Remix.run Logo
The likely cheapest home-made Michelson interferometer(guille.site)
99 points by LolWolf 6 days ago | 58 comments

https://web.archive.org/web/20260109203451/https://guille.si...

zokier 14 hours ago | parent | next [-]

> The two beams (split in the beam splitter—the little rectangle with a diagonal in the center) add or subtract constructively at the output, which yields fringes that are visible to the naked eye. These fringes will move as light from one of the arms of the interferometer takes a longer or shorter path

This explanation is bit incomplete. If you align the interferometer perfectly then it should not have any fringes, the fringes indicate that there is some angle between the light beams. If you get the interferometer aligned then the beam intensity varies as function of the difference of beam path lengths.

LolWolf 13 hours ago | parent [-]

For sure! I didn't _quite_ want to get into the mechanics of it (assuming anyone who is interested would just take a peek at the wikipedia or any YT video).

But yes :)

RationPhantoms 15 hours ago | parent | prev | next [-]

Same physics principles used to measure gravitational waves at LIGO (Laser Interferometer Gravitational-Wave Observatory)but just much, much smaller. Very neat!

mturmon 11 hours ago | parent | next [-]

Michelson interferometry is also used to measure the spatially-resolved velocity and magnetic field of the solar photosphere: https://en.wikipedia.org/wiki/Solar_Dynamics_Observatory -> Instruments

LolWolf 13 hours ago | parent | prev [-]

thanks !

observationist 12 hours ago | parent | prev | next [-]

HN hug of death, oh no!

Wayback machine to the rescue:

https://web.archive.org/web/20260109203451/https://guille.si...

Very cool project!

LolWolf 12 hours ago | parent [-]

oops! ran out of netlify credits. should be good now !

echelon 11 hours ago | parent [-]

This was really cool.

Please do the delayed choice quantum eraser next and post it here!

LolWolf 10 hours ago | parent [-]

thank you!

and ha, that one requires a little more fancy equipment :)

boothby 8 hours ago | parent [-]

Wikipedia links to an article with a fairly approachable looking setup for that!

https://arxiv.org/abs/quant-ph/0501010

LolWolf 8 hours ago | parent [-]

I think getting an electron source and creating a robust-ish adjustable set up is v doable, but is definitely more of a Real Project(TM) than this silly little interferometer :)

DoctorOetker 12 hours ago | parent | prev | next [-]

you can go much cheaper if you use a microscope slide or similar as the "beam splitter". Its not 50/50 so the fringe contrast will be lower, but in interferometry one is typically more interested in maintaining the position of a peak or number of peaks traversed...

alternatively one can use a more grazing sharper angle of incidence to bring it closer to 50/50 beam splitting, but then the internal reflections become stronger and the setup is no longer a nice orthogonal one (but how often is that really necessary for a task?)

LolWolf 12 hours ago | parent [-]

yeah I cheated a little bit, but

> Ok, time to confess: I did cheat a little in calling it the “cheapest” Michelson interferometer, since technically even this beam splitter is like 16 USD, but it is very possible to use a microscope slide instead at the cost of some contrast, which will net out to < 20 cents, even at pretty expensive per-unit prices.

:)

alnwlsn 14 hours ago | parent | prev | next [-]

Neat. Perhaps one can use it to see PLA creep in real time.

(not a joke, I'd actually like to see that)

zokier 14 hours ago | parent | next [-]

Not creep, but thermal expansion is definitely noticeable with interferometry: https://www.youtube.com/watch?v=vupIq4epCQA

LolWolf 12 hours ago | parent [-]

yeah, even breathing nearby this thing (or putting a soldering iron near the paths) will show a visible change!

LolWolf 13 hours ago | parent | prev [-]

Oh absolutely!

I would not actually use this for uhhh, repeatable measurements over any extended period of time!

(If that were the case, I'd recommend re-printing it in a slightly more stable material, or just CNC milling the mounts out of aluminum using some of the ~$1-2K aluminum desktop mills and using some aluminum extrusions as the base.)

bigiain 8 hours ago | parent [-]

I suspect they'd be fairly cheap to get made at somewhere like JLCCNC. Or maybe get them to 3D print them in metal for you?

I wonder if they have enough different metal choices that you could build a thermal expansion compensated version? https://patents.google.com/patent/US8292537B2/en

LolWolf 8 hours ago | parent [-]

i’d probably just get it milled only if you had the CNC handy, the complexity isn’t enough to justify 3d printing it in metal (probably a decent bit more expensive too!)

but at “get it done in JLCCNC” prices I think a thorlabs mount is probably in your future :)

bigiain 8 hours ago | parent [-]

I'd be surprised? If you have spare time, I'd be fascinated to see what their website magic quote tool tells you if you just upload your 3d files. Based on other projects I've seen using them I'd guess under $100...

LolWolf 8 hours ago | parent [-]

sure ! I’ll give it a try a little later once I’m at a computer !

(you can too, if you’d like, the CAD files are all online as .step files :)

ge96 14 hours ago | parent | prev | next [-]

That was a cool physics lab I did (disproving ether) and also measuring the speed of light through lucite... other cooler ones like helmholtz coils, wax lens focusing microwaves, measuring gravity, etc...

Wish we worked on IMUs, I still need to get down quaternions

LolWolf 12 hours ago | parent | next [-]

Yes! Michelson interferometers are an amazing first lab experiment since it teaches you the basics of a bunch more techniques which are handy in more advanced experiments, while still having a satisfying outcome when done correctly (which is not as finicky as other experiments with less fun outcomes).

adrian_b 14 hours ago | parent | prev [-]

I wish that humans were not so easily duped into believing that things for which someone else uses the same name are really the same and things for which someone else uses different names are really different.

The Michelson–Morley experiment was indeed very important, but it has not proved in any way the non-existence of ether. It has just proved that the ether does not behave as it was previously supposed, i.e. like the materials with which humans are familiar.

It does not matter at all what names are used for it, one may choose to name it "ether", "vacuum", "electromagnetic field", "force field" or anything else, but all the modern physics, since James Clerk Maxwell and William Thomson, is built on the assumption that the space is not empty, but it is completely filled with something that mediates all the interactions between things.

Only before the middle of the 19th century, the dominant theories of physics assumed the existence of true vacuum. The existence of true vacuum is possible only in the theories based on action at a distance, like the Newtonian theory of gravity or the electromagnetic theory of Wilhelm Eduard Weber, but not in field-based theories, like the electromagnetic theory of Maxwell or the gravitational theory of Einstein.

It is rather shameful for physics that the main result of the Michelson-Morley experiment has been the replacement of the word "ether" by "vacuum", as if a change of name would change the thing to which the name is applied, instead of focusing on a better understanding of the properties of the thing for which the name is used.

SAI_Peregrinus 13 hours ago | parent | next [-]

"Ether" is a hypothetical substance with certain properties. The Michelson-Morley experiment proved that no substance with those properties existed. There's something else with different properties, so it makes perfect sense to use a different name.

TeMPOraL 13 hours ago | parent [-]

In context of approximately everyone's education, the history goes like this: in the past people believe there's something in empty space, and used the name "ether" for that. You learn that, then you learn that MM showed there's no "something", no "ether", but that empty space is, in fact, empty, which is what "vacuum" means. And then if you pay attention or any interest to the topic, you learn that there in fact is no pure vacuum, there's always "something" in empty space.

The obvious question to ask at this point is, "so ether is back on the table?".

Turns out the mistake is, as GP said, thinking MM proved space is empty; it only disproved a particular class of substances with particular properties. But that's not how they tell you about it in school.

lisper 12 hours ago | parent | next [-]

More specifically MM showed that earth is not moving relative to a hypothetical medium through which electromagnetic waves propagate. So either the universe is geocentric or there is no such medium.

zmgsabst 7 hours ago | parent [-]

Another interpretation is that the apparatus and not just light is made from ether — and so the signal is lost because the measuring apparatus is also subject to the local distortion.

That interpretation is also consistent with LIGO: we can detect those ether disturbances because the distortion of our motion on the apparatus doesn’t cancel the signal in the same way.

jfengel 12 hours ago | parent | prev [-]

It's more than just the lack of material. It demonstrates that light propagates in a specific way that is different from any ordinary material. Light moving in a vacuum is different from a baseball moving in a vacuum. The speed of light is independent of your own motion, which is not true of anything with mass.

ge96 14 hours ago | parent | prev [-]

I thought the whole point was if it did exist the motion goes faster in one direction than the other.

edit: not sure if you're referring to dark matter

yeah I gotta read your comment more thoroughly

lutusp 6 hours ago | parent [-]

> I thought the whole point was if it did exist the motion goes faster in one direction than the other.

No, the idea was that, in a space filled with the hypothetical ether, Earth's velocity through the ether should have been detectable by comparing light beams traveling in different directions.

The null result was very important -- it didn't prove the absence of an ether, it only showed that it wasn't a factor in light propagation.

CamperBob2 12 hours ago | parent | prev | next [-]

Over $3K for a similar setup from Thor Labs (1). Wow, you can buy everything here including the 3D printer for that. Good work!

1: https://www.thorlabs.com/michelson-interferometer-educationa...

dekhn 12 hours ago | parent | next [-]

With products from a vendor like Thor Labs, you're getting a lot of quality and knowledge built into the system. Mechanical engineers, electrical engineers, optical engineers... all of which means an edu kit like that will train a student to be useful in most grad research labs (which often build their systems out of thor labs components).

It sort of depends on what your goal is; personally, I live to see something expensive on Thorlabs, and make a simplified, less accurate, and far cheaper alternative in my home lab. But that's rarely how folks in labs do it- instead, they will focus on getting people to be useful for performing state of the art research, which usually depends on applying hundreds of years of experience to make some tiny marginal improvement, which frequently depends on having extremely precise and accurate gear.

LolWolf 12 hours ago | parent | next [-]

Hopefully you enjoyed the post then!

I think there's just such a huge middle ground that's missing (for funny historical reasons[1]) between "children's toy" and "lab-grade equipment" especially in optics, which is why I was excited to make this my first foray into making a fully 3d printed "useful-ish" thing that doesn't really exist otherwise.

---

[1] This is because most lab equipment was made _in the lab_ back in the 60s or so, and having this technical ability was a huge advantage for many labs. Now, personnel cost/hours are much more expensive relative to equipment, so people will pretty much pay whatever to get lab-grade stuff.

dekhn 11 hours ago | parent [-]

I haven't read the post yet because my work blocks access to the domain.

I agree there is a huge middle ground- for example, I make hobby microscopes at home, and much of my work has been making accurate and precise 2D/3D stages. It's easy to buy great, simple (non-motorized) scopes with good optical quality, but as soon as you start adding motorized stages, or any sort of complicated illumination or filtering, it gets challenging quickly. My actual goal is to track microbes in real time using computer vision, but the professional hardware to do so is out of my price range.

I have spent literally thousands of hours fiddling with one part or another Today, I'm working on a high speed flash illuminator that is coupled to the camera, and it's one problem after another. Reality has a fractal level of detail.

Since I haven't been able to look at your project yet, I don't know if you worked on this area, but I found it really useful to clone the Thorlabs cage system components: https://www.thorlabs.com/optical-cage-systems and specifically https://www.thorlabs.com/item/CXY2A (you can download their 3D model and see that the mechanism isn't that complex).

Another thing I've ended up doing is prototyping in plastic and then having it machined at a place like JLCPCB out of aluminum. PLA is just flexible enough (especially under load) that it can make the results very frustrating.

LolWolf 9 hours ago | parent [-]

ah I see!

Yes this is very cool (hope you make it open source :) but have you taken a peek at the openflexure project? They make a fully motorized 3-axis microscope that is 3d printed and relatively inexpensive (parts+motors+electronics - PLA net out to USD 250?)

it’s very cool! maybe you can also take some ideas from there :)

https://openflexure.org/

dekhn 8 hours ago | parent [-]

I don't particularly like the openflexure approach. Instead, I build stages similar to standard XY stages: https://www.asiimaging.com/products/stages/xy-inverted-stage... which uses standard motion components (linear rail and linear bearings, 3d printer style steppers), all of which gets mounted on aluminum extrusion (4040). Then the illuminator and camera just get mounted on the aluminum extrusion.

It's pretty close to the Thorlabs Cerna, https://www.thorlabs.com/cerna-r-series-modular-microscopy-s... although I just do low-magnification light microscopy.

LolWolf 8 hours ago | parent [-]

I see, what’s the min step size you’re going for?

dekhn 7 hours ago | parent [-]

typically around one micron (which is about the same as one pixel with my objective and camera). I am not trying to take min steps. I'd rather have smooth motion, with fairly high accuracy.

The current system I am building is mainly optimized around scanning large areas quickly; I have already demonstrated that I can create accurate stitches by moving, stopping, taking photo, repeat, but it's slow (due to the stopping) so I am working on an approach that keeps the stage moving constantly, but triggers a bright flash that freezes intermediate exposures. This gives a good 10X speedup over the simpler model of move/stop/photo/repeat but brings in a number of other challenges.

LolWolf 7 hours ago | parent [-]

ah very cool, I see!

good luck! are you posting updates anywhere?

zipy124 10 hours ago | parent | prev | next [-]

As a student at a top worldwide university, I can tell you we order a lot more stuff off Amazon and eBay than you'd think. There's an awkward middle ground where you either buy something cheap or make it yourself because labour is basically free in academia thanks to the large amount of students and staff but grant money is not.

LolWolf 9 hours ago | parent | next [-]

yes ! but it also assumes you have: a good optical breadboard + bench + dampeners, a beautiful set of lenses, all sorts of nice lasers and kinematic mounts and linear stages etc etc

so yes, we _also_ (back in my phd lab) built equipment in that sense, but there was a pretty good foundation of Fairly Fancy stuff already sitting around !

dekhn 10 hours ago | parent | prev [-]

But what do you down with the Thorlabs Lab Snacks? I thought that was the main reason grad students ordered things from them.

CamperBob2 12 hours ago | parent | prev [-]

In some cases, you'll learn more from crappy hardware than you will from lab-grade gear. This is probably one of those cases. This build will require thoughtful attention and debugging/optimization on the student's part in ways that the Thor Labs kit might not.

I mean, it's practically the most basic optical experiment that you can perform. Nobody needs to pay $3K to learn how an interferometer works. It's not a MoT or something exotic like that, it's a beam splitter and a couple of mirrors.

Put another way, it's the difference between building a Heathkit and putting a bunch of parts together that you salvaged from other stuff, for those who are old enough to grasp that analogy.

LolWolf 8 hours ago | parent [-]

I think that’s very much right :)

bloggie 10 hours ago | parent | prev | next [-]

Ironic isn't it, since Thorlabs brought down the cost of optical tooling and made components more accessible - they are the Amazon of optics and remain a cost leader.

CamperBob2 10 hours ago | parent [-]

Yeah, I don't mean to beat up on Thor Labs. As they point out, they don't even make any extra profit on those kits beyond the components themselves. And then there are the free snacks. :-P

If you are putting together some more advanced educational or experimental apparatus, they are pretty much a no-brainer supplier, as you say. But their level of quality, support, and system integration just isn't necessary for something like this.

analog31 7 hours ago | parent | prev | next [-]

I was a teaching assistant for freshman physics lab while in grad school, almost 40 years ago. My co-teachers were all theoreticians, so I bore the brunt of helping the students troubleshoot their setups.

There's a balance that has to be struck between: 1) Equipment that's so perfect that students learn nothing about the effort to get an experiment working. 2) Or so crappy that it's an obstacle to learning anything at all.

Also, the crappy-ness is multiplied by 30 for the number of setups needed for a class of 60 students, assuming they work in pairs.

Oh, the crappy oscilloscopes. They were cheap "student scopes" and their controls were worn out, so they behaved erratically. Since then I've met other people who took freshman physics lab, and they remember the "oscilloscope lab" with disgust.

LolWolf 12 hours ago | parent | prev [-]

ha, thanks!

that one has uhh substantially less drift for what it's worth, but reprinting in more stable material would help that a ton (and still be quite cheap!)

touchadinosaur 15 hours ago | parent | prev | next [-]

Really cool project. I’ve seen sometimes cheap lab clean outs on ebay which can be cool toys, but this is on a different level.

LolWolf 13 hours ago | parent [-]

Thanks ! Still need to upload it to Printables etc. Maybe this will finally get my ass a bit more into gear.

NetMageSCW 12 hours ago | parent | prev | next [-]

I kind of wished there were more details around the tension springs and steel rods.

LolWolf 12 hours ago | parent [-]

what type of details would you like to see? happy to add more if helpful!

bigiain 8 hours ago | parent [-]

I have a related question, how far are you turning the thumbscrews when making adjustments?

I think M3 standard thread pitch is 0.5mm, so to a first approximation that almost 1000 wavelengths (if I have the SI units right in my head, and I'm not 3 orders of magnitude out?). I suspect the left/right and up/down adjustments have as fairly high lever ratio, but I can't imagine you could successfully adjust the in/out distance with any precision (not in the 690nm sense anyway)? Is the in/out distance not important so long as the beams are aligned?

Dunno if you've seen it, but there's a great youtube video explaining how the actuators to align the James Webb mirrors work: https://m.youtube.com/watch?v=5MxH1sfJLBQ including a 3D printable version of them: https://www.thingiverse.com/thing:5232214 "This is a replica of JWST's mirror actuators. Six of these actuators are paired into a hexapod / stewart platform arrangement and used to control 6 degrees of freedom of each mirror segment (tip, tilt, roll, x, y, z translation)."

LolWolf 8 hours ago | parent [-]

Ooh, great question. Usually fractions (~1/20th?) of a turn for alignment, it’s hard to go below that since the mounts are so small and the springs don’t have the tension to keep it super stable. (This is plenty for such a “coarse” set up like a Michaelson but might not be up to par for more delicate ones. This can be improved very easily but it was enough for this experiment!) If you want to observe something on the outputs, you have to do something like exhale on one of the arms or put a soldering iron near one of them—merely touching one of the screws gives you indiscernible output, even if the mirrors are aligned.

Very interesting re: JWT, I will definitely take a peek, thanks !

aj7 8 hours ago | parent | prev | next [-]

I’m an optomechanical engineer and I’m sorry. I’m not impressed. In a Michelson, the single most important requirement is that the optical path lengths of the two arms do not drift with respect to each other, either in length or angle. Having that path length determined by a fused deposition polymer is about the LAST choice I would make. I have a suggestion. Fused quartz rods are relatively cheap. Buy some 6mm rods and a thin diamond blade to cut them to size. Use the 3D printer to make plate-like parts, ‘replicating’ a cage structure. The polymer parts should be used only for components PERPENDICULAR to the optical path. You could even experiment with using embedded rods to stabilize the plates in various directions. So much of hobbyist activity amounts to a kind of adult coloring or copying. Rather, at your own level, try to be a scientist.

LolWolf 8 hours ago | parent [-]

it’s an interferometer with like 5cm arms made for 3 bucks, it’s not made to be anything other than a basic demonstration !

lutusp 7 hours ago | parent | prev [-]

Nice project! This video: https://www.youtube.com/watch?v=5nBY4Y0bicM explains why the original interferometer was designed in the first place: to detect a hypothetical luminiferous ether, before Relativity theory made it unnecessary.

It's important to say the original interferometer was much less elegant -- there were no lasers in 1887. But it did have a solid stone "boat" floating in a little lake of mercury. Not making this up.