great video. working as a Surveyor, we have from time to time do work on some optical tooling benches in Labs and such, using level in a similar what to set the bench level to start then to set the equipment mounted on it level. A good level is a very powerful tool in an workshop.
Tom, when using a level to plot flatness, you have to move the level so that the center of the following pads land on the previous center of the leading pads. Your kind of just moving the level around willy nilly there. The moves need to be incremental based on the center of the pads. Your plot would only have 3 dots or 2 flat measurements. I've been manually plotting flatness with bubble and electronic levels for almost 30 years, not sure this really shows you anything. Just my two cents
i agree and would add that there ought to be contact points only at the ends of the level or what would work is to consecutively shim the level at increments keeping the level stationary
But then the contact points can fall into small "holes" and throw of the measurement. Because you want to measure averages over certain intervals, not surface roughness.
he wasn't moving the level willy nilly, he was moving it incrementally. i think what you describe is the minimum number of steps (least overlap), as in the moody method, which is good for ease of calculation and wise when checking large stuff like lathe ways etc. he had a one meter beam, so he went with larger overlap and took more measurements, but they were still regularly spaced. the data set obtained through your method is fully included in tom's data set. that being said, tom mentions at 13:15 that it would be interesting to flip the level for the other end of the beam, but i think that doesn't make sense since the level had already been there. he would just take the last few measurements again while looking at it from the other side. so he did in fact check the whole beam, each number corresponding to the midpoint of the level. which means that his beam is lopsided. anyway, the thing should have sagged in the middle, not humped.
Great video, I really hope our future-machinists see some of your videos. I never appreciated the real-life applications of advanced math classes. I wish I had paid more attention in school.
It occurred to me when you showed the graph that if you invert the values you are remotely measuring the wear of the machine used to prepare that surface. The wear pattern looks consistent with what you might expect in a milling machine with most use near the center of the table.
When I built my plate table, I made the Airy points adjustable in height so I can level the plate with ease. It just seemed like a good idea at the time. Now I know why I did it! ;D
Of course that is a pretty typical milled profile since the cutters tend to take a bit extra off the ends because of the changes in tool pressure based on the amount of cutter engagement, close to .0005" over that length isn't too bad really I think.
Tom, I have a few questions, your videos always do that to me...., that I hope you can explore. I used a similar technique in trying to understand the overall flatness of a way surface once with blocks under the heel and toe of the level as it was not relieved in the middle like yours. One of my thoughts then and questions now is about the reading at any given point. I think it just becomes a measurement of the incline or decline of the level at that point and not really an "elevation" reading relative to the initial point of zeroing. The reason I thought this was because any "low" point along the surface can create both an incline and decline reading as the level is moved along; the toe drops in the depression first (declining) and then two or three segments later the heel drops in (inclining) that point creating the opposite affect. I could never figure out how to take each reading as an angle, independently, showing the change- between the two points on the level, along the surface and come up with a method to accurately reflect the surface profile...Maybe someone can? The other question I had about the straightedge was about aging it. Are you a believer in the process and will you be doing it? I've read about and talked to folks who say that all castings, when new, are under stress and any scrapping, or lapping in your case, would be negatively affected as the stress in the part is relieved over time. Burying castings in the ground is one proposed method of aging, some have even suggested that it gets tossed in a pile of gravel, in the back of a truck and left there for some time (weeks) to be subjected to the vibrations and weather changes to help season it... Thoughts? Thanks for the Video! Take Care.
I am in the measuring business also, as a land surveyor, and I think you are right. I think the answer to your question has to do with exactly where the "feet" of the level are, or rather, exactly where they touch the bar being measured. That is indeterminate in this example, because the feet are broad. Imagine that the level is 10 cm/inches/units (unit doesn't matter to get the idea) in length. I live in SC, so even though I prefer metric I'm gonna go with inches. So, your level is (lets say) 10 inches long, and the feet start at one end, and are two inches long. Then, there is a six inch "gap" in the base, and another two inches of feet. Now imagine you are testing a supposedly flat bar that is actually concave along its length (but ya don't know that at this point). Here we must stop and attend something else. The level wasn't zeroed, so the original starting zero that was set before measuring commenced was an arbitrary angle off level (unless the tester was very very lucky). He should first have marked the ends of the level on the bar, then zeroed the bubble (as he did). Then, he SHOULD have picked the level up, reversed the ends, and noted how far off the bubble was. Lets say the bubble was three marks off when the level was reversed. To make the level (not the bubble, but the level itself!) level, he would then have needed to raise the bar AND the level up on one end so that both the bar and the level (but actually only the level becuz the bar is concave) such that the level read "off" one and a half marks. Then he would turn the level back to it's original orientation and it would still read "off" one and a half marks. At this point the level would be truly level, because the bubble is reading the same way regardless of which way it was placed on the bar. Now, without moving the level, the bubble needs to have one end raised or lowered so that it reads zero (so that it is centered). Once that is done, the bubble will stay centered when set down in either direction. NOW FINALLY we know that the level is truly level, the bubble is zeroed, and the tiny portion of the bar we are checking is also level maybe (except we secretly know it isn't cause it is concave). NOW we can get to my point, which is that, if the level feet are perfectly flat, when the bar is concave, the feet will touch the bar at two points 10 inches apart. But, if the bar is convex, the level will touch the bar at two points 6 inches apart. But we don't know the shape of the bar, so WE DON"T know the distance between the two parts of the level base that touch the bar. So, what do we do? The first thing we need to make an accurate map of the bar top is to isolate the exact point of foot contact. In this example, I would think it would be best to put a precise and accurate guage block of identical size at each end of the level, the smaller the better. If you could get two IDENTICAL tiny bits of square bar stock about 1/8" cross section, a couple of inches long, and put those under each end of the level, you would then finally know that when you measure change in level, you are measuring 10 inches of the bar. Now, because you truly zeroed the level (assuming the level has no sag, and your tiny bars match perfectly) when you measure the first 10 inches of bar, you get a usable (both accurate and precise) angle. In the above video, the angles are precise (repeatable) but not accurate, because the level was not zeroed (checked against itself by reversal). I am pretty sure that the rest of the demonstration and math is valid. I am almost positive that if the level had been properly zeroed, then the plotted line would have wandered above and below the imaginary perfect level line on the chart, rather than being entirely "down". The reason I am pretty sure about all this leveling stuff, even though I know nothing about machining, is that in my profession, leveling is very important, and our survey instruments have to be, well, VERY level. To accomplish this, we level with a bubble, and so before we can do anything else, we have to check that the bubble itself is truly indicating level when centered. The first thing we learn is, point the instrument crosshair at something, and center the level bubble, then, we point in the exact opposite direction (first north, then south for example) and we look at the bubble. Is it still centered? If yes, the bubble indicates level when centered. If no, we fix that by removing exactly half of the "uncenteredness" of the bubble, by raising or lowering one end of the bubble (which our instruments are designed so as to make that simple). If half the uncenterdness has been removed, the bubble will read "off" center the same amount no matter where you point the crosshair. At that point, the instrument is level even though the bubble says it isn't. Then, you raise or lower one end of the bubble tube till it reads centered, and your level bubble will indicate true level for a week, or if lucky, for a month. Knowing all this is also the reason that when I use a carpenters level, I check something for level with it, then I reverse it and check again. If the readings don't match, the level gets replaced, and becomes just a carpenters straight-edge. In other words, this video was very helpful, but the proper pre-calibration of the level device wasn't understood (or just wasn't explained) and the uncertainty regarding exactly where the bottom of the level was touching the bar made it impossible to know how much of the bar you were measuring with each measurement. If the bar was concave (sagging) a 10 inch level would measure 10 inches at a time, if not, there is no way to know, unless you add tiny feet that are a set distance apart. I hope this makes sense to somebody, it took me a while to type.....
@@johnwetzel5248 Wow... Thank you! That was a very well thought out answer and it made sense to me. Before getting to the part of your note about surveying equipment I was thinking that your line of work is really just a macro version of what Tom was trying to demonstrate. These words, level and flat, are used often as an absolute and most times when I hear them used like that, I just smile... Thanks for the response! Take Care.
Hey Tom, As the level moves along, are you not changing the datum? The first measurement sets the heel as the datum, and the toe as the variable. As you move the level, the heel is now sitting on a new datum, which effects the angle. If the heel just moved from a high spot to a low spot, and the toe has not changed height since the last measurement, the angle change reflects the change at the heel, not the toe. It can look like a high spot at the toe, when the heel actually dropped into a low. In order to make a more accurate graph would you also need to factor in the difference in height from one heel position to another, ie: the measurement made 6" back?
+SJWs & Betas Killer I'm sure you're right, but I don't understand it either. Imagine for example that the surface dipped down a thou at the two inch point, then back up at the four inch point, stayed at 0 at the six inch point, then dipped down a thou at the 8 inches. If you moved the level along at 2 inch intervals, wouldn't you measure level at each interval, even though the surface was wobbling all over? BTW thanks a million for all the work you put into these videos, Tom, they're better than any commercial tv I've ever seen.
The angle measured is the rate of change of the surface height, so to get the variation in height you'd need to integrate along the straightedge - so yes, to get the height at any point you would need to take into account the angles measured up to that point
When you have a 6" level, you have to start from 0 to 6. Then from 6 to 12 and so on. By doing this you know exactly where the angle changes from the previous point. I think what we have seen in this video is a wrong way to measure straightness
Interesting as always, but I was surprised when it got to the end because I knew it was supposed to be nearly 15 minutes, yet surely only 5 or 6 had elapsed! Such an interesting topic that I need to recalibrate my perception of the passage of time, apparently.
Can you give more info on the surface plate cleaner you used? My video screen I watched this on had a lot of glare on the bottle. I couldn't make out the brand, etc.
Hi Tom, it would be interesting to mount the level on dowels in order to get more precise measuring points, then you could rely on the resting distance of the level to the surface, but, I understand the point of the demonstration... ;) Cheers, Pierre
Cool. It never occurred to me to measure iteratively at such small regular intervals, like a time series. One quibble, for the benefit of any high schoolers watching: At about the 11:30 mark, you don't take the tangent of the ratio (0.001/10) to find the angle. You take the arctangent or inverse tangent or tan-1. You take the tangent of the angle, of course, to find the ratio of opposite/adjacent, as Mr. Ox does in the very next step. Just ordered a master precision level (0.0002"/10"), which will drive me insane if I use it to level a pool table. Promises to be a fun trip. And a short one.
Tom, cool stuff, always good to gather useless information (Just Kidding). I'm happy you are still on the level, and not a complete deviate as some of our wonderful TH-cam Creators Group have tried to insinuate..... not naming names..... but the secret agent Dale Derry comes to mind every time haha! Looking forward to hanging out with you next month, and having some fun together on the road and with our mob of creators down at the ranch, I can taste those steaks and single malt from here. BTW I bet the test would be different if you took the level and the plate and straight edge to the Bermuda Triangle! Cheers and hugs to Sargamo. B.
Sorry, I couldn't read the bottle of surface-plate cleaner. I think the white bottle is a little too over-exposed in the video, and there is some glare from an overhead light? Could you just post some info like the name and/or brand for the cleaner?
Is there any practical way to scrape a machine without having a professional straight edge of any sort? I have a very old mill the knee moves up n down very easily on the upper part of its range, but the lower is rather hard (minimal wear there I assume) I would like to spend some time with this machine in the future trying to fix it up a lil bit (no intention of making it perfect to the 1/2 thou) its never going to be as nice as a modern Bridgeport, but would just like to make it a lil better. Any thoughts?
I was down my big box hardware (bunnings in Australia) and i didnt have me glasses on, so i picked up one of those laser levels and misread its accuracy. I thought it said such and such over Km but it it actually said such and such over meters. I spent 25 minutes opening the box, reading the instructions and setting it up. So im all good and walk over to the wall to mark my line and the laser was about 3/4 of an inch thick! You may as well have marked which of the boys pissed furtherest up the wall! So i left me glasses at home and took it back for a refund. I figured the kid behind the counter was just a thick as i was, so i slapped it on the desk turned it on and said to the kid go look at how thick the line is buddy! I walked out with a cash refund 💰💰💰 LOL!
But what is the significance of the fact that the level base is bigger than the intervals of measurement? Because the level base will then average over the whole length of the level base against the straightedge, but at the same time it will startle over small "valleys" and rock on high sections.
Not sure why you had to over complicate it by converting to angle 0.001" per 10" is the resolution correct? so if you get 1 division per 6" that is simply 6/10 x 0.001 = 0.0006 " (0.015mm) deviation occurred on 6" on that straight edge correct?
I have this problem or should i say doubt on how to check the straight edge using laser and surface plate. If I lay the straight edge flat on the surface plate and the surface i want to check is perpendicular on the surface plate, should i set the parameter to x or to z?
I don't think Tom ever mentioned what the name of the surface plate cleaner he recommended was, so I figured it out. It's Standridge Premium Granite Cleaner.
How long is a rope? It could be dead flat or you could serve soup on it... Even if it were dead flat, 1 1/8 inch thick is not going to be particularly rigid in a metrology sense. Also, temperature fluctuations would be a significant factor. Managing expectations is your key here. If you can get consistent enough measurements for your purposes, it is what it is.. You can get an 3 inch thick 18x24 plate from MSC for 165 bucks which would be vastly better. Edit: current MSC flyer has a lower grade import the same size for 120... Still way better.
It may or may not be what you think it is. There are 'natural' granite slabs which are sawed and finished as one piece; and there are composites. Composites are epoxies filled with materials like crushed granite, ground or crushed glass (very pretty), and just about everything else you can think of. As you might suspect, composites are far more flexible and susceptible to temperature variations. They are also prone to some degree of permanent shape change - insignificant for countertops, but not reliable if you're measuring to .001 or better. They have many uses, but metrology is not something they're very good for.
I have a 1-1/2" thick granite slab that I use, and I don't have an autocollimator, but I can tell that it's not flat with a strait edge and a flashlight. If I stick an indicator on a 'sled' and move it around with the stem of the indicator a few inches away, there's all kinds of inconsistencies.
Rather then expecting a section of granite counter top to be flat enough to use as a reference, why don't you just buy a decent smaller surface plate? Have you actually looked at prices? You might be surprised to find they are quite affordable unless you buy a gigantic one. Another good point: "DO NOT" buy used surface plates unless.... they are like 5 bucks, as you are 99.9% chance buying a well used and thus JUNK piece of rock and expecting it to actually be flat. Never ceases to amaze me why people buy used surface plates and pay good $$. If you do buy a used surface plate, you might as well go down to your local granite counter top company and have then cut you out one. With the counter top granite you have no clue what your getting and the same exact thing goes for "used" surface plates. If you want something to set a piece of sandpaper on to make your wood project flat..... yes, use a section of granite counter top... or that used surface plate you paid no more then 20.00 for at the machine tool auction. If you are an actual "Machinist" or "wannabe machinist" buy a actual certified granite surface plate. Then, take care of it, and it will be there for you and help you and be far far far better then a section of counter top, unless your looking for something to sand your woodshop hobby project..... ~
What I'd like to know is how to level a C channel rectangular frame so that it's flat. I've seen so many people making welding tables but no one talks about flatness which is as important, if not more, than squareness.
A welding table, you say? Would that be the version of welding where things get hot? Where the table is made of a material with a non-zero coefficient of thermal expansion? And you're heating some parts of the workpiece to red heat? Depending on the degree of flatness required, a lump hammer may be an ideal tool to help. A bit more specification in your query would help to give a useful reply.
Just my 2cents, but wouldn't it make more sense to simply support the straightedge on the surface plate with parallels or 1-2-3 blocks at the Bessel points, then run your indicator over the surface of the straightedge (referencing the surface plate) to get a map of deviation all over? That way you can get a more detailed picture of what the surface looks like. I would question the errors that could creep in due to the length of the level and the wide but limited coverage (there's no way to run the level off the end to get readings beyond the large footprint). Also, you are looking at two points simultaneously for each reading, which could be a numerical nightmare to analyze.
Josh Myer I think he said yesterday that it is graduated in arc seconds. (The small one.) The big one is graduated in 0.0001" height difference per inch of length.
Since those levels tend to be expensive, I wanted to note that you can buy the bubble vials itself for cheap from the "usual suspects", with specs like, say 10" (1/3600 degree), for sub $10. Then I guess you can easily make your own level by attaching the bubble to, I don't know, a parallel, or 123 block, with some screw adjustments to zero it, and "Bob's your uncle" (?) I did not actually make a level like that yet, just ordered a vial - any comments on this idea?
if you place the level on a flat plate with 2 roles, like sine bare roles , that are closer together you can get a very small measurement as close together as you can balance the lot. and as long as you know the distance of the tuch points you can do the math
Just an innocent (hopefully) suggestion but in instead of using of using angles to figure out how much one division of the level equals in 6" just use the ratio i.e., .001"/10" = a"/6" therefore a = .0006 Sorry you actually did that in the explanation. My bad.
If the straight edge is metal, in my opinion you should have lightly stoned the top with a hard arkansas stone to remove any dings and then clean it with your special cleaner.
If all the numbers were negative, shouldn't your graph show a constant slope down to the right? You measured the slope at the points along the length and need to take the integral to get the actual heights.
Not going into what the actual ticks are showing but think of it like this - assume that these ticks are what you got: -1, 0, 0, -1, 1, 0 These are the slopes at 6 points (assume 1" apart). You need to multiply them by the distance between them and add to a running total. Starting with a zero point and using these slopes as a representation of the surface between the points we get these 7 heights: 0 -1 -1 -1 -2 -1 -1
hay now i never said flat. where i live its hillie and holie and thers dam trees so i cant see a thing. and its dark half the time ,and a lama bit me this one time . thats where i stand , where the lama did the biting.
Sir can you give me some information about this particular level which you are using in this vedio. What's exact name of this instrument & How I can read this instrument easily
That was basically my question: does it matter if the surface plate is out of level for the measurements Tom was doing?... If we say the plate is flat but out of level and the straightedge is unknown flatness, then zeroing the bubble anywhere on the straightedge creates a reference point that takes into account the surface plate tilt... no?
I would guess the dial only has a so many increments and that the granite would go off scale meaning he would have to decalibrate his level in order to get measurements
This works only if your straightedge is less than 60 feet long. (Earth's curvature is about .001" in 60ft) Yes, I did the math. Yes, I'm a nerd. What's your point? :P Thanks for the video!
If the surface plate is moving when you put your weight on it, how can you be sure the surface plate is not tilting as you move the weight of the level? I would want to put something as heavy on the other side of the surface plate to quantify the tilt effect on the surface plate, and then again on the straight edge to quantify the deflection of the straight edge.
To be sure the surface plate was not itself tilting, Tom could have used one of his more sensitive levels on the plate to note any changes in the plate due to the weight of the level on the straight edge.
Ok, so you do a great job cleaning the surface plate but then you don’t ensure any other thing you put on it is dust/grime free. For all I know your arm dropped a hair that got caught under the parts and that gave the errors. I tend to run my hand over the plate every time before I put anything on it and I run my hand under anything I’ll put on it.
glad to see the back of your balding head again this episode seems to be building tword something measuring out the shape and collecting data forming a hypothesis and predicting an outcome notso fresh has some interesting post below by the way ive never used a level in such a matter so complicated but I know enough there's something to learn
I think this is a wrong way to measure straightness. When you have a 5 inch level, you have to start from zero to 5 inch. The next point starts from 5 to 10 and so on.
I think your brain is so big, it has pushed all the hair follicles off the surface of your scalp. I am also folically challenged but I dont think it has anything to do with the size of my brain. Being self employed I understand completely how hectic work can get and your missing videos, but it sure is nice to see the red dot for new videos on your subscription. Miss ya.
Thumbs up if you had a near death experience watching this.... Tom . . your a professional, sir!
great video.
working as a Surveyor, we have from time to time do work on some optical tooling benches in Labs and such, using level in a similar what to set the bench level to start then to set the equipment mounted on it level.
A good level is a very powerful tool in an workshop.
It would be nice to rerun the test after you are finished scraping it! Thanks for the demo, you seem to always come with interesting ideas!!
Great Show Mr Wizard! As always thanks for sharing your time and many talents! Best... Chuck
Tom, when using a level to plot flatness, you have to move the level so that the center of the following pads land on the previous center of the leading pads. Your kind of just moving the level around willy nilly there. The moves need to be incremental based on the center of the pads. Your plot would only have 3 dots or 2 flat measurements. I've been manually plotting flatness with bubble and electronic levels for almost 30 years, not sure this really shows you anything. Just my two cents
i agree and would add that there ought to be contact points only at the ends of the level
or what would work is to consecutively shim the level at increments keeping the level stationary
Lone Wolf Garrett 87 Welsh national railways
But then the contact points can fall into small "holes" and throw of the measurement. Because you want to measure averages over certain intervals, not surface roughness.
how big you think the contact points would be?
he wasn't moving the level willy nilly, he was moving it incrementally.
i think what you describe is the minimum number of steps (least overlap), as in the moody method, which is good for ease of calculation and wise when checking large stuff like lathe ways etc. he had a one meter beam, so he went with larger overlap and took more measurements, but they were still regularly spaced. the data set obtained through your method is fully included in tom's data set.
that being said, tom mentions at 13:15 that it would be interesting to flip the level for the other end of the beam, but i think that doesn't make sense since the level had already been there. he would just take the last few measurements again while looking at it from the other side.
so he did in fact check the whole beam, each number corresponding to the midpoint of the level. which means that his beam is lopsided. anyway, the thing should have sagged in the middle, not humped.
Great video, I really hope our future-machinists see some of your videos. I never appreciated the real-life applications of advanced math classes. I wish I had paid more attention in school.
Your a great teacher Tom, thanks for sharing
It occurred to me when you showed the graph that if you invert the values you are remotely measuring the wear of the machine used to prepare that surface. The wear pattern looks consistent with what you might expect in a milling machine with most use near the center of the table.
I learned something today - I can go home early! Thanks Tom. :-)
When I built my plate table, I made the Airy points adjustable in height so I can level the plate with ease. It just seemed like a good idea at the time. Now I know why I did it! ;D
Of course that is a pretty typical milled profile since the cutters tend to take a bit extra off the ends because of the changes in tool pressure based on the amount of cutter engagement, close to .0005" over that length isn't too bad really I think.
Nice level, more accurate than any work I will probably ever produce :)
Tom, I have a few questions, your videos always do that to me...., that I hope you can explore. I used a similar technique in trying to understand the overall flatness of a way surface once with blocks under the heel and toe of the level as it was not relieved in the middle like yours. One of my thoughts then and questions now is about the reading at any given point. I think it just becomes a measurement of the incline or decline of the level at that point and not really an "elevation" reading relative to the initial point of zeroing. The reason I thought this was because any "low" point along the surface can create both an incline and decline reading as the level is moved along; the toe drops in the depression first (declining) and then two or three segments later the heel drops in (inclining) that point creating the opposite affect. I could never figure out how to take each reading as an angle, independently, showing the change- between the two points on the level, along the surface and come up with a method to accurately reflect the surface profile...Maybe someone can? The other question I had about the straightedge was about aging it. Are you a believer in the process and will you be doing it? I've read about and talked to folks who say that all castings, when new, are under stress and any scrapping, or lapping in your case, would be negatively affected as the stress in the part is relieved over time. Burying castings in the ground is one proposed method of aging, some have even suggested that it gets tossed in a pile of gravel, in the back of a truck and left there for some time (weeks) to be subjected to the vibrations and weather changes to help season it... Thoughts? Thanks for the Video! Take Care.
I am in the measuring business also, as a land surveyor, and I think you are right. I think the answer to your question has to do with exactly where the "feet" of the level are, or rather, exactly where they touch the bar being measured. That is indeterminate in this example, because the feet are broad. Imagine that the level is 10 cm/inches/units (unit doesn't matter to get the idea) in length. I live in SC, so even though I prefer metric I'm gonna go with inches. So, your level is (lets say) 10 inches long, and the feet start at one end, and are two inches long. Then, there is a six inch "gap" in the base, and another two inches of feet. Now imagine you are testing a supposedly flat bar that is actually concave along its length (but ya don't know that at this point). Here we must stop and attend something else. The level wasn't zeroed, so the original starting zero that was set before measuring commenced was an arbitrary angle off level (unless the tester was very very lucky). He should first have marked the ends of the level on the bar, then zeroed the bubble (as he did). Then, he SHOULD have picked the level up, reversed the ends, and noted how far off the bubble was. Lets say the bubble was three marks off when the level was reversed. To make the level (not the bubble, but the level itself!) level, he would then have needed to raise the bar AND the level up on one end so that both the bar and the level (but actually only the level becuz the bar is concave) such that the level read "off" one and a half marks. Then he would turn the level back to it's original orientation and it would still read "off" one and a half marks. At this point the level would be truly level, because the bubble is reading the same way regardless of which way it was placed on the bar. Now, without moving the level, the bubble needs to have one end raised or lowered so that it reads zero (so that it is centered). Once that is done, the bubble will stay centered when set down in either direction. NOW FINALLY we know that the level is truly level, the bubble is zeroed, and the tiny portion of the bar we are checking is also level maybe (except we secretly know it isn't cause it is concave). NOW we can get to my point, which is that, if the level feet are perfectly flat, when the bar is concave, the feet will touch the bar at two points 10 inches apart. But, if the bar is convex, the level will touch the bar at two points 6 inches apart. But we don't know the shape of the bar, so WE DON"T know the distance between the two parts of the level base that touch the bar. So, what do we do? The first thing we need to make an accurate map of the bar top is to isolate the exact point of foot contact. In this example, I would think it would be best to put a precise and accurate guage block of identical size at each end of the level, the smaller the better. If you could get two IDENTICAL tiny bits of square bar stock about 1/8" cross section, a couple of inches long, and put those under each end of the level, you would then finally know that when you measure change in level, you are measuring 10 inches of the bar. Now, because you truly zeroed the level (assuming the level has no sag, and your tiny bars match perfectly) when you measure the first 10 inches of bar, you get a usable (both accurate and precise) angle. In the above video, the angles are precise (repeatable) but not accurate, because the level was not zeroed (checked against itself by reversal). I am pretty sure that the rest of the demonstration and math is valid. I am almost positive that if the level had been properly zeroed, then the plotted line would have wandered above and below the imaginary perfect level line on the chart, rather than being entirely "down". The reason I am pretty sure about all this leveling stuff, even though I know nothing about machining, is that in my profession, leveling is very important, and our survey instruments have to be, well, VERY level. To accomplish this, we level with a bubble, and so before we can do anything else, we have to check that the bubble itself is truly indicating level when centered. The first thing we learn is, point the instrument crosshair at something, and center the level bubble, then, we point in the exact opposite direction (first north, then south for example) and we look at the bubble. Is it still centered? If yes, the bubble indicates level when centered. If no, we fix that by removing exactly half of the "uncenteredness" of the bubble, by raising or lowering one end of the bubble (which our instruments are designed so as to make that simple). If half the uncenterdness has been removed, the bubble will read "off" center the same amount no matter where you point the crosshair. At that point, the instrument is level even though the bubble says it isn't. Then, you raise or lower one end of the bubble tube till it reads centered, and your level bubble will indicate true level for a week, or if lucky, for a month. Knowing all this is also the reason that when I use a carpenters level, I check something for level with it, then I reverse it and check again. If the readings don't match, the level gets replaced, and becomes just a carpenters straight-edge. In other words, this video was very helpful, but the proper pre-calibration of the level device wasn't understood (or just wasn't explained) and the uncertainty regarding exactly where the bottom of the level was touching the bar made it impossible to know how much of the bar you were measuring with each measurement. If the bar was concave (sagging) a 10 inch level would measure 10 inches at a time, if not, there is no way to know, unless you add tiny feet that are a set distance apart. I hope this makes sense to somebody, it took me a while to type.....
@@johnwetzel5248 Wow... Thank you! That was a very well thought out answer and it made sense to me. Before getting to the part of your note about surveying equipment I was thinking that your line of work is really just a macro version of what Tom was trying to demonstrate. These words, level and flat, are used often as an absolute and most times when I hear them used like that, I just smile... Thanks for the response! Take Care.
I would be very curious to see if this pattern shows when bluing is applied from the surface plate.
Hey Tom, As the level moves along, are you not changing the datum? The first measurement sets the heel as the datum, and the toe as the variable. As you move the level, the heel is now sitting on a new datum, which effects the angle. If the heel just moved from a high spot to a low spot, and the toe has not changed height since the last measurement, the angle change reflects the change at the heel, not the toe. It can look like a high spot at the toe, when the heel actually dropped into a low. In order to make a more accurate graph would you also need to factor in the difference in height from one heel position to another, ie: the measurement made 6" back?
This was simple demonstration of relative measurement. "Datum" doesn't matter.
+SJWs & Betas Killer I'm sure you're right, but I don't understand it either. Imagine for example that the surface dipped down a thou at the two inch point, then back up at the four inch point, stayed at 0 at the six inch point, then dipped down a thou at the 8 inches.
If you moved the level along at 2 inch intervals, wouldn't you measure level at each interval, even though the surface was wobbling all over?
BTW thanks a million for all the work you put into these videos, Tom, they're better than any commercial tv I've ever seen.
The angle measured is the rate of change of the surface height, so to get the variation in height you'd need to integrate along the straightedge - so yes, to get the height at any point you would need to take into account the angles measured up to that point
When you have a 6" level, you have to start from 0 to 6. Then from 6 to 12 and so on. By doing this you know exactly where the angle changes from the previous point. I think what we have seen in this video is a wrong way to measure straightness
Very Nice Tom. Very educational. Thanks for the video.
Question: Could you have rotated the parallel on the surface plate till you found the level plane?
Nice demonstration Tom!
As a framing carpenter I check all my 2x4s to .001/10"
As we all should.
Nice video. Wondering how the weight of the level changes things as you move it away from the supports on the straight edge.
good observation.
Interesting as always, but I was surprised when it got to the end because I knew it was supposed to be nearly 15 minutes, yet surely only 5 or 6 had elapsed! Such an interesting topic that I need to recalibrate my perception of the passage of time, apparently.
Can you give more info on the surface plate cleaner you used? My video screen I watched this on had a lot of glare on the bottle. I couldn't make out the brand, etc.
Looks like Standridge Premium Granite Cleaner.
Thanks!
I remember Tom's video where he measured the thickness of a Sharpie mark, and then ponder what leaving a lanolin film on the surface plate does...
I was thinking the same thing...'why would you smear lanoline on a surface you need to be flat?'
Hi Tom, it would be interesting to mount the level on dowels in order to get more precise measuring points, then you could rely on the resting distance of the level to the surface, but, I understand the point of the demonstration... ;)
Cheers, Pierre
Why not just use a sine bar?
Do you have a next step video on how to straighten the straight edge now that you know the peaks and valleys on the surface?
what was that cleaner? the exposure was blown out and i couldnt see thr bottle.
www.standridgegranite.com/products/cases-covers-cleaner
Please share the name of the best surface plate cleaner!
Mr Lipton you are a smart Cat!
Cool. It never occurred to me to measure iteratively at such small regular intervals, like a time series. One quibble, for the benefit of any high schoolers watching: At about the 11:30 mark, you don't take the tangent of the ratio (0.001/10) to find the angle. You take the arctangent or inverse tangent or tan-1. You take the tangent of the angle, of course, to find the ratio of opposite/adjacent, as Mr. Ox does in the very next step. Just ordered a master precision level (0.0002"/10"), which will drive me insane if I use it to level a pool table. Promises to be a fun trip. And a short one.
Do you have a link for that cleaner that you used? Thanks
The brand is the same as the surface plate: Standridge Premium Granite Cleaner. www.standridgegranite.com/products/cases-covers-cleaner
Tom, cool stuff, always good to gather useless information (Just Kidding). I'm happy you are still on the level, and not a complete deviate as some of our wonderful TH-cam Creators Group have tried to insinuate..... not naming names..... but the secret agent Dale Derry comes to mind every time haha!
Looking forward to hanging out with you next month, and having some fun together on the road and with our mob of creators down at the ranch, I can taste those steaks and single malt from here.
BTW I bet the test would be different if you took the level and the plate and straight edge to the Bermuda Triangle!
Cheers and hugs to Sargamo. B.
Thanks Tom, quite interesting, didn't know a level could do all that. God Bless Ya Dave
Hi. Would it be possible to measure the curvature of the Earth with these instruments?
what is a good recommendation for a Precision level?
Sorry, I couldn't read the bottle of surface-plate cleaner. I think the white bottle is a little too over-exposed in the video, and there is some glare from an overhead light?
Could you just post some info like the name and/or brand for the cleaner?
Standridge
Excellent info and very interesting. Thanks!
Nice demo, enjoyed.
How much does the weight of the level deflect the straightness of the straight edge???
Is there any practical way to scrape a machine without having a professional straight edge of any sort? I have a very old mill the knee moves up n down very easily on the upper part of its range, but the lower is rather hard (minimal wear there I assume) I would like to spend some time with this machine in the future trying to fix it up a lil bit (no intention of making it perfect to the 1/2 thou) its never going to be as nice as a modern Bridgeport, but would just like to make it a lil better. Any thoughts?
I was down my big box hardware (bunnings in Australia) and i didnt have me glasses on, so i picked up one of those laser levels and misread its accuracy. I thought it said such and such over Km but it it actually said such and such over meters.
I spent 25 minutes opening the box, reading the instructions and setting it up. So im all good and walk over to the wall to mark my line and the laser was about 3/4 of an inch thick!
You may as well have marked which of the boys pissed furtherest up the wall!
So i left me glasses at home and took it back for a refund. I figured the kid behind the counter was just a thick as i was, so i slapped it on the desk turned it on and said to the kid go look at how thick the line is buddy! I walked out with a cash refund 💰💰💰 LOL!
But what is the significance of the fact that the level base is bigger than the intervals of measurement? Because the level base will then average over the whole length of the level base against the straightedge, but at the same time it will startle over small "valleys" and rock on high sections.
Level headed. Interesting Tom.
Not sure why you had to over complicate it by converting to angle
0.001" per 10" is the resolution correct?
so if you get 1 division per 6" that is simply 6/10 x 0.001 = 0.0006 " (0.015mm) deviation occurred on 6" on that straight edge correct?
I have this problem or should i say doubt on how to check the straight edge using laser and surface plate. If I lay the straight edge flat on the surface plate and the surface i want to check is perpendicular on the surface plate, should i set the parameter to x or to z?
How important is that amount of accuracy to the home hobbyist or machinist?
This is super interesting. Is there a way to do this without a level that has an adjustment dial like yours?
Just level the plate(?)
Ah gotcha - the adjustment isn't used to measure, just for initial setup.
But, when you put the gauge down, it looks like it's saying that the granite plate isn't level either? (At 2 different points in video.)
I don't think Tom ever mentioned what the name of the surface plate cleaner he recommended was, so I figured it out. It's Standridge Premium Granite Cleaner.
Always a bit of interesting information, thanks.
I have a piece of granite cutout from my sink counter top. I use it as a surface plate. Very flat. Wondering how strait it is. Any thoughts?
It is only 1 1/8" thick so I have made a sand box to support it.
How long is a rope? It could be dead flat or you could serve soup on it... Even if it were dead flat, 1 1/8 inch thick is not going to be particularly rigid in a metrology sense. Also, temperature fluctuations would be a significant factor. Managing expectations is your key here. If you can get consistent enough measurements for your purposes, it is what it is..
You can get an 3 inch thick 18x24 plate from MSC for 165 bucks which would be vastly better.
Edit: current MSC flyer has a lower grade import the same size for 120... Still way better.
It may or may not be what you think it is. There are 'natural' granite slabs which are sawed and finished as one piece; and there are composites. Composites are epoxies filled with materials like crushed granite, ground or crushed glass (very pretty), and just about everything else you can think of.
As you might suspect, composites are far more flexible and susceptible to temperature variations. They are also prone to some degree of permanent shape change - insignificant for countertops, but not reliable if you're measuring to .001 or better. They have many uses, but metrology is not something they're very good for.
I have a 1-1/2" thick granite slab that I use, and I don't have an autocollimator, but I can tell that it's not flat with a strait edge and a flashlight. If I stick an indicator on a 'sled' and move it around with the stem of the indicator a few inches away, there's all kinds of inconsistencies.
Rather then expecting a section of granite counter top to be flat enough to use as a reference, why don't you just buy a decent smaller surface plate? Have you actually looked at prices? You might be surprised to find they are quite affordable unless you buy a gigantic one.
Another good point: "DO NOT" buy used surface plates unless.... they are like 5 bucks, as you are 99.9% chance buying a well used and thus JUNK piece of rock and expecting it to actually be flat.
Never ceases to amaze me why people buy used surface plates and pay good $$. If you do buy a used surface plate, you might as well go down to your local granite counter top company and have then cut you out one. With the counter top granite you have no clue what your getting and the same exact thing goes for "used" surface plates.
If you want something to set a piece of sandpaper on to make your wood project flat..... yes, use a section of granite counter top... or that used surface plate you paid no more then 20.00 for at the machine tool auction.
If you are an actual "Machinist" or "wannabe machinist" buy a actual certified granite surface plate. Then, take care of it, and it will be there for you and help you and be far far far better then a section of counter top, unless your looking for something to sand your woodshop hobby project.....
~
What I'd like to know is how to level a C channel rectangular frame so that it's flat. I've seen so many people making welding tables but no one talks about flatness which is as important, if not more, than squareness.
A welding table, you say? Would that be the version of welding where things get hot? Where the table is made of a material with a non-zero coefficient of thermal expansion? And you're heating some parts of the workpiece to red heat? Depending on the degree of flatness required, a lump hammer may be an ideal tool to help. A bit more specification in your query would help to give a useful reply.
Two string lines across the diagonals. Space one up by the thickness of the string.
Cool, thanks for sharing.
i truly miss the machining videos
Just my 2cents, but wouldn't it make more sense to simply support the straightedge on the surface plate with parallels or 1-2-3 blocks at the Bessel points, then run your indicator over the surface of the straightedge (referencing the surface plate) to get a map of deviation all over? That way you can get a more detailed picture of what the surface looks like.
I would question the errors that could creep in due to the length of the level and the wide but limited coverage (there's no way to run the level off the end to get readings beyond the large footprint). Also, you are looking at two points simultaneously for each reading, which could be a numerical nightmare to analyze.
It really doesn't matter down at this scale, but: is that level graduated in sine or tangent ticks? (A pre-caffeine question if ever there was one.)
Josh Myer I think he said yesterday that it is graduated in arc seconds. (The small one.) The big one is graduated in 0.0001" height difference per inch of length.
Good stuff Tom.
Totally disagree, the checks aren’t relative because you’re increments are shorter than the level length hence your measuring from an unknown spot?
This was excellent !! Do you think that you could have measured the difference in the surface if you had supported it by the ends??
May be so, maybe not.
By my calculations, over 6 inches, the direction of gravity will change by 4.934 milliarcseconds
Since those levels tend to be expensive, I wanted to note that you can buy the bubble vials itself for cheap from the "usual suspects", with specs like, say 10" (1/3600 degree), for sub $10. Then I guess you can easily make your own level by attaching the bubble to, I don't know, a parallel, or 123 block, with some screw adjustments to zero it, and "Bob's your uncle" (?) I did not actually make a level like that yet, just ordered a vial - any comments on this idea?
if you place the level on a flat plate with 2 roles, like sine bare roles , that are closer together you can get a very small measurement as close together as you can balance the lot. and as long as you know the distance of the tuch points you can do the math
This is the best test to show the earth is flat.. Well done
Very informative to loved the video i learned some thing thank you
Tom.. How do you know what part of the pad on the level is actually contacting the surface ??? Wouldn't this change the angles in your calculations?
In case you didn't noticed, this was simple demonstration, not going into details. So your question does not apply.
Nice! Now I just need to get a precision level ;)
Thanks for sharing, neat to know.
Awesome video
I think he should have blued it up, and done a rub on the surface plate to see if
the results match up..
What results? You don't get any numbers from what you are suggesting, just a sense of where the high spots are. So there is nothing to compare.
interesting. you know you have made it exteremly difficulty to nab a good engineers level on ebay or the like. :)
And to think all these years, I've just been using levels to hang bookshelves.
Welcome to 0.00000000xtool.com.
I recommend using screws.
😃
Just an innocent (hopefully) suggestion but in instead of using of using angles to figure out how much one division of the level equals in 6" just use the ratio i.e., .001"/10" = a"/6" therefore a = .0006
Sorry you actually did that in the explanation. My bad.
Lol just figure it out after doing the inverse sine and cosine of my 0.02/m metric level , waist of time really.
Always adjust for maximum "fancy pants" !....lol..... Good Job !
If the straight edge is metal, in my opinion you should have lightly stoned the top with a hard arkansas stone to remove any dings and then clean it with your special cleaner.
If all the numbers were negative, shouldn't your graph show a constant slope down to the right? You measured the slope at the points along the length and need to take the integral to get the actual heights.
Not going into what the actual ticks are showing but think of it like this - assume that these ticks are what you got:
-1, 0, 0, -1, 1, 0
These are the slopes at 6 points (assume 1" apart). You need to multiply them by the distance between them and add to a running total. Starting with a zero point and using these slopes as a representation of the surface between the points we get these 7 heights:
0 -1 -1 -1 -2 -1 -1
So ... the level always points to the center of the earth. Straightness would be "limited"
to a circle with a radius of 4000 miles. ;-)
if the earth were round , its not .
Noway Jerk I hope you don't believe that
It's better approximated as an oblate spheroid. Also, the pull of gravity will depend on the density of the surrounding crustal rock, etc.
Paul Langenkamp I'll except that, ,thought you were a flat earther.
hay now i never said flat. where i live its hillie and holie and thers dam trees so i cant see a thing. and its dark half the time ,and a lama bit me this one time . thats where i stand , where the lama did the biting.
Can u fix it please. Make it flat and level!,
Autocollimator analysis seems inevitable now.
I am for it - +1 vote here:)
0.1mm/mrt. that is carpenter style!!
Всё доступно рассказал, спасибо!
Sir can you give me some information about this particular level which you are using in this vedio.
What's exact name of this instrument &
How I can read this instrument easily
Interesting. This is cheaper rather than using the collimeter.
Nice but I don't get why it was necessary to use adjustable parallels??
as he said, to level out the straightedge
ajtrvll Because the surface plate is out about 0.400" over the length. he just wanted to get it close.
That was basically my question: does it matter if the surface plate is out of level for the measurements Tom was doing?... If we say the plate is flat but out of level and the straightedge is unknown flatness, then zeroing the bubble anywhere on the straightedge creates a reference point that takes into account the surface plate tilt... no?
I would guess the dial only has a so many increments and that the granite would go off scale meaning he would have to decalibrate his level in order to get measurements
Great catch madinatore!... LOL that's the difference between a theoretical mind and a practical one.
Tom nice one
so with a good level one can check a surface plate aswell. probably not to the accuracy of optic but for the basics.
That level is gorgeous!
Yeah. I'm a nerd. sorry man.
You sound like you're having fun
Arc seconds sounds like you doping a scope.
This works only if your straightedge is less than 60 feet long. (Earth's curvature is about .001" in 60ft)
Yes, I did the math. Yes, I'm a nerd. What's your point? :P
Thanks for the video!
If the surface plate is moving when you put your weight on it, how can you be sure the surface plate is not tilting as you move the weight of the level? I would want to put something as heavy on the other side of the surface plate to quantify the tilt effect on the surface plate, and then again on the straight edge to quantify the deflection of the straight edge.
This was simple demonstration, so chill out, smart-ass.
To be sure the surface plate was not itself tilting, Tom could have used one of his more sensitive levels on the plate to note any changes in the plate due to the weight of the level on the straight edge.
No offense or disrespect was intended.
Windex brand surface plate cleaner
Ok, so you do a great job cleaning the surface plate but then you don’t ensure any other thing you put on it is dust/grime free. For all I know your arm dropped a hair that got caught under the parts and that gave the errors. I tend to run my hand over the plate every time before I put anything on it and I run my hand under anything I’ll put on it.
glad to see the back of your balding head again this episode seems to be building tword something measuring out the shape and collecting data forming a hypothesis and predicting an outcome notso fresh has some interesting post below by the way ive never used a level in such a matter so complicated but I know enough there's something to learn
like the blue origins sticker :)
I hope Mrs.Ox doesn't find out about you & that plate
Too bad Tom is not doing much Y-T content these days.
I think this is a wrong way to measure straightness. When you have a 5 inch level, you have to start from zero to 5 inch. The next point starts from 5 to 10 and so on.
I think your brain is so big, it has pushed all the hair follicles off the surface of your scalp. I am also folically challenged but I dont think it has anything to do with the size of my brain. Being self employed I understand completely how hectic work can get and your missing videos, but it sure is nice to see the red dot for new videos on your subscription. Miss ya.
Randy Richard utube
Randy Richard utube