Man, there's a lot to like in this video. The description of the problem, the description of the tool, design considerations, future-proofing. It's one thing to understand what a tool is and how it works, but understanding the _why_ of the implementation brings many layers of depth to understand why many other things are they way they are, and generally how to think better. I realize how late this message is to the party, but on the off chance you are able, could you describe why the flexure slot is cut diagonally? Is it so that the flex is concentrated at one corner of the slot? What if the bottom of the slot were a different shape, like being round? Outstanding video, I got a lot out of this instruction. Thank you!
Thank you mr Lipton, although I might not want to make a repeat o meter, there is a lot of good tips and advice in your video for future builds I might consider. Once again, thank you for taking the time to make the video, and sharing your knowledge! Ian
Very nice Tom! My only suggestion is to drill and ream a hole thru the base to form the flexure thickness and then just band saw into the hole. The sharp edged slot as a flexure point is a stress riser as you well know. Best, Robin
That seems like a likely first cringe assessment, but I wonder how many of these have actually failed (and they have been around for a while). My guess is that the flexure stays far enough within the elastic range for the stress riser to not matter, since the linkage limits it to a very narrow range of motion...
Interesting, thanks Tom. Is there a reason for the slot to be on an angle? I dont see make it any difference from a slot at right angle. I would probably just drill and ream the end of the slot and cut it with the bandsaw or a wheelcutter...
I am very new to precision machining and I had to replay it 3 times to hear "million'ths?" ! I can hold 0.001" on my less stellar Chinese lathe and mill, but after 4 zeros I have nothing to quantify it! :) I'll go back to learning the basics.....Thanks for the videos, Gerry
I was thinking why not use a ball-nose mill to make round pockets, then glue in 3 ceramic balls (from a bearing or just loose one can buy on ebay/ali/etc..,) but I guess there must be some reason for making them flat in this video... Maybe round ball would be more prone to scratching the surface plate due to small contact area?
The balls would drop into all the tiny pits and valleys in the surface plate. One reason granite is used for surface plates instead of steel or other metal is that it chips when something is dropped on it, leaving a pit without any raised material around it, where the steel would deform leaving a ding with raised material around it.
Another thought would be to get a commercial wheatstone bridge strain gage and build the meter to use a strain gage, with an Arduino using a high-res ADC to read the output. You could then calibrate the actual deflection of the strain gage using a precision indicator.
I am not a machinist or metrologist but I have a couple of suggestions. 1. The Part No. 2 of the upper should be even with the end of Part No. 1 to provide a more secure alignment of the two pieces. The thickness may need to be about 1/4 of Part No. 1 to provide the flexure. This will reduce the stresses to the screws holding them together. 2. The flexure of the bottom piece be a spring steel blade attached with screws but the links will still be needed to limit the amount of travel to prevent damage to the flexure. I think your design is very good and makes sense.
Tom, thanks for the great explanation, you are always clear and show the important details. "Make it swoopy", and you have a deft hand at making beautiful objects. Please update us on the etching press, where you are making a precise heavy duty machine with beautiful details.
You've created a new segment...."Tom's two cents". Thanks for the videos, which I had to hunt for due to RU-vid failing to notify of the posting. You and several other subs. I have that I'm not getting notifications for. Thanks RU-vid.....
Great video mr.Lipton,sensational logic on the subject; i had a thought on the fine adjust mechanism instead of a counterbore and a nut keeping the nub in place haow about a thin snapring or even better a Standard E clip keeping the shaft in place(in a reemed hole) this way you do not have to split the mechanism in two parts.Thank you for the lessons you provide!
Why not have a two-piece base with full-width spring steel "bridge" between them as flexure? Recesses in undersides of two base pieces would allow room for fasteners. Concept is used on both surface gauges you showed. Would allow adjustability of flexure strength with different thicknesses of spring steel flexure material, and if ever damaged, flexure could easily be replaced without junking the whole tool. Love the tool design stuff!
Kettletrigger I agree, this is very much in the spirit of avoiding the other saw cut. allows the flexure to be replaced if it falls, too. I suspect the original has the angled cut due to limits in their tooling, perhaps the saw Arbor gets in the way when the work is parallel to the spindle axis.
basically it is a design problem. the simple cut solution provides all that is required, is simpler to construct and last the expected life of the tool - the fatigue failure of the flexure is not likely to be the weak point in the design. dropping it will be a more likely failure. As stated above, the strain is well within the elastic range and not likely to fail, so why complicate the design and manufacture.
Here's an idea for a variation that could let this tool do the job of an autocollimator: put two of the rear feet in line with the front one under the gauge, put the other foot halfway between the two rear ones and offset to the side (same arrangement as used on the bases of autocollimator mirrors), then the tool effectively measures deviations along a straight line. Do some Excel magic on the numbers and global flatness for the whole plate can be mapped out. If the distance from the rear to middle foot and from there to the one under the gauge is 150mm, you'd get about 0.69 arc seconds resolution (with a 0.5um or 20 millionths indicator), which is good enough for mapping a plate. Increase the foot spacing and you can measure down to whatever resolution you want...
AFIK you would have to zero the meter on a perfectly flat master to do what you describe. Even with the three feet in line it cannot determine flatness, only uniformity of radius (flat is infinite radius)
Correct. Another caveat is the distance between mapped points would be pretty big (ex 6", 150mm rather than the 2", 50mm you could expect with a mirror base). Perhaps there is a self proving method of establishing the calibration for zero (ie flat) without a perfect reference, but I haven't brainstormed that yet...
@chris0tube - I know what you refer to and yes, it is a tool that has stood the test of time. My suggestion above though is effectively for a slightly different tool/function. My proposal is for a tool to precisely measure angles of surface deviations, rather than just localized flatness and repeat-ability. It should be seen as a separate sub-project, rather than a suggestion for how Tom and John should build theirs...
There is something to be said for the mental exercise of coming up with a more elegant/precise/economical/aesthetically pleasing solution to a existing design. The majority of my engineering career is doing just that starting with competitors products that already do what they were designed to do.
Regarding the flat calibration surface to set zero: A self proving setup can indeed be built. For a flatness reference to self-prove, you would need two perfectly parallel faces, and a means to adjust the straightness between 3 points. You then measure one side with a repeat-o-meter, flip the reference around and measure the other side too and adjust until you get a zero reading on both sides. Construction would be tricky; think 3 gauge blocks on a rod, with a lockable hinge on one segment of the rod (so angle between the two segments can be adjusted for zero). The gauge blocks are of equal thickness and have parallel faces. The tricky part would be to twist them into the same plane and to hold everything rigidly. Almost easier to lap up a reference using 3 chunks of cast iron...
Love the high precision content lately! This is a really cool project. Glad there are people passing on these concepts that have been around for a long time.
Excellent video Tom. I really enjoyed the off the cuff on the fly back of the envelope engineering of what should be a real cool tool for the obsessed machinist (Micron Retentive?)
I see a problem with the fine adjust, how you want to put it together. The screwed connection will flex very heavily but the inidcater will be mounted to the more ridged upper half. If you look in detail you will see that the upper half of the fine adjust is thinner that the lower part. I would make the hole tool not from three parts (lower part, distant block, upper part) but from 5 parts - lower part with the fex-joint - distant-block to add weight and give hight - thick lower part of the fine adjust - very thin distant plate - thin upper part of the fine ajdust This this construction reduces the amount of complicated machining. All parts are flat without any steps in them. By the way i think the slot is cut in an angle to make sure the flexing joint is an straight and even as possible. ther is only one thinnest point with an angle. with a 90° cut there would be plane and the point of flexing ist not very well defined
Agree. Good way. I might have used an additional piece of spring steel for the bottom flexure since the shop was rough on these types of gages. The spring would be attached from the top w/4 screws on the member & thru bolts on the body. Spring steel is easy to replace - 3/4 hr total, including writing work request. New lower piece w/integrated flexure about 2 days thru the tool room. I use to design & build many different types of flexure gages, some w/4 axis flexures (X,Y,Z,A) at a time. Have a 2 axis flex (X,Y) in my office, 28yrs old, lots of abuse & still going strong.
I have been thinking of making a sort of Repeat o Meter. I have 2 Mahr Millimess Mechanical Comparators, a C E J Mikrokator and a few LVDT units. I was thinking of using a spring steel Flexures (if you get it wrong you can change it - unlike a saw cut in the main part). I enjoyed the video and the comment from Forrest Addy . My contribution is for fine adjustment - I have a precision stand that uses a fine screw with a gently tapering point about 1 1/2 inches long - ¼ inch OD - the screw floats and the taper acts as a fine wedge - think of taper pin but hardened. As I don't have a surface grinder I was going to use a couple of cheap 123 blocks as the main elements, many people would want to test small surface plates too. I was going to use gauge blocks or carbide inserts as feet. Only the single front foot is important for size and shape as the others are assumed to be on a good plane surface. As I understand it the practical reason for a Repeat o Meter is finding local low spots on a surface about the size of the foot on a typical measuring instrument - not tiny chips or large inaccuracies of the surface. If you put a measuring device (particularly one with 3 feet) with one foot on the low spot it will completely ignore the true plain of the rest of the surface and produce errors and confusing measurements. The link to stop the instrument from breaking itself sounds like an afterthought - I agreed that some simple restraint could be designed in just using sloppy fitting screw.
Hey Tom very interesting, sounds good to me. I love your screw holder, I cut a lot of screws in my radio controlled cars to make them ( odd lengths) work. Neat feetsies, the bearings seem the best choice. Enjoyed this video, lots of good information, thanks. God Bless Ya Dave
Hi Tom, I love your videos. I'm working my way up through your playlist down from the oldest one. At the moment I'm about 2 years out, you just showed off your new surface grinder. But my progress through your list is slow, partly because you also inspired me to make my own videos for youtube. Sometimes I sneak your new videos into my queue because you also sparked my interest in the precision stuff, and inspired me to go learn scraping. I'm hoping to learn and make/buy the tools and get the experience I need to overhaul and scrape my small lathe into something a bit better and well aligned. I saw one of these repeat-o-meters being used by the guys who reworked your surface plates. (The way the one guy talked about all of it was just excellent) Then I wondered if one could make one of these, but I guess you'll answer all my questions about it. (one of the reasons why I've resisted responding/asking questions until I was up to speed on your vids until now) I have a new surface plate it's about 16x10" certified at 2,32 um which is DIN 876 grade 00 (max 3 um). I intend to use it to scrape in the cheap Chinese plate which I already had (It's a solid 10 ;) ) I was planning to make tools like height gages with scraped undersides from parts of an old cast iron machine member, would that be a bad idea? The information you give in your meatloaf and other series is invaluable, and has been a real help to me. (I'm not a trained machinist) You never cease to amaze me in how you show and tell stuff about things I'm about to encounter, or answer questions I have in your next 2 year old video. For instance, there was a big tool auction here recently, it was a real honeypot of quality tools. All the stuff you showed me helped me sort out the interesting stuff. One of the things on my table right now is a lot with random carbide inserts, with some plane square ones too. I was contemplating about soldering/brazing them to a scraper-like tool (for something I'm working on with really hard to access areas) but I just knew they would serve a purpose :) Thanks for all the hard work, and keep it up! Greetings, Martijn
The Rahn is normally a 5" measuring radius. If you were to keep the ball-foot at 5 inches and increase the length of the foot to 10 inches, you would double the effective resolution of the indicator, reducing the requirement for a $1k indicator (new price of course). You could also change the kinematics of the gaging foot to add a lever, or multiplier, to increase the movement of the indicator. Let's say you used a tenths indicator and applied a 1:10 multiplier, then the direct reading would be 10 millionths for every 1 tenth of indicator movement.
Wouldn't you just measure the section from a different position so the larger body is over another portion of the plate ? I mean the things are moved around over the surface so whilst I see it being larger might mean it takes more effort to cover everywhere - it would be possible ? If I were NYCNC and looking to sell the bodies - I would want to make them as accessible to as many shops as possible. That being said - maybe the meters are not so rare over in the USA.
We are measuring one point from three points. The three points need to delineate the reference surface while the measuring point measures the deviation. Ideally, the three reference points are moved to different locations for measuring each spot of the plane, so a 3D map can be inferred. Also, some surface plates are small, and need to be checked as well. There are trade offs...
Even with a 10 inch unit, the points are 5 inches apart. However, the 1:2 ratio isn't enough to eliminate the need for an expensive indicator. You need a 10x multiplier to do that. That would require some sort of linkage, which potentially makes it less accurate overall.
I would consider ending the flexure slice in a round hole rather than sharp corner. Less strain concentration, and if someone did have to saw it, it would give a nice even thickness on the flexure. If they were bandsawing, they could just come straight in.
My count was 122 mkay's and okay's, for an overall average of one every 17.7 seconds, not counting "alright" nor grammatically useful okay's ("... I think that would be okay."). And still... an *excellent* video, thoroughly interesting and useful! Go figure. I just take them as "Got it? Ya with me?" :) Yup. Yup, I am! Mkay?
Differential thread adjusters can offer on the order of 25 um per rev adjustment. Such as Stanley uses them on planes for blade depth adjustment as well. Having a 32tpi right thread paired with a 28 tpi right thread gives you effectively a 224 tpi screw or 0.00446" per rev. Combining them with left and right threads you also get multipliers. 32tpi and 28tpi combined to give 14.93tpi efectively.
I've thought through just such a project. Here are my thoughts after viewing your video: You're right. the Rahm Repeat o Meter is a lame design. It works but it could be done simpler and more economically. You will probably wind up using a 20 millionth resolution indicator or an LVTD and gage amp. A 1/4 - 28 adjustor thread will be plain too coarse. 20 millionths on a 28 TPI adjustor knob is 12 arc minutes of rotation per graduation. Unless there is a separate fine adjustment, you most certainly need coarse and fine or a fine adjust and a lift function. The differential thread is a good choice but a PITA to execute. Maybe a 28 TPI working with 32 TPI: 0.0045" per rev. 1.6 degrees of known revolution per gage graduation. Touchy, but do-able. An integral flexure with sharp internal corners is metal fatigue waiting to happen. I suggest a drilled and reamed hole to start then machine to your desired wall followed by the slit. The angled slit is BS. Use a slit vertical to the flexure wall. Better, use a spring steel flexure clamped with a couple retainer plates. The elevation of the flexure above the reference plane is not big deal but close is better The link stop limiting the follower movement doesn't have to be a link. A link in this application is finicky to fit and without a way to adjust for wear. A set screw and a cap screw travel limiter will do the job, be adjustable, and simpler to execute. I suggest the base have three feet and the follower one. Finally: the basic concept. a plane is defined by three points. Add a fourth and you define a sphere. If it was me, I'd call a Repeat-o-Meter a "sphereometer" if the name wasn't taken by another gadget. Anyway a Repeat o Meter detects localized departure from a spherical radius. It may be said a special case of a sphere is one having infinite radius, IOW a flat plane. As you said, you have to reference the Repeat o Meter from a known flat surface, that is, one of infinit radius - then go forth to locate hollows and humps and rolled-off edges in dubious surface plates jubilantly declaring "A-ha!" with every discovery. Properly set, the Saunders-Lipton (or is it Lipton-Saunders?) Improved Repeat o meter will absolutely define the straight and level and the truth of the plane. Too bad we haven't such an instrument to screen politicians in advance of election.
Mike Cordi I haven't taught in a formal setting for some time. I'm getting old and creaky along with grouchy and critical. Endearingly grouchy of course: everybody thinks it indicative of wisdom and fondly consider my fulminations amusing and love me for it. Other people who try for the same vibe fail and are despised. Grouchy is an old fart's preogative.
Forrest Addy, oddly enough I'm grouchy, and cranky. I've read your post for years and was too young to get to your classes when you were doing them at the Precision Museum. I hold you in the highest respect. Thank you for all the information you've shared. It seems some of old PM subjects are getting rehashed here. Soon enough there will be a Kingway build. I'm glad to see them presented
The reason behind an angled cut instead of a vertical or square is possibly because you have a single hinge instead of two, It also has less or no planar twist. It makes a stronger hinge.
looks interesting. I was looking at making one of these at some point. there is an old federal spec for surface plates that showed diagrams for building a repeat gauge that was basically the same thing as a repeat-o-meter only the bottom wasn't one piece it was cut all the way through with spring steel in between the pieces but i had it on a scanned pdf and can't seem to be able to find it again. so i will be looking forward to watching this thing come together.
Why not use a light gauge metal flexture set in a pocket with a block to hold it? This would allow you to be able to replace it. if damaged or adjust the sensitive if need. To a wider gap for more sensitive less of a gap for less. My thinking of doing this is for more options without having to build a new base each time.
seems like it would be a lot more robust to cut all the way through the flex point then inset a spring steel flexure on the bottom. it would eliminate the need for the links i would think as well.
Thanks for addressing some of the basics that are taken for granted like flatness and squareness. Would a v-block design be a better one size fits all for mounting an indicator? No need for bushings/tolerance stacking.
I have to ask, when you were discussing the threaded attachment device for the end of the link, you said, 'We want something like a shoulder bolt.' Why not just use a shoulder bolt? You get all of the features you need, 'out of the box', so to speak, and if you want something prettier than blk ox, get stainless. Oh, yeah, the sample of spring steel look quite like my long time source of spring steel, binding bands from units of lumber. I made 'slim jims', (old style locked door openers for autos), out of banding material for years. It is getting harder to scavenge, though, as a lot of the banding has gone to plastic.
Wouldn't it be a good idea for the bottom spring relief cut to use a radiused edge slitting saw, leaving the cutout stiffer and less of a stress riser to snap from the sharp normal 90 degree corners of a standard slitting saw, and just make the cut slightly deeper to lessen the extra stiffness, as a balance? I love your coverage of base feet, the 0.015" press fit spheres are a nice touch.
Nice Tom. Do you find the balls you use are always through hardened? Some are case hardened, but I haven't used them enough like that to know how common to expect that. The base of the unit is actually 3 feet and the fourth is a contact point. Maybe that helps to explain it to others, maybe not ;) I like your indicator stand. I have a comparator stand I will one day copy, somewhat scaled down, as the traditional indicator stands are generally rather pathetic for surface plate work. Nicely done on yours.
The stiffness of the upper arm, makes ur tighten bolt has a 1 turn depth. If the upper can bend easy it will make the bolt push less than 1 turn depht, cos the bolt push upward to. I think it will have more less depth at the measurement. Great design i think. Cos you may not thread bolt less then 0.7mm with the ordinary taping tool
Hello Thank you very much for your helpful videos. Ihave one qoetion: Iwill build one small ,but the qoetion is the hardness feet on the bottom must be exactly the same diameter? Or it’s not important if one or two of them be bigger ? Thank you Best regards sina von Österreich 🇦🇹
Really nice analysis Tom, well done! I'll look forward to seeing what they look like :) I wonder if the back feet will fit on your 6" lapping plates for that step; I'm also looking forward to seeing those in action :)
I may have missed it. Did you explain how a "repeat-o-meter" checks flatness? It is designed to measure repeatability. The auto-collimator is the instrument to measure flatness.
You need a known flat surface to set the R-O-Meter to an index zero, then use it to check the flatness of other things. I happen to watch Tom's vid about having his surface plates re-calibrated, just before I watched this one, and the tech doing the work explained it to Tom, and us, then.
The fine adjust in the original device looked like some differential screw. That's perhaps the most interesting part of the design, why did you get rid of it? Won't your crude alternative be too coarse for the end task?
Hello Thank you very much because of your helpful videos. Ihave a question? Iwill make one ,do you think if the feet’s on the bottom if be bigger is better or smaller?(diameter) I’m waiting for your response Best regards sina Von Österreich 🇦🇹
Why not use an e-clip and groove on the adjuster thread, instead of a big klunky locknut? Just use a shoulder bolt that has been modified with a groove for the e-clip. The adjuster knob can be press fit to the head of the shoulder bolt.
I thought you'd go more into technical detail on how the flexure works. Since it's just gravity making it flex, how heavy does it have to be to work? How much material must be left in the flexure point? How far does the fourth contact point have to be from the center? (It looks to be about 5.25" in the picture)
Curious....using the lock nut in the pocket in lieu of a double thread, doesn't that create slop in the fine adjustment when adjusting? Tighten to untighten?...that slop doesn't matter in your screw holding tool.....don't think it matters in the meter but could be bothersome in the fine tune?
Chuck, That is the beauty of the lock nut. Unlike E clips and snap rings you can remove ALL the axial backlash and set up almost any amount of knob resistance you desire without shimming and fussing about. Best, Tom
I suspect that the flexure is cut at an angle to make the area that is flexing smaller, maybe. Cut at an angle the thinest point of the flexure has a "V" profile this rather than a flat profile (|_|).
Mike B I thought the same. If Rahn's slot were vertical, the flexure would be a strip of mild steel whose length equaled the slot width, and whose thickness was 1/16 or whatever was left after cutting the slot. Gut feeling suggests that's too stiff. Angling the slot reduces the flexure to essentially zero length and limbers it--as well as stressing it as many have said. I too like the idea of rounding it. While I can't imagine it happening with a vertical slot, the flexure-end of the follower ought not have any vertical freedom. Maybe this was a design influence? It would make me think twice about using a separate spring here. Unless there's enough downforce on the end of the follower to render it moot, what's to keep flaws in the design or execution of the flexure from influencing readings? Seems like any excessive stiffness or upward springiness would destroy the instrument's honesty, dealing in millionths. Does one aim for "plenty flexible" and move on?
I don't think it's quite true that mild steel and spring steel have the same rigidity, is it? Perhaps Tom mean't they're very similar, or was talking about annealed vs hardened, rather than steels of different composition
? what about the flex on upper and lower arms long term to lift the gauge top plate will need to be thinner than lower. just .02¢ more. good job looks like fun. bill
Wasn't there a kit out there that supposedly allowed you to check your plate with an indicator? Either i was dreaming or i remember something akin to a button tip with a weird looking base.
I love your fine adjustement mecanism, do you preload to lock washer with a beville string or something like that to prevent back clash from the tread? It make so much more sens than just having a spring preloading a pivot wich I never really find them repeatable. Thanks for the tips Tom put them in my designer toolbox.
Hi Simon, No need for the spring washer. You can take out all the axial backlash by carefully adjusting the lock nut. Thanks for the comment. Cheers, Tom
How fine of a reading ( .001, .0001, .00001) can you have before the arc of the flexure and the arc of the meter fine adjust affect the reading on a significant level? I guess you might need to account for cosine error like on a tense indicator? I would think using a piece of spring steel between the main body and the forward reading foot (think like a biscuit in a wood joint) would be more effective.
Is Permatex better or worse or same as Loctite? I've been wanting to stock up on a few different ones. What's the typical shelf-life of the "RED"s, "BLUE"s, "GREEN"s and adhesives like the Loctite 380?
They vary a lot, from only a few months to indefinite. You'll need to check with the manufacturer (not the supplier, they'll tell you anything to sell it). I have some Locite products that are twenty years old and still good. You do need to keep them away from temperature extremes (not on the window sill). 380 certainly keeps for >2 years if kept cool and sealed. Beware buying from small sellers who may have very old stock. Cyanoacrylate cures from humidity so it depends how often you open the bottle, and the RH at the time, and expose it to fresh supply of moisture. Mostly only time will tell!
I wonder if there is a standard series of shoulder screws that would work with the strap instead of the floating spacer. That part is going to be relatively tiny thus a pita to make and keep track of. Thinning a strap or a custom hole would be relatively easy. Along the same line is it possible to add something like a 10:1 lever to the indicator? Let 0.0001 = 0.001 at the other end of the lever. Just some idle thoughts of a non machinist.
With that name I immediately imagined someone selling it on a Shopping Channel: "But, wait - there's more! Call now and with your Repeat-O-Meter you will get THREE extra brass bushings for free!". Ok, I guess no one would buy it, but that name just makes me lol:)
Hey Tom, you should pop over to the "Megabots" facebook page and take a look at the cover photo for a second :P To that end, anyone who reads this should too.
I think you drop a couple of zeros, unless you meant 100 millionths. A millionth of an inch is 0.000001". ( I had to look it up, to make sure; never needed the dimension, before.) The reason? If they have a surface plate which is calibrated, the Repeat-O-Meter can be set to zero on the plate, then used to check the flatness of other things, like lathe or mill tables, ways, or any number of blocks, angle plates, etc. It wouldn't be used to measure a millionth, but to check flatness. I'm setting up a lathe for the first time, and the 'straight' edges I used for 40+ years, are suddenly not straight enough for a lathe capable of turning to accuracy in the range of 0.001" to 0.0001".