Sciencemadness Discussion Board - DIY FTIR

DIY FTIR

SupaVillain - 3-11-2014 at 20:28

Okay so on all the other FTIR threads I found in the search here and on google errbody's like "the hardest part is building an interferometer" and searching into these interferometers all I get is these

http://www.instructables.com/id/Do-It-Yourself-Lab-Ware/step...

now whats the difference between that and what is really required

Marvin - 9-11-2014 at 04:36

That set up gets you interference fringes, which is equivalent to the information in a few micrometers of path difference. To get a useful FTIR spectrum needs of the order of 10mm path difference. This is usually done with a travelling mirror tracked within a fraction of a micrometer and absolutely parallel. Far IR is absorbed by glass, so it also needs different optics. The KBr/Ge beamsplitter used in commercial instruments is sensitive to moisture.

Building an optical interferometer with a travelling arm is feasible and the tolerances for FIR are easier to reach but that may be balanced by the low availability of FIR optics. Then you need to add FIR sensors, electronics and computer control. I think this would be a big project and the kind of thing that with a big effort would work reliably once.

SupaVillain - 9-11-2014 at 08:19

Honestly the hardest part is finding out what I actually need in order to build one.

If I know what I need to get and what I need to do, then it's made. Everything electronic and computerized is not hard for me to accomplish, Im an IT major and know a lot of people in the field.

SupaVillain - 9-11-2014 at 12:34

FOR THE SLIDING MIRROR

okay so Ive found theres simple programmed arduinos that can control linear actuators, or small electric motors.

im thinking small electric motor would prolly be better for this "fractions of a micrometer" movement using VERY FINE AND TEENY TINY TOOTHED GEARS and can be tracked simply by how it is commanded, in programming this would tell the motor to move it "x" micrometers and a couple lines of code could tell it to retain the amount it moved and display it back to you, even though you already know its movements if youre typing them into command it. Im assuming these movements would be prepogrammed anyways though instead of entered in by the operator.

Now as far as planning for every other component, i need to know details, for example exactly what spectrum, size, and intensity of IR laser light for the laser. This would be easier if i could reverse engineer a broken FTIR but the parts are so expensive there really is no such thing i dont believe. Havent done extensive research but have only found pdfs to be like user manuals and not designs or specifications n such

SupaVillain - 9-11-2014 at 12:37

Just realized how small micro meters are, gonna have to see how others do it in order to make something that can be that accurate

SupaVillain - 9-11-2014 at 12:41

Annnnd google helps me find out about "nanometer positioning stages" so still proving to be possible. Learning as im going.

so actually now im seeing what you mean by optics, an "optical linear encoder" preferably via arduino is gonna make it happen

[Edited on 9-11-2014 by SupaVillain]

[Edited on 9-11-2014 by SupaVillain]

SupaVillain - 9-11-2014 at 18:50

So as far as the infrared detector, there arent any sources for this info on wikipedia but i guess it suggests the following....

"In IR spectrometers the pyroelectric detectors are the most widespread."

-For the sliding mirror again, this may actually be moved by something that doesnt move it by nanometers per second for example, and could just be moved by an electric motor, i think as long as the movements are TRACKED nanometer by nanometer

[Edited on 10-11-2014 by SupaVillain]

smaerd - 11-11-2014 at 10:28

This is one of those big dreams I've had for a long time. There are some hurdles but it is very possible. The biggest hurdle imo is the detector. Those puppies are expensive. I'm not sure what could be built in a home experimenters budget from commercial parts.

You're right in that you don't have to position the mirror nanometer by nanometer but instead can 'track' the movements to that resolution while recording information. An idea for this that comes to mind would be a very fine screw thread. Yes there would be hysterisis but if secured properly it could play the part. The ones I've seen/fixed in commercial FT-IR's at my university had some sort of linear actuator with mounted springs to reduce vibrations and made their 'scan routines' very quickly. A home instrument could afford to sacrifice some time (longer scan's) if everything else can be mounted and handled properly. This would likely be important for data chucking over USB to a computer for example.

Another big issue is simply housing the optics/mechanics securely. It's really easy to over-look this but before making the plunge for parts have a game-plan for this. Even the most basic dynamic optic paths are not easy to make reliable.

There's a lot to consider but it's totally possible. I'd love to see some attempts here. If I had the free-time and the money and didn't have a million other projects to finish I'd be right there with you.

Edit - also, I'm not sure if a laser source is necessary. I think heating certain metals (nichrome type stuff through joule heating) is a common way. Those are called Globar sources. A laser can be used to align the optics though so it's not flying blind. Would be a lot cheaper as well. A Globar source should be very cheap to construct with a little research.

Edit2- yea know what I might need to make a globar source for a different project I have been working on. If I go that route I will share my findings/efforts.

[Edited on 11-11-2014 by smaerd]

Marvin - 14-11-2014 at 13:44

The HeNe laser in an FTIR has a use, by tracking the interferogram of the this line it can be precise about the position of the mirror. Tricks in the optical path can do this in phase quadrature. Runners made from metal dowelling and ball bearings and pushed by a micrometer have been used for optical interferometers with success. These parts are usually made to high tolerances. Might be that a CD-ROM sled would work. There is apparently something to be said for angle aluminium glued to a flat surface to fabricate the rest.

WGTR - 14-11-2014 at 15:23

I know nothing about FTIR, but for nano-positioning, using piezoelectric actuators is a standard way of doing it. For example:

http://www.physikinstrumente.com/technology.html

The displacement of the piezoelectric stack changes with applied voltage. Different configurations work in different ways. Some devices flex, others compress. You don't have to use the world's most expensive actuator:

http://www.steminc.com/PZT/en/stack-piezo-actuators

Little_Ghost_again - 15-11-2014 at 11:59

Piezo is also used in some microscope stages like some electron microscopes. They can be very very precise

SupaVillain - 15-11-2014 at 16:49

Thank you so much for the help i will make sure that i show ya'll what I come up with.

i think ive decided for the optics i can probably buy the KBr windows (so that i dont have to cut crystals perfectly or polish them) and coat them with germanium myself (using PVD of some sort and a cheap calotester to assure the perfect coating job is done) - this will be the beamsplitter which is the most expensive part i think, therefore the most vital part to work on making cheaper. For $100 i think i can order a couple KBr windows.

yes the fact that i dont have to move it with nanometers but just TRACK it by some amount of nanometers is a huge help.

I would love to start a small business just doing analysis with this FTIR and maybe move on to making UV-Vis and other spectrometers.

a person has informed me of a problem called "chirping" that is fixed with a "compensator" in which the person had to cut an identical piece of KBr crystal to place before the detector.

the hardest part is finding all the required information in high detail, no other task is as difficult when determination is involved.

SupaVillain - 15-11-2014 at 17:20

Im pretty sure theres a lot of schematics/wiring diagrams for FTIR's online via google images. At first i had imagined circuit boards and software but now that i see how it works and is built i think i may just need a switch to turn every component on and activate it, then from the detector the information is sent to the interferogram which im sure theres open source programs for

but maybe this needs its on switch and then an actual button to activate an arduino that controls the mirror, and the completing of the full range of motion that the mirror makes will determine the start and finish of the scanning of the chemical sample.

IrC - 16-11-2014 at 21:18

Quote: Originally posted by Little_Ghost_again

Piezo is also used in some microscope stages like some electron microscopes. They can be very very precise

I wonder if the parts from a DVD optical assembly could be used. The lens is adjusted by current variation on a coil which magnetically moves the focus. Also I found a link which may be useful on this subject.

http://www.fc.up.pt/pessoas/peter.eaton/tutorial/webCT/

Little_Ghost_again - 17-11-2014 at 06:35

Quote: Originally posted by IrC

Quote: Originally posted by Little_Ghost_again

Piezo is also used in some microscope stages like some electron microscopes. They can be very very precise

It might be worth looking at hard drive heads as well.

SupaVillain - 18-11-2014 at 20:33

Infrared Source:

globar is best, but will require water cooling or something or the sort.
"Advances in ceramic metal alloys have lead to the production of Globars that no longer require water cooling. However these newer Globars are typically not operated at as high a temperature as 1300K"

Beamsplitter:
i've actually found in catalog, rectangular KBr crystals that are like almost an inch tall and almost two inches long, for $40 so im thinking of cutting these in half and making $20 KBr windows that are about a square inch big. Then with prolly "Ge" magnetron sputtering I'll put the germanium coating on them.

As far as testing the coating, every coating thickness test like calo test and such seem to require some kind of liquid to work like diamond paste or whatever and i worry about the KBr window taking in all the moisture from that. Either way I would only translate these test results to the light intensity of the transmission half and the reflection half, and to the amount of time spent in "Ge" magnetron sputtering. So what im saying is it's pointless to test the thickness if i can just compare seconds spent in sputtering with the actual correct performance of the beamsplitter. This would be tested with a photometer, comparing light intensities of the relflected and transmitted beams.

also, ebay is a very good friend for finding things cheap, but always remember when ordering KBr type stuff, it's safer bought from the manufacturer because of how hygroscopic it is

SupaVillain - 18-11-2014 at 20:42

Globars are also called non inductive resistors it seems, maybe theyre a type of those. Only problem is that these things look like straight tubes, how do I hook up and use something like that????

Either way the most modern IR source is a globar that does not need water cooling. All this heat and water cooling combined with KBr and electronics has me worried...

after looking at this page: http://chemwiki.ucdavis.edu/Analytical_Chemistry/Instrumenta...
i realize ill have to look into "air bearings" as they say they use for the sliding mirror

also a "collimator" can be addd to the front of a globar or any IR source so that it can be more direced in a path rather than omnidirectional

[Edited on 19-11-2014 by SupaVillain]

SupaVillain - 20-11-2014 at 08:21

Theres something about having had to use mechanical bearings and ruby ball bearings for the sliding mirror to be able to oscillate every millisecond, and now currently they use air bearings because theres air in between the parts making them longer lasting from not having friction and crap like that, however ive heard otherwise about them....

Now theres also the sample compartment that im gonna have to look at but i think thatll be a simpler part compared to others.

Now the worst and most hardest part possibly.... is the detector. This is because you cant just buy a little infrared detector, those only do like 100um to 1mm or something, right before the wavelength of microwaves. Not even in the same ballpark of what we need. In order to get what I need, im either going to have to "liquid phase epitax" some HgCdTe, or the cheapest and most likely thing that I do is to make a pyroelectric bolometer with PZT or LiTaO3 crystals then figure out how to hook that up to electronics that convert the beam into electronic signal that goes to the interferogram. Deuterium for DTGS detector aint happenin.

SupaVillain - 20-11-2014 at 14:26

Check out this file for a basic idea of what to use for a detector and why,

http://shop.perkinelmer.com/Content/applicationnotes/app_lit...

so basically DTGS is a standard, but LiTaO3 is "at least as good" except for S/N which can be made up for by running more scans with a little more time. The benefits massively destroy DTGS as a competitor. Also cheap cheap cheap! Or cheapER at least...

SupaVillain - 23-11-2014 at 10:17

Ive realized i can get a small lithium tantalate crystal for like $50 bucks but then i still have to build the tiny electrical component with its electrodes and field effect transistor crap, at the time i dont know anything about electronics and if i just pay $50-60 bucks more i can get a lithium tantalate detector that I know will work and will be far more effective than what I can make myself due to the fact that these manufacturers have all put their little signature thin films and other such pieces on that they have designed through extensive research.

ive pretty much figured something out for every single component in my head but now im trying to see how everything will work as an assembly with the electronics

im also thinking of keeping the IR source as a separate outer part instead of it being right up close with all the other parts because of how hot it is

ive found HeNe lasers at 632nm and .8mw to 5mw for like 40 bucks or so, are these strong enough? There are lasers just like these available for $1000 which is ridiculous, i hope i dont need a certain strength of laser

OK, so this kind of defeats the purpose

quantumcorespacealchemyst - 28-11-2014 at 09:51

but http://www.ebay.com/itm/PERKIN-ELMER-1600-FTIR-SOFTWARE-AND-...

SupaVillain - 28-11-2014 at 12:55

I dont understand why you would post that it doesnt defeat any purpose it's just spam. What you posted is literally exactly the reason why this thread exists. Im in the 300-500 dollar price range after pricing almost every single part and Im gaining and will soon share the knowledge of how to prepare this entire machine for cheap.

You have achieved nothing but insist that my efforts are valuable in the thousand dollar range. From what Ive learned and the capabilities one gains from the experience of building something like this, this FTIR, and several spectrometers of other kinds can be made for the same price as buying a single machine that you have no reason to believe can work as it's bought from a stranger off the internet. They always say never buy a used beamsplitter so theres also a high risk of having to repair that, as well as other certain parts. Also consider, why are they selling it? Maybe because their laboratory cannot build enough profits to pay off their bills as they have thrown too much money around on making themselves look fancy with crap like this. Also, they have to pay to repair this thing, which isnt cheap. If something breaks in a machine that you built yourself, youre going to know EXACTLY what went wrong and how to fix it. Then you dont have to have your business inoperable until the repairman can schedule an appointment with you and order repair parts from a facility that takes days to process and ship your part.

I can probably make an entire other FTIR for the price of one repair after labor, parts cost, and not losing business over waiting for service. Maybe Nicolet has a warranty that says theyll work on your machine for free. Still gonna take a long time before you can operate it for your business.

And no i dont care if anyone is going to come to me and tell me that the way Im doing things works for me and not for others.

I dont do things MY way i do them the RIGHT way, in which all else is inferior.

also, not everyone can afford to buy something that overpriced or wont be willing to gamble with taking a loan out for one. In this way however, it can be affordable and operable for those who are willing to do the work it takes.

Id like to put on my resume that i can beat the price of these big companies and make the exact same thing they do with the same quality analysis.

smaerd - 28-11-2014 at 17:20

I am pretty gungho on home made instruments for a few reasons. Go ahead buy an old FT-IR on e-bay. I would if I could afford one. The real downside is say, if a detector goes bad a replacement is 3k+ or an opportunistic(or should I say optimistic) ebay purchase. Whereas a home instrument can be modified/improved on and fixed when something goes wrong. Almost always for significantly cheaper. There are always drawbacks though. Cost is the most prohibitive for me, however, lately time is becoming another. But to get back on track...

I don't believe the laser needs to be high power ( <= 1mw). I think it's moreso about the beam quality/character. Maybe also the pulse time (probably not though). Keep in mind the purpose of the beam.

[Edited on 29-11-2014 by smaerd]

SupaVillain - 5-12-2014 at 18:28

Yea i mean i think all that matters for the laser is that it's "632 nm", i mean the stronger ones burn things and are used for engravings. Not trying to burn every component and cost hundreds in damage, so it must not need to be one of those.

DOES ANYBODY KNOW HOW FAST THE SLIDING MIRROR MOVES??? what is the full action between the moving mirror and the detector? For example, does the detector receive the input signal as a single picture as the mirror moves from point A to point B or does it "take a picture" every single micrometer of movement until the whole movement is made? This is pretty much the area in which i know nothing about and dont know how to find. The electronics are the most complex but, like i said i will learn them with time.

smaerd - 5-12-2014 at 19:26

I don't see how the velocity of the mirror would really matter. Seems to me, more like a variable you should tune based on limitations of your design.

Edit - Unless you choose to design the instrument by interpolating the mirrors position from say it's velocity. Then you'd want to tune the velocity of the mirror based on how fast you can sample/process the data. I don't think the 'flying blind' approach would ever be acceptable. Even a lot of position tracking method's have downsides which can give outright wrong measurements. So definitely look into how you'll handle that.

A good starting point might be to first see how long of a throw you will need, and then see what your options are for inducing the motion. Then see what your options are for tracking.

Keep in mind that mirror's motion is the essence of the instrument. Definitely recommend looking up some of the math and crunching some numbers with available components before moving forward.

Good starting point -
http://www.shimadzu.com/an/ftir/support/tips/letter8/tech.ht...

Great over-view -
http://mmrc.caltech.edu/FTIR/Understanding%20FTIR.pdf

[Edited on 6-12-2014 by smaerd]

[Edited on 6-12-2014 by smaerd]

SupaVillain - 5-12-2014 at 20:18

Apparently in order to learn something I have to be stumped and then post that i cant find it, then immediately after i find it!.

https://www.youtube.com/watch?v=3hlp401JN0M

- this video is a jackpot of info because it actually tears down an FTIR (from 1989) and explains a lot about its parts. I had no idea that the laser was used for tracking the movement of the sliding mirror and that therefore an expensive air bearing with expensive programmable control isnt necessary. That makes the electronics part way easier. Setup of a nichrome wire thingy may be simpler because i dont know anything about setting up a globar. Back to electronics, they are more complicated than the current times and reality that we live in today because in this video the computer hardware of this old machine wasnt as advanced back then and im hooking this thing up to my computer anyways instead of having it use a computer of its own. Nowadays, 1 or 2 chips of RAM could replace all that mess of RAM chips they have... Also this video sheds light on the requirements of the resolution for the ADC chip (analog to digital for detector to computer connection), being 16 bit, but the guy in the video said this was an NaCl and ZnSe FTIR instead of KBr and the detector quality didnt seem to impress me so Im going to say that the ADC chip needs AT LEAST 16 bits of resolution... more resolution cant hurt but looking at these chips, more of that means more things gotta be hooked up to all kinds of crap.

Another video i saw on youtube introduced the OMNIC software which i believe is open source (free) and the guy in this video explained that the interferogram is more of a recorded thing over a short span of time like a few seconds. He stated that the moving mirror moves slowly. Ill be looking into more of the details on the moving mirror (how far it travels and how its movement affects the interferogram output so that i find out the tech specs of it and therefore see how i need to design it

SupaVillain - 5-12-2014 at 21:19

Smaerd i didnt see your post until after i posted again, i will look into those things thanks, another thing i have realized might gimme info on how the moving mirror affects the interferogram is to actually download the OMNIC software and tinker with it, also looking into OMNIC software tutorials might show users tweaking the instructions of the parts and the displying an outcome that will tell me something.

I also keep forgetting to address the component of the sample compartment in which I found the attenuated total refraction (ATR) is the best for this because it totally takes away sample preparation but I don't know how expensive this is compared to traditional sample compartments. This link below is valuable in that it shows a diagram of how the beam will pass though it and lead back into direction of the detector.

http://chemwiki.ucdavis.edu/Analytical_Chemistry/Instrumenta...

smaerd - 6-12-2014 at 07:33

ATR is wonderful but expensive. Home-made ATR is definitely possible, still probably expensive, http://link.springer.com/article/10.1007%2Fs00897970140-X /

"‘Do-It-Yourself’ Attenuated Total Reflectance Cell Designed and Constructed in a Laboratory Course: A Versatile and Economical Alternative to Commercial Designs". The Chemical Educator
October 1997, Volume 2, Issue 4, pp 1-16.

If I had access to springer I would have gotten the article for you.

For testing the instrument prior to sample cell development thin sheets of polystyrene are often used as sample and are of course very cheap and easy to find.

I think the mirror's speed depends on what the operation is. For scans it's probably pretty slow(but a nice constant rate to achieve a uniform dataset), but when I fixed a few FT-IRs they were doing a quick jerking motion during the start-up routine. Maybe it was a calibration routine to find where the mirror was located.

SupaVillain - 6-12-2014 at 13:19

Honestly i think that i feel pretty confident about the sliding mirror from its movement to the final product of the interferogram, from the links you provided. Thank you! Seems like the ultimate output interferogram is the averaged or high resolution result made from thousands of little lower resolution scans from across the whole movement of the mirror.

all that really maters.... is that the final interferogram is made of a large amount of detector scans, for example, a scan at every micron increment passed by the mirror if your actuator has micron resolution. If you really think about it, theres no reason why making this slower or faster would make a difference in the final product. The highest resolution is found by taking a detector scan at every smallest unit amount of distance that the mirror can take steps by. However, the mirror's movement doesnt have to be a bunch of little steps, it could be a continuous movement in which each detector scan is paired with the recorded distance it has traveled at the time of the scan. This may be more difficult due to lagging of the processing made by the electronics.

As far as tracking the movement though, somehow this supposed to be done with the HeNe laser and a "fringe" thingy. I will look more into that next. I really hope to make a cheap sliding track and have all the technical tracking figured out by the hopefully strong and high resolution capabilities of this laser fringe combo thing.

aga - 6-12-2014 at 14:03

The ADC sampling time will be the bottleneck in a low-cost rig.

If the mirror moves faster than that, you'll just miss samples, lowering your overall resolution.

E.g. for a PIC16F1823 that i happen to be using at the moment, a reliable ADC reading takes 1us.
From experience, i would take at least 16 readings, bubble sort, and use the median value, so an absolute minimum of 1uS, more like 100us per reading, so the mirror has to move slower than than between positions.

Is the 'IR Source' really just a heated wire ?

SupaVillain - 6-12-2014 at 16:22

Yes nichrome wire, i mean its a heating element, IR radiation is just heat basically.

By 1us you mean 1 microsecond right?

Yea honestly though a lot of the price of the cost of an FTIR is in the wage hours taken to design it, make it, and test it. But since I love doing this crap i dont count that in on my own self. Im counting the cost of materials though, and thats the honest cost of em, with even mostly highest quality components, for example a Ge thin film on KBr substrate beamsplitter

SupaVillain - 7-12-2014 at 12:18

It's hard to find a way to count the fringes since they dont really sell something for that commonly.

Edit - lmao and then i immediately find out that what im labeling the as is incorrect for rhe industry terminology. Ebay and google searches produce a wealth of info for "laser distance sensors".

[Edited on 7-12-2014 by SupaVillain]

m1tanker78 - 7-12-2014 at 16:52

SupaVillain: What do you plan to do for optics? The expense has always put me off. I planned to attach the moving mirror to a speaker cone (instead of a proper linear actuator) with enough dynamic range, calculate latency from momentum and electrical path(s), plug that into the software, and 'track' the position of the mirror by the phase of the sine wave applied to the speaker. Naturally, the zero crossing point would be the center of travel. With some tweaking, one could find an ideal frequency or range of frequencies to drive the speaker. Amplitude (deflection) would also be a controllable parameter.

With a 16 bit DAC, I figure the resolution would be more than adequate once the dynamic quirks (jitter) of the speaker and driver are worked out.

Tank

SupaVillain - 7-12-2014 at 21:04

Not sure which part of the optics you mean lol im new to this stuff, i plan to use highly reflective mirrors that i can buy or make myself with PVD or CVD, a Ge on KBr beamsplitter and i think they call them "parabolic windows" which take the IR radiation and focus it into a skinnier beam so that it fits the size of the detector. Ill be ordering my KBr crystal lenses instead of trying to cut my own but ill do my own PVD magnetron sputtering to deposit the thin film of Ge on them. I mean theres a site you can order KBr crystals like almost 1 inch by almost 2 inches for like 40 bucks and i was gonna cut that in half. Not sure what the cost of magnetron sputtering is but ive seen some pretty ghetto DIY setups on youtube

So far ive planned that the detector will be pyroelectric lithium tantalate crystal from here http://www.silverlight.ch/laser.php

as far as the tracking of the moveable mirror... that is what i have just found out a great deal about...

the HeNe laser i used entirely to generate the whole "interference" part of the interferometer. The thing about interference is that it's basically the two split beams giving off these waves called "fringes". As you move the sliding mirror in its linear path you will find these fringes to keep coming out from the center of the beam.... and counting these fringes helps you track the movement of the mirror. In order to digitalize this, a phototube, more specifically a PHOTOMULTIPLIER is used to count the fringes somehow. Im currently looking at how these photomultipliers are hooked up and used for this purpose. Now thats tracking the movement, and as far as actually MAKING the movement, im going to try to go as cheap as possible with something like an arduino step motor or maybe do a periscope type thing like in the FTIR teardown video i posted. I personally think the quality of movement doesnt matter its how high resolution you can track it that matters, because that simply will see all that was done in the movement and pair that data with what was scanned by the detector

smaerd - 7-12-2014 at 21:51

Jeese I don't know supavillian. I would really reccomend using more things off the shelf, especially when it comes to optics. I'm sure that you're aware that KBr will degrade in moist environments (IE air) over time.

I would also recommend avoiding a photomultiplier tube at almost all costs. I don't recall seeing one in the FT-IR's I've peaked in too. I'm sure there's another way to handle what you're trying to do.

Can you find any spectral data on the detectors you listed? I didn't see any on the website.

aga - 8-12-2014 at 01:49

ISTR that digital vernier calipers use a fringe-counting technique, and they're really cheap these days.

Obviously they achieve pretty fine and repeatable resolution.

http://www.ebay.co.uk/itm/Digital-Vernier-Caliper-Gauge-Micr...

SupaVillain - 8-12-2014 at 11:08

Hmmmm. Yes i mean the whole point of the laser is to make the fringes and make them countable upon the moving mirror but it really isnt required. Ive just been curious as to see if thats how this could really be done.

Yes but smaerd why wouldnt a photomultiplier work? I think the cheapest movement tracking is to get one of those analog knobs that displays how far its been turned, as it turns, upon an arduino, very simple cheap setup. Unless i can figure out how to get the "caliper tong teeth things" from the caliper thing that aga has mentioned, to be moveable along with the mirror, and i guess a hookup could be made from its electronics to plug the distance values into the FTIR software instead of sending it to its little LCD display

i think im gonna just give up on the photomultiplier because i not only would have to figure out from a scarce info resources how to hook it up and put it in an environment in which it could actually read the fringes.

Might not even have a point for the HeNe laser at all anymore!!! Is the laser used for the analysis as well or just for the mirror movement tracking? It is valuable to remove things from the recipe list that you wouldve had to buy before!

SupaVillain - 8-12-2014 at 11:26

So yes the vernier calipers do provide measurements as you move the prongs continuously as you would need, but i may look into higher definition alternatives for this.

[Edited on 8-12-2014 by SupaVillain]

aga - 8-12-2014 at 11:54

higher definition = more gears maybe ?

i suppose there's a limit, but increase the gearing ratio by 100:1 should mean increasing the resolution by a factor of 100.

[Edited on 8-12-2014 by aga]

m1tanker78 - 8-12-2014 at 12:43

SupaVillain, it seems you're a visual learner. Here is a random video I pulled from yt (though probably a bit extreme). Apparently, the audio guys call it 'excursion'. I call it deflection but whatever. Imagine the moving mirror attached to the center of the speaker cone. You'd have to google terms like 'sine wave', 'phase', etc. You could theoretically do away with optical or mechanical tracking of the moving mirror by simply keeping track of the angular phase (google) of the signal that drives the speaker. Any laser light that operates in the visual range of the spectrum could be used to align the optics. Even those $2 cheapies from the corner store will work.

The fringe thing as I understand is simply a standing wave but could be used to confirm that the incident beam does in fact interfere with the phase-shifted beam.

https://www.youtube.com/watch?v=ZI6C_4KHtnM

SupaVillain - 8-12-2014 at 12:43

Yes either optical encoder or gears like that, sometimes i worry that these gears may not be exact or precise enough but then when I think about a manual watch or clock, i guess there's nothing to worry about

SupaVillain - 8-12-2014 at 12:47

Tanker i didnt see your post until i posted, what youre saying sounds very interesting but yes i am a visual learner so much googling is about to happen of the terms you just brought forth

SupaVillain - 8-12-2014 at 13:07

Yo tanker i think i found what youre talkin about with the mirror at the center of the speaker cone and the laser projecting off the excursion measurement, in this yt video here, skip to like 3:30 in the video to see and yea i had no idea what you were talking about until i paired it up with this visual representation lol

https://www.youtube.com/watch?v=EjReQE9B_7E

SupaVillain - 8-12-2014 at 13:34

Okay what I need to know is can the excursion of the speaker be done slowly and make only one movement? Like from point a to point b once instead of this repetitive motion?

aga - 8-12-2014 at 13:42

Hmm.
It *can* be, but to get smooth operation, you would really want to drive it with a pure sine wave.

That would make it move in a repeatable way, without stuttering.

Avoiding saturation at the ends would be necessary, or it'll smoothly bang into end-limit, giving 'strangeness' at that point.

Google for 'pure sine wave oscillator' and there's loads of example circuits.

If you go for a Big speaker, google again for Class A audio amplifiers (class A gives the least distortion).

SupaVillain - 8-12-2014 at 13:52

Yea i dont know this sounds more complicated and not as precise as gears with a ratio and an optical encoder

m1tanker78 - 8-12-2014 at 13:55

Quote: Originally posted by SupaVillain

Okay what I need to know is can the excursion of the speaker be done slowly and make only one movement? Like from point a to point b once instead of this repetitive motion?

Definitely. The resolution is going to depend on your drive circuit. If you use a microcontroller with 16 bit digital to analog converter, you can increment the deflection or 'excursion' of the speaker cone by more than 32,000 increments forward and more than 32,000 increments back. Your AVR or whatever you use would cycle through and create one [almost] smooth motion. You can't drive a speaker directly. You'd need to drive an amplifier of some sort and account for dynamic irregularities of the speaker. Pretty much all this can be done in software.

As you saw in the video, you can measure the excursion directly but if you're going to be measuring very small changes, you could angle the backboard to give higher resolution.

By how much does a typical commercial FTIR move the mirror?

Tank

aga - 8-12-2014 at 13:57

It's just a bit of electronics !

Probably more reliable than a mechanical system too.

Having said that, you could try the mechanical system, and then try the electronic one later, and compare results.

SupaVillain - 8-12-2014 at 18:25

I have massive respect for what methods ya'll are talkin about but i dont even know how to research that stuff lol. I kept diving into dense photomulitiplier research and eventually it led away and towards other more modern technologies. I came up with info about using photo diode arrays or CCD's and such to take a picture of the fringes once theyre projected off by an optical flat or some similar setup. Once you get these fringes as a picture at certain points you can basically use OPENFRINGE software to analyze them and turn it into some kind of displacement data eventually. Theres other software like metropro & mx, labview, and any other "static fringe analysis" or "phase measuring interferogram" software. I eventually basically came to a point where once i got to the fringes analysis, everything turned into software names, and sometimes people talked of the interferogram being made of this, where actually the interferogram and spectrum are made of a combo of this fringe analysis program and the data found from the FTIR's detector. Detecting these fringes and measuring with them can apparently provide measurement at a "fraction of a micrometer" i guess is the best way to say.

SupaVillain - 9-12-2014 at 12:29

Well i think im gonna go with traditional sample compartment since its easy and simple and ATR cant even do gas analysis. However they use a small diamond with the popular "brilliant" cut and do a "single bounce" refection thru it, meaning laser pointed at the bottom right corner and reflects to bounce against the top of it from inside and then back out the bottom left corner, if looking at a horizontal diagram... pretty simple but you may need to focus the radiation down into a point to fit in the tiny diamond (top of diamond where sample sits is usually 1.5 mm diameter) and then the top has some simple thing that screws down onto the sample and applies a little pressure....

but yea im probably going to just do the traditional sample compartment cause i guess i could change it at any time

smaerd - 9-12-2014 at 16:39

That's a really creative solution m1tanker. I guess what I am confused about is this. A speaker is in effect, a solenoid? Why fuss around working indirectly with a solenoid. Sure speakers are cheaper(unless you wrap the solenoid yourself) but if you have to buy an amplifier, tune circuitry, and write soft-ware to correct for a lot of the 'speaker induced issues' why not choose a solenoid directly?

I would be afraid that there would be some lateral or longitudinal oscillations added from a speaker, I have no real evidence to suggest that would happen but I don't think of speaker heads as firm things.

The real question is, how far of a throw do you need SupaVillain? Then an actual mechanism can be thought up.

The idea about using one of those harbor freight digital calipers is a great one, if it provides the resolution necessary. If I finish all the projects I'm working on and no one has done it yet, maybe I'll crunch some numbers for the interferometer specifications... That'd probably be in a few weeks though.

smaerd - 9-12-2014 at 17:03

Supavillain, I don't want to sound disenchanting but, yes ATR attachments are simple in an optical diagram. Constructing one is an entirely different story. Not that they are super complicated things, but I've been surprised by the difficulty of turning pen and paper into reality with even 'simple' projects. Aligning optics and everything else... Also ATR crystals do not have to be diamond at all. Most of the reasonably priced ones are ZnSe iirc.

I'd like to see more about this fringe analysis stuff. Do you have any links?

I recently found one of these researching another users posts. Super interesting, depending on detector costs, might be the way to go.

http://www.pyreos.com/products/handheld-mid-ir-spectrometer....

http://www.azom.com/article.aspx?ArticleID=9726

m1tanker78 - 9-12-2014 at 18:31

Quote: Originally posted by smaerd

A speaker is in effect, a solenoid?

True. A glorified solenoid with a cone attached to it (sort of)

Quote: Originally posted by smaerd

I would be afraid that there would be some lateral or longitudinal oscillations added from a speaker, I have no real evidence to suggest that would happen but I don't think of speaker heads as firm things.

That's a valid point.

Quote: Originally posted by smaerd

why not choose a solenoid directly?

A solenoid usually is happy if it's either fully on or fully off. In between becomes somewhat undefined and therefore, clunky.

A typical speaker has a sleeve that rides on a core. Usually the sleeve and core are manufactured to a pretty tight tolerance (not perfect but surprisingly tight). A typical speaker also has a stabilizing baffle (sorry, I forget what they're really called) at the rim to which the cone is attached.

The result is if you gently press down on one side of the cone, generally speaking the whole cone moves down pretty uniformly rather than badly skewing.

Quote: Originally posted by smaerd

The real question is, how far of a throw do you need SupaVillain? Then an actual mechanism can be thought up.

I'm not sure he even knows. IR occupies a huge chunk of EM spectrum. Best case would be 1mm. Worst case would be ~700nm (0.0007mm!!). I would have jumped on this project if CaF optics weren't so damned expensive.

Could someone explain the detector part of the interferometer? If IR intensity is being measured in the time domain then run through a Fourier Transform to obtain the frequency domain representation, why not use an inexpensive detector? Is it because the cheap detector's window won't pass all the frequencies in the IR range?

Tank

smaerd - 9-12-2014 at 20:04

I'm far from knowledgeable about FT-IR detectors but I do know that the window is important for the over-all range that can be observed (transmission).

I have also heard DLaTGS (Deuterated Alanine Doped Triglycine Sulphate) detectors are the bee's knee's when it comes to mid-IR pyroelectric detectors - http://www.microwattonline.com/images/files/MWA5000.pdf
http://infrared.als.lbl.gov/content/equipment/49-detectors/8... (KBr window)

I think a lot of it really has to do with sensitivity and the semiconductor material they are made out of (cost-wise). I'm not sure by what you mean by an inexpensive detector. I'd like to learn more about suitable detectors in general so if you have any examples I'd like to see. Some of them have built in temperature regulation and things like that which would be really nice (if not critical).

I'm having a hard time even finding pricing on these kinds of detectors. Little worried they aren't off the shelf type components. Anyone have any information to the contrary? I bet there's something out there off the shelf with a low cost that could be used for FT-IR though, even if it required popping a little KBr disc over it.

I see what you mean about the speaker now, that is a cool idea. Might not need too much of an amplifier if the throw is small. Could probably even build a simple amp circuit.

Edit - Sorry forgot to lower-case a letter in the DLaTGS thing and I know someone out there would write a snippy post about it.

[Edited on 10-12-2014 by smaerd]

SupaVillain - 12-12-2014 at 18:57

Yea i honestly dont really know yet, how far the displacement is, however, those DLaTGS detectors would be amazing but i cant "deuterize" things or find deuterium for sale or anything like that. No idea of any of that stuff. Im going for this "eltec model 420-0" lithium tantalate pyroelectric detector at http://silverlight.ch/laser.php and with no window it somehow has full spectrum optical bandwidth.

I would think were using more expensive detectors because of the resolution of the spectrum, also those HgCdTe detectors are hard to make lol. I coul make my own cheap litium tantalate detector but i dont know enough about the little mechanisms of the electronic components in it, like the FET, FET gate, the capacitor setup, im lost on that and i dont want to "wing" the cutting of a crystal and hope i do it perfectly flat.

As far as tracking the thingy with fringes i have been lucky enough to locate the aerial view picture of the exact same FTIR as in the video i posted of the teardown, this picture here, http://www.umich.edu/~orgolab/ir/pic/ftir.jpg , shows you a LOT of useful information. Im talking about the laser path!

The laser fringe sensor is purely and widely CCD stuff and sometimes photodiode array but im just gonna go with CCD. Google "CCD arduino spectrometer" and theres a lot of projects posted about, of course nothing at this level of quality and expense of a project, but it's nice to see how theyre hooked up and respond to light.

Oh and it turns out, the OMNIC software has to be paid for, but as it also turns out, "openfringe" isnt just for analysis of fringes and can take the interferogram and do all the FFT analysis as well. However i do not think it has a library and im still looking for sonething about that, if i have to i could download all the pictures of spectrums from a source like sigma aldrich or those chemical info web sites that just happen to have FTIR spectrum data available, and then place them in an image search engine, for example I'd have the spectrum at hand entered in the search field and it would choose the closest matches in there. To see this work, go try to use google image search, however this would be an offline library for myself.

Thinking of using the pendulum periscope idea for the moving mirror as well.

SupaVillain - 11-1-2015 at 18:42

Well if youre looking to build your own FTIR i think this thread covers all the info you need, otherwise you need to do more research. Not gonna summarize this