Killer deal on USB spectrometer! (update: legitimate!)

[ManCave]

Those are the same procedures. It is a bit easier with a grating, but the exact same method is used. Try binning a couple dozen laser diodes using that method. Also, you get much better results if you use longer distances. Ten cm is too short.

I agree,
Still, these kind of measuring methods are a good introduction to the concept for the uninitiated. And they teach math which is always a good thing
Cheers!

 

[FlatBroke]

Yeesh! I didn't actually look for it. Has a 1 mm slit. Professional spectrometers have 25-10 um slits. They claim an accuracy of 5nm, which is horrible. I wouldn't expect it to be that accurate though. Wonder if it uses the same two point calibration that other cheap units have.

I have one. They quote 5nm accuracy but mine is neaerr to 1nm. The pro model comes with an interchangeabl;e slit plate allowing smaller slits to be used. Yes, it uses Theremino software which uses 2-point calibration. That seems to give very accurate measurements for my laser pointers, form deeep blue to red, and my IR laser and my TV remove around (I thnink I recall) at 890nm. I can read the sodium spectrum from distant street lights nd I've found invisible security cameras from their IR flood lights (just for fun). It can do absorption too, though I have not tried it. It is sensitive from 400 nm (quoted - mines goes down to 360 nm) up to 1000 nm. Now - it doesn't do absolute intensity, and the webcam driver means you have to know how to get the best from it. It is a compromise between cost and functionality. It won't do everything anyone might ask of a spectrometer but then it's less than 1% of the cost of a typical Ocvean Optics spec and does many of the same things just as well. It suits many people like me, who want to do things like testing power LEDs and other growing lights. See the videos and the reviews. It's not a high-end instrument - DUHHH - but neither it is a toy, though at this price you could certainly buy one for your kids to play with.

 

[paul1598419]

I agree,
Still, these kind of measuring methods are a good introduction to the concept for the uninitiated. And they teach math which is always a good thing
Cheers!

Yeah, I have used this method in the past too. I have a large sheet diffraction grating that is 1000 l/mm. It works well for this purpose and I can cut smaller sheets out of it. I have sent some out to other members here.

I have one. They quote 5nm accuracy but mine is neaerr to 1nm. The pro model comes with an interchangeabl;e slit plate allowing smaller slits to be used. Yes, it uses Theremino software which uses 2-point calibration. That seems to give very accurate measurements for my laser pointers, form deeep blue to red, and my IR laser and my TV remove around (I thnink I recall) at 890nm. I can read the sodium spectrum from distant street lights nd I've found invisible security cameras from their IR flood lights (just for fun). It can do absorption too, though I have not tried it. It is sensitive from 400 nm (quoted - mines goes down to 360 nm) up to 1000 nm. Now - it doesn't do absolute intensity, and the webcam driver means you have to know how to get the best from it. It is a compromise between cost and functionality. It won't do everything anyone might ask of a spectrometer but then it's less than 1% of the cost of a typical Ocvean Optics spec and does many of the same things just as well. It suits many people like me, who want to do things like testing power LEDs and other growing lights. See the videos and the reviews. It's not a high-end instrument - DUHHH - but neither it is a toy, though at this price you could certainly buy one for your kids to play with.

A spectrometer is not a linear device. You cannot calibrate one using only two points. It's not possible. You can believe what you like, but I know that thing is not accurate to 5nm. It couldn't be as described at their site and by you. I typically calibrate mine using at least 7 points. That is the minimum and the more you can add the greater the accuracy.

 

[Encap]

A spectrometer is not a linear device. You cannot calibrate one using only two points. It's not possible. You can believe what you like, but I know that thing is not accurate to 5nm. It couldn't be as described at their site and by you. I typically calibrate mine using at least 7 points. That is the minimum and the more you can add the greater the accuracy.

The above is 100% spot on ^^^

A toy is a toy--loosely low level educational or not, an entertainment device--like a mobile over a baby's crib is---yeah they both do something that's entertaining but...

What can anyone expect for $60 retail price that cost maybe $20 to make? It works along side a chemistry set for kids yes as something to learn from---not far beyond that.

In the land of the blind, the one eyed man is king, so it seems .

 

[hakzaw1]

great stuff from Paul and Encap
will +rep when sys allows.....

 

[FlatBroke]

Heh. Such energy and blind conviction! This isn't a religious discussion, y'know.

The physics behind the spectrometer is relatively simple. At the grating:

d sin theta = m lamda

- d and m ae constants, so the deflection is proprtional to the wavelength. This is a linear relationship.

If you project the emergent beams (or the spetral image) on a spherical (2D) or cylindrical (1D) plane, then the corrolation between wavelength and beam displacement is linear. If however, you project the emergent beams onto a flat plane to form an image for a camera then the linearirty is gone, but the software could easily compensate for that. I don't actually know if it does or not.

Of course, poor lens quality in the camera can introduce other errors in wavelength, but empirically, these seem minimal.

It's mildly offensive to hear that you know I'm lying, but - hey ho - I guess you forked out a fortune for a fantastic spectrometer (good for you) and are feeling threatened. I don't have that klind of money so I couldn't, but even if I could, I certainly wouldn't, because my £100 unit does exactly what I need it to do.

The i-Phos is made by a guy who has a Youtube channel, and that channel shows the i-Phos in use in a variety of settings. One of his videos shows a direct comparison from a Scandiavian customer comparing his source's spectrum as collected by i-Phos and by an Ocean Optics machine and correlation was pretty impressive.

 

[paul1598419]

You missed my point entirely. Yes for a single wavelength the diffraction grating uses a linear equation to work out the wavelength. But, a spectrometer is not linear across its wavelength range. If fact, the equation used to calibrate them is on the order of X^3. Very nonlinear.

 

[FlatBroke]

Actually, I didn't miss your point at all.
As I said, the equationm describing the physics in play is:

d sin theta = m Lamda

lamda is the wavelength of the light. Theta is the angle of diffraction at the grating.
For ANY AND ALL WAVELENGTHS - the angle of diffraction is proportional to it.

In practice, of course, only some wavelenths are visible and well-behaved in the optical path, but that is true for all visible light and nearby IR and UV too.

But linearity is not even the issue. In the case of the i-Phos, the camera sensor simply captures an image of the spectrum.
It is the software which INTERPRETS that image to infer wavelengths, which it then displays on a dynamic graph,.
The raw physics package could deliver extremely non-linear responses and it wouldn't matter a tiny bit
because the software can characterise the response of the opics and correctly calculate and display wavelengths.

This is exactly what happens in the Ocean Optics machines and other lab instruments.
ALthough Ocean Optics use a linear array CCD sensor, the principle is the same.
The webcam sensor is simply a 2D array which is using only 1D, except that in fact, it uses more than one pixel strip and in this way is able to dramatically improve sensitivity, integrate over time for long exposures, improve noise immunity and more.

With the i-Phos software you can choose how many pixel strips are used to collec the espectrum data and it will integrate them for you.
Typically, I choose 10 strips to integrate for a nortmal exposure but when I want the best resolution I go down to a single strip.
This improves resolution by minimizing lens artifacts but it lowers sensitivity.

 

[paul1598419]

No, you really don't get it. The response from any diffraction grating is nonlinear across the spectrum, so if you aren't using the third order equation to calibrate it, no software in the world can compensate for your linear representation of a very nonlinear response. You are expecting something that is impossible. Go to one of the several threads on spectrometers and you will find a good explanation of this and why it is what it is. I have done this in threads, but am not going to do it again just to show you the math involved that you are unaware of presently.

 

[hwang21]

Here's one such paper that briefly mentions an Ocean Optics spectrometer and its nonlinear response: https://www.americanlaboratory.com/...ecision-Calibration-Method-for-Spectrometers/

 

[paul1598419]

That is a good article for people using fewer lines to calibrate a spectrometer than the minimum 7 lines, usually more, that I use. Read the last sentence in the conclusion if you still don't believe.

 

[FlatBroke]

OK well thanks for your time.

 

[kecked]

I used 19 lines to calibrate mine. Nice having a dye laser and a Nicolette spec next to it.

 

[paul1598419]

Yeah, I would have loved to have that to calibrate mine, but I had to use what I have on hand. It is important to get your two outer limits set as these don't extrapolate well. None of the ones I have worked with do.

 

[kecked]

Use a mercury lamp and a neon bulb. Add a hene. That gives you at least ten lines. People make the mistake of using dpss and diode sources that drift around. Atomic transitions do not move. You can then check accuracy by looking for doublets to see if you can split them.

 

[paul1598419]

I have. Also a Na vapor as well.