Build: a highly sensitive precision level (0.01 mm/m)

[@andybelieves]

Outstanding build and presentation, excellent work!

[@TKC_]

I have a level of this accuracy. It has a similar adjusting mechanism. The knob is marked in .01mm per div. Additionally the lever has a coarse scale to keep track of the millimeters. The fine threaded nut is split and tapped to control backlash. The craziest thing on it is above the spirit level is a prism and viewing window that shows both ends of the bubble in split screen. The reading is accurate when the 2 bubble ends align to each other in the widow. This differential bubble movement doubles the sensitivity as does the magnification it provides. This means the vial is not graduated it doesn’t need to be. You can find a few KZbin videos on it. It was made by Zeiss and can be found with a search of Zeiss adjustable spirit level. I disassembled mine to clean and calibrate it as the instruments grease was hard as rock. It’s an East German model and my example was long forgotten. I took some pictures of the insides, it’s pretty amazing inside. It’s basically a machinist tool made by an optics company.

[@Michel-Uphoff]

Very interesting! Zeiss is famous for its quality and precision. Splendid idea that optical addition. I have the required optics I think. So I will experiment some myself.

[@TalRohan]

this build is so thoughtful and beautifully adjustable....I feel your pain with the cracked vial..I can imagine the bad language that went with that. Thankyou very much for sharing.

[@Michel-Uphoff]

Yes, I cursed quite a bit when that happened 😁

[@TalRohan]

@@Michel-Uphoff I've just bought one of those Vials I think mine is 0.001, it was well packaged but I was being super careful even getting it out of the box. Did you bed this one or is it help on metal surfaces?...I feel like putting them on a softer bedding material (I'm considering silk) will upset the final accuracy because you can't garauntee the thickness of the material when you press the vial into it.

[@Michel-Uphoff]

@@TalRohan I made two steel holders, one for both ends of the vial. The holes in these holders are 1mm too wide. I very gently stuffed the small gap between holder and vial with putty. The holders are not fixated on the vial bed. I used two neodymium magnets between the holders and the bed to 'click' the vial onto the bed. Works like a charm this way. I've done this after the publishing of this video.

[@DaveEtchells]

Beautiful build! I looked them up, was astonished by how cheap the raw vials are, and the extraordinary precision available. It must be a very simple manufacturing process.

[@ikkentonda]

“Well THAT’S an instant subscribe” I thought … only to discover I’d already subscribed! Very nice work. Precisions levels are an often overlooked tool.

[@leanmixture]

Thanks for posting. I enjoy working with my PD 400 as well when I get a chance. Very nice machine. I’m grateful to have one.

[@bkoholliston]

That's great project--very nice! I have watched a few of your videos and subscribed recently and have really enjoyed them. Particularly the relaxing new-age style music. I'm off to search for level vials now!

[@Michel-Uphoff]

Thank you, especially for the appreciation of the music.

[@najroe]

i was taught to use something flexible to hold the vial in a slightly oversized hole so it would not deform (or break). we used mixture of beeswax, talcum powder and pinepitch with consistency of stiff spent chewingum, I likely still have the recipe in a book or notebook somewhere.

[@Michel-Uphoff]

I used some pieces of thick paper to fill the 0.2 mm gap between the vial and the POM holders. I wanted removal to be easy while building/testing. Worked ok, but your solution is better for the final assembly. I think I have a "modern" alternative for it.

[@najroe]

@@Michel-Uphoff that mixture is old but I have worked on scientific instruments that was well over 100 years old and still in good working order with very little work besides cleaning and perhaps replacing a worn part or two, sadly often worn by careless use and bad maintenance.

[@jrkorman]

An excellent build and will be looking forward to a new vial and a test run on the surface plate.

[@usopenplayer]

Really mesmerizing to watch. Thanks for sharing!

[@neffk]

Really nice work. I like your test setup.

[@float32]

So what happened at 17:34 that caused to break? It looks like you somewhat gently put in a little something to hold it in. Are these vials extremely thin?

[@Michel-Uphoff]

Ouch.. It still hurts.. In the video at 17:34 I very carefully pressed two pieces of sturdy cardboard between the POM holders and the vial. That was fine, the grip was OK. But a little further on in the video you can still see a small white edge of the cardboard. So, after finishing the build and off camera, I wanted to correct that small flaw. Should have just pushed it in a little further with a second piece of cardboard, but I grabbed a micro screwdriver with my stupid head and started pushing with it. Crack! 😑 Wall thickness is just over a millimeter.

[@russelldold4827]

​@@Michel-Uphoff Excellent build and video. Your description of how the vial broke means my suggestion of a rotating metal protective sleeve probably would not have prevented the incident. Add 1 to the Box of Learning 😉

[@spambot7110]

3:17 just curious about how the 100mm distance was established, it looks like you just kinda plonked the indicator down. was there some measuring or positioning off screen, or like a mark on the rocker i'm just not seeing? it looks like the indicator stand might be indexing into some sorta hole, but i don't see what's constraining its rotation (and that aside, the position of the pivot point isn't indexed)

[@Michel-Uphoff]

I measured it. Can't show everything when cramming several days of work in less than half an hour of video.

[@petersilva4242]

I just bought a surface plate exact same as your one. Exact same scraping pattern. Mine was way out of flat. I measured 30um deviation when the calibration cert specified 6um. Can I ask how good was your one. I started rescraping but it's full of internal stresses so it's going to sit for a year to settle before I come back to it. Another great vid mate, cheers

[@Michel-Uphoff]

Mine was within the error specifications (7 µm). If your plate is indeed full of internal stresses, you can try to reduce them by keeping the plate in a kitchen oven for a few hours at 200 degrees C. Then turn the oven off and let it cool down slowly. The paint may come off, but that can be repaired quickly. Chances are, that the plate will warp more, so its your decision.

[@peirossmallhomemachineshop5364]

Great project. Great design. Could you share where you’ve bought the spirit level vial? Thanks

[@Michel-Uphoff]

Thank you! I don't think my links will work in your country, so do a Google search for: haccury 0.02 vial Btw, they now have a 0.01 vial for 17 euro's Hope this helps.

[@g.tucker8682]

Thanks Michel, this is a fascinating project. I just ordered a couple of vials, will try my own versions.

[@Michel-Uphoff]

Nice! Keep us posted 🙂

[@DudleyToolwright]

Wonderful design.

[@Michel-Uphoff]

Thank you!

[@lohikarhu734]

I thought that it might be a plan to make the 10x (or 20x) pivot by making a compliant hinge point directly in the steel, eliminating the need for the pin and drilled holes, no need to worry about precision fit, lubrication or wear... To get a really precise 10x (or arbitrary), you could mount the adjustment screw on a sliding mount, to set the length of the pivot arm by calibration with the dial gauge... If you drill the threaded hole for the adjustment screw at a 90° angle to the middle angle if the adjustment "arm", it eliminates most of the change in legthbof the arm as the screw tip " slides" along the arm part.. Pretty tiny change, but, while you're at it, maybe little "tweaks" are worth doing... On a telescope tracking mount, one even mounts the threaded mount so that it pivots, as the adjustment screw runs in and out, so that the angular adjustment over thread travel stays very constant per turn... OK, maybe too much 😸, but, as they say, the 👹 is in the details..

[@Michel-Uphoff]

These may indeed be some alternative constructions. I am not sure whether they work better in practice, but I do know that they are a bit more difficult to implement. In itself, this construction with the rocker and hardened pins is very stable and easy to make. I don't think wear and tear will be a problem. The amount of movement is very small and those two sturdy springs eliminate any play in the hinge points and threading.

[@jobkneppers]

Weer een prachtig project Michel! Het valt mij wel op dat het vele gebruik van magneten zijn sporen of spanen nalaat...Ik probeer het zelf te vermijden. Vooral op het meetgereedschap vind ik ze erg irritant. Ik denk dat jij er ook van terug gaat komen ;-) Dank en hartelijke groet! Job

[@Michel-Uphoff]

Hé Job! 🤗Goed nieuwjaar! Ik zeg nooit nooit, en zeker niet tegen iemand met zoveel ervaring. Maar ik gebruik magnetische hulpstukjes al heel lang. Mijn hele Michelson interferometer die ik 10 jaar geleden bouwde is op magnetische montage gebaseerd, daar zitten er wel 50 in. Het is wel opletten, want krasjes wil je niet. Allereerst zijn de magneten zelf altijd of iets verzonken of afgedekt met een dun plaatje metaal. Daarnaast houd ik ze met buddies (die poster kauwgum) altijd goed vrij van spaantjes want die worden op raadselachtige wijze altijd wel verzameld. En inderdaad raakt er regelmatig iets magnetisch, maar dan zijn twee halen over mijn demagnetiseerder voldoende om dat te verhelpen. Zo'n ongebruikelijke meetopstelling als ik hier toonde, is met wat magneetblokjes flexibel, stabiel en snel in elkaar te zetten, daar kan m.i. geen andere oplossing tegenop.

[@marcusplanlos2037]

Really, really good!! It's amazing what's possible with these small machines. I hope that someday I'll have as much talent as you! Warm regards from snowy Germany.

[@garymucher4082]

When you gently adjust the vile, via the fine thread knob, that makes the vile level but not necessarily the surface it is sitting on. So where do you look to get it truly level and set to use to check other surfaces? I guess what I am saying is what surface becomes the standard to use to set it and basically lock it in to check other surfaces... Thumbs Up!

[@lohikarhu734]

basically, i think that the typical use is for "differential" measurements", where you move this device along a "track" and measure variations along the track, to generate a map of local/global variation, then move "sideways" to the next track; you then do the same on tracks at 90°, and thus generate a 3d surface map, so that you can take the numbers from any point, use that as the "reference"

[@Michel-Uphoff]

@@lohikarhu734 That's right!

[@Michel-Uphoff]

@garymucher4082: It works like with any spirit level: An excellently flat and roughly level surface, that's all you need. If the spirit level shows the same deviation after being rotated 180 degrees, then the spirit level has been adjusted. But that is not the main purpose for this spirit level.

[@gordonwalsh5652]

One neat thing about levels is that they are self-checking. Basically, the only slope that measures the same in both directions is true horizontal. So if you can find a place where you can zero the level, flip it around, and still read zero, that's your horizontal reference. Additionally, on any tilted plane, there will be some position that is level (even if there might be a bit of side-to-side roll, but that's usually not important). This isn't really a guide to finding that horizontal position, but I think it can be done reasonably efficiently with a bit of cleverness and splitting the difference, etc.

[@traitorouskin7492]

I had to rewatch bits then pause and have a think but I get how it all works now. Thanks. I enjoyed watching this.

[@RedDogForge]

Using magnets to hold setups, absolutely brilliant!

[@thechumpsbeendumped.7797]

All this because his wife said the shelf he put up wasn’t level.

[@mobile.cristhianguedes]

😂😂😂😂😂😂😂😂😂😂😂

[@BobbyJobling]

She said: "it doesn't look like it's leveled" He said:" Hold my 🍺 "

[@Michel-Uphoff]

🤣😂🤣😂

[@tjj5726]

Excellent work... didn't know you can just buy these vials. Time to make my own haha

[@donaldvantongeren8385]

You keep amazing me with your utterly precise work, love watching your videos! Not sure about the sequence of the braking though, did it have too much tension and broke later!?

[@Michel-Uphoff]

See my reply to the question from float32 11 hours earlier.

[@colincreedtattoomachines]

Another interesting project Michel, Well done!! So I gather the machining for the various components were done on the Mill "after" adding the epoxy to the column..? Was there a substantial difference with the milling of these, to your previous jobs?? Looking forward to what you'll do with this Precision Level.

[@Michel-Uphoff]

Yes Colin, that's right. I did some fairly heavy (for this small machine) cutting and yes I know there is an improvement. But I still have to measure it to give a more objective evaluation. I hope to shed some more light on wat I had planned for this precision instrument in an upcoming video.

[@car9167]

I always wondered if you can put a microscope camera on the level with graduations and zoom in maybe 10-20 times, put that on a monitor screen and have a mask/raster of some sort so a division spreads over 100mm or something and that way you can increase the precision to maybe 1 micron per 1000mm. Would that even be possible?

[@Michel-Uphoff]

Excellent remark. And yes, that's possible. Before I destroyed (🤫) the vial, I was able to detect movements of less than a mm on the screen under an 80x magnification. That amounts to roughly 1/80 of 0.01 mm, or 0.125µm (125 nm) per meter. But I suspect it then becomes much more difficult to determine what causes those very small movements. Unfortunately I haven't had the chance to dive deeper in that yet, and I'll try to show more about it in my next video.

[@car9167]

@@Michel-Uphoff 5 microns at 1 meter s about 1 arc sec so yours 10 microns at 1 meter is about 2 arc sec. Not bad! A magnification of 10 would give you about 0.2 arcsec pretty good I must say for shop made, comparable with a good autocollimator

[@g.tucker8682]

Interesting, but would you really get the accuracy to match your improved resolution? The vial is produced to deliver an accurate measurement within defined limits. I think of those cheap micrometers with a vernier scale - they may have a resolution of 2 microns (or 0.0001"), but everyone knows they are only accurate to 10 or 20 microns.

[@car9167]

@@g.tucker8682 I know what you mean but not in this case. The only worry you can have maybe is linearity. This is an "analog" device not relying on a magnetic strip or raster on a glass scale which is machined within a precision level. And you don't really "measure" a distance but do comparison within short range

[@g.tucker8682]

​@@car9167Fair point, I see what you mean - this is not for measuring an angle. Instead accuracy, I really am asking about repeatability I guess. Thanks for your input.

[@appatula]

Is that birchwood casey for the cold blue or your own formula? The results are fantastic; what prep methods do you prefer prior to application? Also did you scrape that surface plate yourself? I love studying the minute patterns in hand scraped surface plates; sure I'm a fan of the old fashion "moore pattern" but a hand scraped surface plate is quite a beautiful feat regardless! Achieved with just simple tools, time, and skill to create a precision surface. @Michel-Uphoff

[@Michel-Uphoff]

I used Dr. Galva Ultra-7 Streichbrünierung. First I cleaned the steel very thoroughly with acetone. No, that surface plate came scraped.

[@antesdedormirmanu6789]

Nice approach to the problem...

[@ldftrdnk]

Michel, great craftsmanship, well done. If you don’t mind my asking, what was the music from 6:58 to 18:20 in your video?

[@Michel-Uphoff]

That's from Innovations4u named Gift to the world

[@akosbuzogany2752]

I thought about keeping the vial in a triangular recessed bed and pressing it down with a rubber band or a weak steel spring. What is your take on it? Would it work, or it is more important to have a soft rubber/foam band around on the ends to keep it separated from shocks transmitted through a direct contact with metal?

[@Michel-Uphoff]

I would certainly avoid direct contact between the very fragile glass and metal, my screwdriver was a painful example of this... I used POM and not a more malleable material to achieve a reasonable balance between rigidity and flexibility. Maybe I will experiment some more with steel holders with too large an inner diameter, in which I will secure the tube with silicone sealant. In addition, it is not entirely certain that the outer wall of the tube is really straight. The inside is ground to size while applying the curvature, but the outside is unprocessed. So in my opinion there may be a deviation from the shape that's needed for a secure fit in that v-shaped bed. Then you definitely only want to secure it at the ends, so that almost the entire vial floats.

[@rodbutler9864]

Outstanding

[@madsighntist14]

Very Interesting. I have been looking at Sine Bars, and wanting one on a Micro Scale for my Proxxon M70. I will consider (much) adaptation of this video, potentially at a 50 or 75 mm length for my body.... Thank you! philip from the Great Pacific NorthWET Oregon Division, USA

[@josepablolopezaguado6191]

I wonder how you did the fine thread, I saw you turning the very fine pitch but then using a rather coarse tap, is the tap an exact multiple of the pitch? I would like to learn more about making ultra fine screws, could you give me some pointers? Wonderful video btw!🎉

[@Michel-Uphoff]

You mean the tap that is used around 14:00? That wasn't a coarse pitch, maybe it looked that way because of the poor lighting. It is an M8*0.5mm tap. The most important thing is to take very light cuts, and not worry about a bottom and top radius for the threads. Furthermore, if you use a tap for the internal thread, it is important that you first cut it, if necessary in a piece of disposable metal. Then cut the thread single point, small step by step. I never remove more than 0.05 mm at a time. I also shift the threading bit using the compound two hundredths of a mm alternately to the left and right to only cut one side as much as possible and to take out a very sharp bottom (stress riser). At a top angle of 60 degrees (metric thread), the maximum cutting depth is equal to sqr(3)*0.5*pitch, so at 0.5 mm pitch that is 0.433 mm. But I always stop earlier, because the threading tap has a usually unknown margin. Just run a fine emery paper over the thread to remove the burrs, then clean the thread very well, fit with the nut you made, a few hundredths more, fit, etc. If the nut starts to fit but real snugly after a few turns, see if you can get the fit as good as possible with one or more spring passes. It's something you just have to practice to get good enough at it.

[@josepablolopezaguado6191]

Thanks so much for the answer! I went back and payed more attention this time, the thread looked larger the first time! I like a lot your method for cutting threads, specially using a scrap part and controlling the depth to fit the tap, I think my threads will go much better even when using cheap taps! Thanks for the advice and the videos!

[@АлександрСАП-й7ч]

Hello! Where you bought vial? And also - can you share drawings? Thank you!

[@Michel-Uphoff]

I don't think my links will work in your country, so do a Google search for: haccury 0.02 vial or try: nl.aliexpress.com/item/32870886774.html The only drawings I have are published in the video.

[@daynosdr]

@@Michel-Uphoff I would also like to build one. You dont have dimensioned drawings? I have a hard time believing that haha. I'm willing to cad them for you. or if I do my own im happy to share them.

[@antontrio6517]

In German we say, Tangentialwasserwaage.

[@Michel-Uphoff]

One of those famous long German nouns that describe its operation very well 🙂 When I was a kid and learned words like Eisenbahnknotenpunkthinundherschieber I fell in love with that kind of words.

[@Rob_65]

@@Michel-UphoffOne pf the nicer German words I remember ever reading: Jahresinhaltsübersichtsverzeichnis Very nice built by the way. I have a cheap Chinese machine level that is a pain in the ... to level. I might copy this tangential design and upgrade mine sp it becmoes a real usable piece of equipment

[@lutze4614]

rekristallisationstemperatur

[@caseytailfly]

Sorry the vial broke, but at least you proved the design was effective, and that you are indeed human! Excellent work.

[@Michel-Uphoff]

Yes! I'm human. My voice... not so 😂

[@billshiff2060]

Interesting but that is not actually a dial indicator, it's a test indicator. Those test indicators are only comparitors they don't give exact measurements because of the angle of the probe.

[@Michel-Uphoff]

Unless, as I did, you position the test pin exactly straight, then you can measure accurately. If the pin cannot be positioned straight, you can compensate for the cosine error.

[@billshiff2060]

@@Michel-Uphoff Each manufacturer specifies the angle(usually 0° but interapid is 12°) at which there is zero cosine error but that is only at a single point. Every deviation from that angle introduces an error such as taking any reading other than zero. In your case that error is minute but it is there and you are measuring minute changes. I'd like it better if it was an actual linear dial indicator that reads directly. Ultimately it makes little difference in your application but it is the principal I'm talking about.

[@Michel-Uphoff]

​@@billshiff2060 Mathematically speaking you're right. In practice however the cosine error is nearly zero in this case. It's somewhere in the 8th digit, way below the accuracy of the instrument.

[@johnlambo135]

Prachtig Michel !!!!!!!!

[@Michel-Uphoff]

Dankjewel

[@melgross]

With all that work, I’m surprised you didn’t put plastic pads on the sides to hold it by so that the heat from your fingers won’t affect the readings. All of my Master levels have something to prevent you from touching the metal body. Yes, folks, I know it’s nitpicking. But if you’re going to all that trouble, you might as well go all the way.

[@Michel-Uphoff]

There are four black pads of PVC foam. See the drawing and the last photo.

[@111smd]

i see 2 flaws minor but flaws none the less 1. you will still need to set something in your work shop as level and use that as your standard - you will need to check that your level is adjusted correctly periodically 2. it would be best to use 3 precision ball bearings as feet not flat plates - if you go this rout you will need a second level to verify that two of the barrings are level to each other or at lest in on the same plane as they where when adjusting for level

[@Michel-Uphoff]

Sorry, I don't see them. I have a sufficient flat surface. The plane is very easy to level (adjust, turn the spirit level 180 degrees, adjust, turn 180 degrees, until no difference is noticeable anymore and the plane is dead level). Of course one has to check his instruments beforehand, I don't see a flaw there. Unfortunately, bearing balls are useless in this case. A scraped surface plate has irregularities which make reliable measurements with point contacts impossible. You just need small very flat surfaces that rest on the peaks of the scraped plate. People can argue about which surface area they should have. I chose 10 mm diameter, so 78.54 mm²

[@The_Juan_Percent]

Ha! Surprise ending.

[@jameshisself9324]

Why the bot voice? I understand that you don't think your English is good enough, but hearing a real person even with a strong accent is so much better than this. On screen text is better than this.

[@Michel-Uphoff]

Sorry, but I don't agree with you. I've listened to too many videos where the spoken English was so bad that I (and I assume many more, especially non-English speakers) could barely understand it. And with such poor pronunciation, the automatic subtitles hardly work. So this is an attempt on my part to provide easy to understand English to everyone who finds my videos interesting. And if that synthetic sound bothers you, turn it of and turn on the subtitles, which work great with "Brian," my digital sidekick.

[@jameshisself9324]

@@Michel-Uphoff In that case the alternative I suggested of on screen text would be far better. I wish I could convey the uneasy feeling that one might feel when they are aware they are listening to bot speech. It really is unnerving.

[@Andreas-Bauer21]

Stephen Hawking rolls over in his grave. Just turn on subtitles and stop nagging.@@jameshisself9324

[@Jan-yc2lr]

@@Michel-UphoffIk ben het wel met hem eens. Zo'n robotstem is erg onprettig om naar te luisteren. Engels met een zwaar accent is helemaal niet erg. Het geeft de video meer menselijkheid en persoonlijke charme. 😉

[@Michel-Uphoff]

Als het, zeker voor niet Engelstaligen, niet goed te verstaan is, werkt de automatische ondertiteling ook slecht. En dan krijg ik daar terecht klachten over. Stoor je je aan een robotstem, dan zet je toch ondertitels aan en geluid uit? Met deze stem heeft de automatische ondertiteling geen enkel probleem (robot luistert naar robot). @@Jan-yc2lr

[@joell439]

👍👍😎👍👍

[@pappaflammyboi5799]

The engineering is a thumbs up, the music, not so much. Just a suggestion of course.