 um by a filter.
>>>
>>>Would similar principles also hold for the selection of a lens in
>>>front of a photodiode?  Can anyone suggest how I might go about
>>>selecting the lens?  Is a good reference work available?
>>
>>  The $20 Fresnel lensed IR thermometer at Harbor Frieght can be micro
>>adjusted (aimed) pretty easily, and at 5cm, you could easily point it
>>only at the device in question.
>>
>> Place a piece of room temp (ambient) matte black painted aluminum in
>>front of the target with like a 1cm hole in it at the target location.
>>Make a set of successively smaller holed plates like that one.
>>
>>  With that, you can see just how well your device is already pointing.
>>
>> With successively smaller holes in the ambient plate, that is.
>>
>>  Also, your device under test has to be hotter than ambient to be a
>>proper pointing determination target, etc.
>
>Why not just sweep it across a point source of heat and see the
>response curve? 
>
>John


  IR black body calibration sources ("calibration ovens")are typically
made such that the heated black body is an Aluminum ingot placed behind a
matte painted sheet of Aluminum with a hole in it slightly smaller than
the ingot.  This keep fringe errors low near the edges of the ingot, and
cuts down on ambient convection currents wisping past your target when
you are trying to calibrate an IR device.

 The ingot is called "the target" and the "target size" equals the size
of the hole or "port" in the Aluminum sheet.

  He already stated what he wanted his targeting capability to be. 3mm.

  My test solves two problems at once.  Targeting capacity, and amplitude
when on target at a given temperature.

On Sat, 18 Dec 2010 20:13:07 -0800, TheGlimmerMan
<justaglimmer@thebarattheendoftheuniverse.org> wrote:

>On Sat, 18 Dec 2010 19:21:13 -0800 (PST), Nicholas Kinar
><n.kinar@usask.ca> wrote:
>
>>Hello,
>>
>>I am using an IR thermometer (MLX90614 from Melexis) to measure the
>>temperature of a surface. This device is in a TO-can, and I am using
>>the standard package version (cf. first page of the datasheet, option
>>code A).  I would like to place a lens in front of this IR thermometer
>>so that I can measure the temperature of the surface with a spot size
>>of 3 mm.
>>
>>How do I select or calculate the required
>>
>>(a) focal length of the lens;
>>(b) lens diameter;
>>(c) distance of the lens to the detector
>>
>>I would like the lens to be situated at a distance of 5 cm from the
>>measurement surface.  I assume that I would require the lens to have
>>an IR response and appropriate coating, since according to the
>>thermometer datasheet, the spectral response of the IR thermometer is
>>limited from 5.5 um to 15 um by a filter.
>>
>>Would similar principles also hold for the selection of a lens in
>>front of a photodiode?  Can anyone suggest how I might go about
>>selecting the lens?  Is a good reference work available?
>
>  The $20 Fresnel lensed IR thermometer at Harbor Frieght can be micro
>adjusted (aimed) pretty easily, and at 5cm, you could easily point it
>only at the device in question.
>
> Place a piece of room temp (ambient) matte black painted aluminum in
>front of the target with like a 1cm hole in it at the target location.
>Make a set of successively smaller holed plates like that one.
>
>  With that, you can see just how well your device is already pointing.
>
> With successively smaller holes in the ambient plate, that is.
>
>  Also, your device under test has to be hotter than ambient to be a
>proper pointing determination target, etc.

Why not just sweep it across a point source of heat and see the
response curve? 

John

> I suspect that IR sensors have fudge factors that account for the mix
> of object radiation and internal (ambient temp) radiation that the
> sensor sees. So if you change the lens, that will change too. I think
> it might amount to introducing an error into the emissivity
> correction.
>
> John

Thanks John; I will have to assess this effect carefully during
calibration of the IR thermometer.

On Sun, 19 Dec 2010 11:39:40 -0500, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

>Nicholas Kinar wrote:
>>
>>> The colour won't make any difference--all paint has high emissivity in
>>> the thermal IR.  Even shiny metal won't do much--you aren't cooling your
>>> detector, so it's going to see a room-temperature ambient field regardless.
>>>
>>> The key is to use a lens that's transparent in the thermal IR, which
>>> doesn't happen by accident!
>>>
>>> Cheers
>>>
>>> Phil Hobbs
>>
>> Thanks, Phil; and thanks also to all who responded.  I've been looking
>> at IR lenses from Newport, and I think that these might be the best
>> way to go (http://www.newport.com/Infrared-Lenses/381068/1033/
>> catalog.aspx).  For budget applications, perhaps the AMPX010 or
>> AMPX013 would be most appropriate, but the other lenses sold by this
>> company may be good as well.
>>
>
>Sensitivity will be better with the 25 mm diameter, 25 mm FL one.
>
>Infrared sensitivity is all about how much solid angle the FOV subtends 
>at the detector.  It's just like a voltage divider: you get (image 
>projected solid angle)/pi
>less sensitivity you'd get with uniform (Lambertian) illumination.
>
>(Projected solid angle is cos theta integrated over the FOV--for a lens 
>of a given NA, it's conveniently
>
>Omega prime = pi NA**2.
>
>For an f/1 lens (NA=0.5) you lose a factor of 4, whereas for f/1.5 (NA = 
>0.33) it's a factor of 9.
>
>Cheers
>
>Phil Hobbs

I suspect that IR sensors have fudge factors that account for the mix
of object radiation and internal (ambient temp) radiation that the
sensor sees. So if you change the lens, that will change too. I think
it might amount to introducing an error into the emissivity
correction.

John

> Sensitivity will be better with the 25 mm diameter, 25 mm FL one.
>
> Infrared sensitivity is all about how much solid angle the FOV subtends
> at the detector. =A0It's just like a voltage divider: you get (image
> projected solid angle)/pi
> less sensitivity you'd get with uniform (Lambertian) illumination.
>
> (Projected solid angle is cos theta integrated over the FOV--for a lens
> of a given NA, it's conveniently
>
> Omega prime =3D pi NA**2.
>
> For an f/1 lens (NA=3D0.5) you lose a factor of 4, whereas for f/1.5 (NA =
=3D
> 0.33) it's a factor of 9.
>
> Cheers
>
> Phil Hobbs

Phil, thank you for posting this.  I think that I follow what you are
saying, and I was able to calculate the loss factors for a number of
different lenses.  (I like the voltage divider analogy.)  If I follow
this correctly, would a high numerical aperture lens be most
appropriate for this application?  Thus, the higher the numerical
aperture of the lens, the more sensitive that it would be, and the
loss factor is consequently lower?  So by increasing the diameter of
the lens and keeping the focal length constant, the sensitivity
increases?  Does the magnification of the lens (expressed as an
effective f-number) play a role in any of this as well?

Nicholas Kinar wrote:
>
>> The colour won't make any difference--all paint has high emissivity in
>> the thermal IR.  Even shiny metal won't do much--you aren't cooling your
>> detector, so it's going to see a room-temperature ambient field regardless.
>>
>> The key is to use a lens that's transparent in the thermal IR, which
>> doesn't happen by accident!
>>
>> Cheers
>>
>> Phil Hobbs
>
> Thanks, Phil; and thanks also to all who responded.  I've been looking
> at IR lenses from Newport, and I think that these might be the best
> way to go (http://www.newport.com/Infrared-Lenses/381068/1033/
> catalog.aspx).  For budget applications, perhaps the AMPX010 or
> AMPX013 would be most appropriate, but the other lenses sold by this
> company may be good as well.
>

Sensitivity will be better with the 25 mm diameter, 25 mm FL one.

Infrared sensitivity is all about how much solid angle the FOV subtends 
at the detector.  It's just like a voltage divider: you get (image 
projected solid angle)/pi
less sensitivity you'd get with uniform (Lambertian) illumination.

(Projected solid angle is cos theta integrated over the FOV--for a lens 
of a given NA, it's conveniently

Omega prime = pi NA**2.

For an f/1 lens (NA=0.5) you lose a factor of 4, whereas for f/1.5 (NA = 
0.33) it's a factor of 9.

Cheers

Phil Hobbs
-- 
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058

email: hobbs (atsign) electrooptical (period) net
http://electrooptical.net

> The colour won't make any difference--all paint has high emissivity in
> the thermal IR. =A0Even shiny metal won't do much--you aren't cooling you=
r
> detector, so it's going to see a room-temperature ambient field regardles=
s.
>
> The key is to use a lens that's transparent in the thermal IR, which
> doesn't happen by accident!
>
> Cheers
>
> Phil Hobbs

Thanks, Phil; and thanks also to all who responded.  I've been looking
at IR lenses from Newport, and I think that these might be the best
way to go (http://www.newport.com/Infrared-Lenses/381068/1033/
catalog.aspx).  For budget applications, perhaps the AMPX010 or
AMPX013 would be most appropriate, but the other lenses sold by this
company may be good as well.

> =A0 We had solid Ge lens, but also had bolometers which had various
> "windows" right on the transducer, which is, of course, the cheapest
> route. =A0Best of class. =A0He stated his already has a filter on it. =A0=
It is
> probably at the transducer level, and if the range he gave is wide,
> further filtration would also mean further attenuation of the signal, so
> hopefully he can compensate for that.

Thanks for your response; yes - you are right about the filter being
at the transducer level.  Since it has been added by Melexis on the
MLX90614, and I did not select it myself, adding another lens will
indeed attenuate the light.

On Sun, 19 Dec 2010 02:59:39 -0800 (PST), Greegor <greegor47@gmail.com>
wrote:

>
>Archie, I'm proud of you!
>Not a nasty aspersion or swearword in that whole post!

  Fuck off, stalk boy!
>
>Would you like this to be another addenda to your
>recent assessment or a good start on your next one?

  Keep posting, bitch.  Maybe I will print them out and use them as
templates for the tattoos I emblazon across your pathetic forehead.

Nicholas Kinar wrote:
> Thanks Phil and TheGlimmerMan; thank you very much for your
> suggestions and for putting me on the right track.  I suppose this is
> simpler than I thought - no need to worry about Gaussian beam theory,
> field of view, or the numerical aperture of the lens.  These were the
> questions in my head when I made the initial post.
>
> One other question: If I am to use a C-mount lens holder for the lens,
> would I have to worry about the color that the lens holder is painted
> if the sensor and lens assembly are taken outdoors?  I've noticed that
> most of these C-mount lens holders are painted black, but this might
> heat up if the thermometer and lens assembly is placed outside in the
> sun.  Would it be beneficial to coat the outside of the lens holder
> with white paint?  This would be the side of the lens holder facing
> the surface to be measured.  (Paint containing barium sulfate or a
> similar chemical might work.)  The lens would be inserted into the C-
> mount holder and the holder then screwed on the end of the tube.  The
> IR sensor would be situated at the end of the tube, aligned with the
> optical axis of the lens.  What color should be the inside of the
> tube?

The colour won't make any difference--all paint has high emissivity in 
the thermal IR.  Even shiny metal won't do much--you aren't cooling your 
detector, so it's going to see a room-temperature ambient field regardless.

The key is to use a lens that's transparent in the thermal IR, which 
doesn't happen by accident!

Cheers

Phil Hobbs

-- 
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058

email: hobbs (atsign) electrooptical (period) net
http://electrooptical.net