Hello,

1. Thermal Sensors

In most cases, a thermal head is the logical choice. It has a relatively high damage threshold and a broadband spectral response. Depending on the gas used for the lamp, it will output wavelength in a broad spectrum. That means it needs to be measured with a device that is wavelength independent.

Thermal sensors are very close to that ideal.

But what about low power lamps? Thermal heads can measure powers in the tens of mW, for the special 3A sensor down to tens of μW. But what about lower power?

2. PD300-BB

In that case, you will have to use a photodiode detector.

The problem is that photodiodes’ response varies far more greatly with wavelength. This is fine for measuring single-wavelength lasers, but will make measuring broadband lamps quite difficult, if not impossible.

For that reason, Ophir developed a special filter that renders its photodiode PD300 sensor nearly flat. We call it the PD300-BB (BB for “broadband”).

In the unique case where your lamp is both low-power and UV, neither of these solutions will work.

(The PD300-BB sensor can only measure from 430 nm and higher.)

Fear not! Not all hope is lost.

If you know the wavelength spectrum emitted by your gas lamp, you can calculate the power.

Good luck!

I don't know a lot about SLEDs, but it seems likely to me that you are seeing what's known as spectral ripple. If it's stable enough, then it should divide out in your measurements, but since you observe it being sensitive to the fiber being touched you should ensure the fiber is not moved or touched throughout the measurements. Also, since it seems to be related to reflections between the fiber and the SLED, you probably would want to minimize direct reflections from the optical components you are measuring into the fibers. Apparently SLEDs are available with low ripple, or you could use an incandescent source like a LabSphere.

https://en.wikipedia.org/wiki/Superluminescent_diode#Spectral_ripple

https://en.wikipedia.org/wiki/File:Spectral_ripple_of_SLED.svg

Karen A. Darbinyan

Dear Karen,

thanks a lot for your suggestions. Also I didn't quite understand how to get the spectrum profile.

If I use a power sensor, then I would need a wavelength filter like a monochromator to measure the power for each wavelength separately. I might eventually do that but is there anyway to solve the instability of the spectrometer?

PS. The source is 50 mW from 800 to 900 nm.

thanks,

You can check the parameters of your device sensors,it's capacity to measure spectrum .

One more advice is use the spectrum of light coming from North direction in evening which contains full spectrum of natural light and advised for taking tooth shade for fabricating the artificial prosthesis

Dennis P Boyle

Dear Dennis,

Thanks a lot for your reply. I'm not sure if they are spectral ripples.

The technical report of the light source states that the ripples are less than 5% which for me would have been ok, but the oscillations I'm seeing after the SMA prob are much bigger nearly 50% or so.

For the reflections I have FC/APC connector coming out of the SLED and the Isolator, so any reflection from optical element should not enter the source right?

thanks,

There are many phenomena that could be contributing to this behaviour. Here are some:

Spaces in fiber connectors acting as Fabry-Perot etalons.

Whispering-gallery modes in the fibers caused by cladding or jacket modes interfering with the light in the core.

Other components behaving as etalons.

Complex effects associated with the isolator.

I'd suggest that you go back to really simple setup and add parts step by step.

Direct a conventional LED direct into the spectrometer (no fibers) - do you get what you expect?

Direct the LED in via the spectrometer fiber...

Direct the SLED in without a fiber (move it away if the signal is too big)...

Direct the SLED in via the spectrometer fiber...

etc.

Keep adding complexity until you get an unexpected result - then try to work out what has gone wrong. Using a white box as an integrating "sphere" can sometimes be helpful in getting a low but consistent launch into fibers.

Fibers vary a lot in their throughput when you move them, bend them, squeeze them etc. Also when the light launched into them varies in angular distribution. To get accurate absorption measurements these phenomena must be controlled. Optical elements that change the direction of rays (anything except a window) will affect the throughput in pretty random ways and what you measure may not be what you want! An integrating sphere might help - put the optical element at the input port and connect the spectrometer fiber to an output port.

Good luck!