What's new
What's new

Foundations of Mechanical Accuracy by Wayne R. Moore

Dennis some suggestions you can take or leave.
You should look for specifications on your electronic gauges, they will often have a spec on change with temperature, there can be a big difference with different designs and manufactures. Some of the newer electronic stuff is surprisingly stable. I think your idea of using a gauge block to verify is good. You can check if leaving the electronics turned on or turning off each time is better.

As far as measuring creep over time I would suggesting adding more data points, record the deflection at various points in time like in sort of an exponential time scale, like 5 sec, 20 sec, 1 min, 5 min, 20 min, 2 hrs, 10 hrs, 1 day, 5 days, 2 weeks, you get the idea.

Another thing to consider is to record the deflection as the force is increased in several steps then decreased in several steps. Repeat the cycle a couple of times. If the deflection is not linear with force, or different between increasing and decreasing or changes test to test it suggests some measurement or setup errors that will be interesting to get to the bottom of. Sometimes there will be a relatively large deflection at a small force then become more linear. In any case having an idea at what force it starts to deform plastically relative to the creep load you are applying.

If you really want to make a science project it would be interesting to flip the bar over and see how the creep or elastic deflection changes as a result of previous loads in the opposite direction. At least food for thought as you design the setup. It has always seemed to be that plastically loading it in one direction results in a different effect when subsequently loading it in the opposite direction.

A long pointed rod, loosely guided to be vertical above the center might be a straightforward way to add dead weights, of coarse what you really want is is a load cell with all force, deflection and temperature data read into a data acquisition system LOL. I expect much of this is way beyond the scope of what you want to do but you did ask for comments...
 
Dennis some suggestions you can take or leave.
You should look for specifications on your electronic gauges, they will often have a spec on change with temperature, there can be a big difference with different designs and manufactures. Some of the newer electronic stuff is surprisingly stable. I think your idea of using a gauge block to verify is good. You can check if leaving the electronics turned on or turning off each time is better.

As far as measuring creep over time I would suggesting adding more data points, record the deflection at various points in time like in sort of an exponential time scale, like 5 sec, 20 sec, 1 min, 5 min, 20 min, 2 hrs, 10 hrs, 1 day, 5 days, 2 weeks, you get the idea.

Another thing to consider is to record the deflection as the force is increased in several steps then decreased in several steps. Repeat the cycle a couple of times. If the deflection is not linear with force, or different between increasing and decreasing or changes test to test it suggests some measurement or setup errors that will be interesting to get to the bottom of. Sometimes there will be a relatively large deflection at a small force then become more linear. In any case having an idea at what force it starts to deform plastically relative to the creep load you are applying.

If you really want to make a science project it would be interesting to flip the bar over and see how the creep or elastic deflection changes as a result of previous loads in the opposite direction. At least food for thought as you design the setup. It has always seemed to be that plastically loading it in one direction results in a different effect when subsequently loading it in the opposite direction.

A long pointed rod, loosely guided to be vertical above the center might be a straightforward way to add dead weights, of coarse what you really want is is a load cell with all force, deflection and temperature data read into a data acquisition system LOL. I expect much of this is way beyond the scope of what you want to do but you did ask for comments...
Gard,

Some great ideas. And you need not apologize for making suggestions!

"You should look for specifications on your electronic gauges, they will often have a spec on change with temperature, there can be a big difference with different designs and manufactures."
The Mitutoyo gage I am using is quite old and the documentation for it has been lost. Mitutoyo's website does not have user manuals in its support section for this gage and the gage itself has only a serial number with no other placarded information.

"As far as measuring creep over time I would suggesting adding more data points,"
I do plan on multiple time intervals and multiple repeat measurements to look for creep* and if I do not see it at the initial weights I plan to use, I'll increase both time and weight until it is either seen or not seen. I am looking at the question of whether under ordinary conditions likely to be seen in a shop storing a casting under less than ideal circumstances, can creep be introduced into the part. The question is not whether I can exceed the elastic limit of the part to bend it. We already know that answer---yes.

"Another thing to consider is to record the deflection as the force is increased in several steps then decreased in several steps. Repeat the cycle a couple of times. If the deflection is not linear with force"
I think what you are saying is to make sure the setup seems to follow expected deflections and is not "squirrelly." And, indeed, this is important as I have just fooled with the setup a few minutes before leaving town and by applying finger pressure here and there to make sure there was no rocking and found one support point to be suspect. So, I will look at that closely and may just go straight to the poured metal supports outlined in post 65 above. If I go that route, there are some other potential sources of error that will have to be watched.

"If you really want to make a science project"
I really do not mean to do any in-depth evaluation. My main focus is to try to induce creep under ordinary circumstances and determine whether it occurs or not.

"A long pointed rod, loosely guided to be vertical above the center might be a straightforward way to add dead weights,"
I really like that idea. It is simple and does just what I need to provide vertical (close enough for this experiment anyway) force that can be easily applied and measured. I will fire up the welder and make a guide that will isolate the weight of the guide from the plate so as not to deflect the plate.

Comment: It has been immensely helpful to float the idea for this test to the members here. Most of the comments have been well-intended and have helped me see things I might not otherwise have seen until time had been wasted.

*Carbidebob provided a lot of new-to-me information on sources of measurement error and how to, first, just accept them and then figure out ways to compensate for them.

Denis
 
Last edited:
Here is an updated version of what I anticipate the setup to be not using an SE and using Gard's weighted spike load applicaton.

1675540845339.png


Denis
 
The more classical way to do a 3 point bending experiment like this is to support the ends on knife edges that are perpendicular to the long axis. In practice they are not really sharp but rounded enough to not cause local deflection of the beam, so small diameter dowel rods might be an easy way.

If there is a large washer welded on the vertical rod, I am sure you will find all kinds of things lying around that have a hole in the center, pulleys, lathe chucks etc. Weigh them and the vertical rod on the kitchen or bathroom scale.

Measuring and plotting force vs deflection at some lower forces is one way to check for sources of error, you may find things act squirrelly up to a certain force then everything is nice and linear. This may help more quickly deciding what to use as a minimum force for the creep measurement. Assuming the squirrels are hard to eliminate.

Your deflection measuring device is under the load so you might want to protect it somehow from accidental overload by putting it on something that can move if overloaded or putting something solid next to it.

Point well taken about looking for creep at shop like storage conditions. If I remember correctly sometimes the creep behavior of materials is researched at a load that is some percentage of yield stress. It was years ago for me but what I remember was first trying to find a force where creep was observable on reasonable time scale like minutes or hours, then going to a slightly lower force test for weeks or months. This can then be extrapolated out to force over the lifetime of the part. Some materials will have small amounts of creep at small forces and others will have a force below which no creep occurs.

One thing I would be interested in is after you finished your creep testing you intentionally plastically deformed the part thus inducing an internal stress, flip it over and repeat the creep test, I suspect you might see creep at very low forces on a part that is not stress relieved. Of course you have the ability to stress relieve the part so this might be a cool way to measure the effect of stress relief.

I have seen a related effect like this when attempting to straighten shafts, it takes a relative high force to bend the shaft the first time but if you go a little too far and need to flip it over, the opposite force is much less. I suspect if I stress relieved it before each step the force required would be the same but I had no way to do it..
 
Years ago Midwestern Machinery in Minneapolis hired me to scrape a "Dummy" which is a surface plate with a V and Flat cast into it, so when they scraped Fellows 12 H Gear Tester. It would go faster then using a camelback straightedge and a level gage like a King-Way to make the V and Flat co-planer. The Foreman at Midwestern designed the pattern and they had it cast and machined. They wanted it scraper to .00005" / 12" . It sat on 3 points cast in at approx 30% from the ends. It was approx. 30" long an 14" wide and 4" thick with cast in ribs. I told them I did not think it was thick enough as a normal Cast Iron surface plate that size was 6 to 8" thick and more deeper ribs. It was sitting in the inspection area and the foreman assured me it would not bend.

I was getting paid by the hour, so I scraped it like they wanted it. After I finished they had me use it to scrape the bed of the Fellows 12H Tester. When your using a Dummy there is no way to hinge or test the pivot of the ways like you can with a camel back as the V traps the Dummy and all you can do is rub it back and forth. I scraped the bed and when I was done I set a camel back on the bed and it was high in the middle. What was happening the plate would sag on the ends convex and the bed got high in the middle. Then I rotated the dummy with a camelback , one time with the dummy, next time with the camel back.

PS: The machine was setting on 3 points cast in at the factory

My point is the structure of the part can change when flipping it over if the design is not rigid.
.1675607878140.png .
 
The more classical way to do a 3 point bending experiment like this is to support the ends on knife edges that are perpendicular to the long axis. In practice they are not really sharp but rounded enough to not cause local deflection of the beam, so small diameter dowel rods might be an easy way.

If there is a large washer welded on the vertical rod, I am sure you will find all kinds of things lying around that have a hole in the center, pulleys, lathe chucks etc. Weigh them and the vertical rod on the kitchen or bathroom scale.

Measuring and plotting force vs deflection at some lower forces is one way to check for sources of error, you may find things act squirrelly up to a certain force then everything is nice and linear. This may help more quickly deciding what to use as a minimum force for the creep measurement. Assuming the squirrels are hard to eliminate.

Your deflection measuring device is under the load so you might want to protect it somehow from accidental overload by putting it on something that can move if overloaded or putting something solid next to it.

Point well taken about looking for creep at shop like storage conditions. If I remember correctly sometimes the creep behavior of materials is researched at a load that is some percentage of yield stress. It was years ago for me but what I remember was first trying to find a force where creep was observable on reasonable time scale like minutes or hours, then going to a slightly lower force test for weeks or months. This can then be extrapolated out to force over the lifetime of the part. Some materials will have small amounts of creep at small forces and others will have a force below which no creep occurs.

One thing I would be interested in is after you finished your creep testing you intentionally plastically deformed the part thus inducing an internal stress, flip it over and repeat the creep test, I suspect you might see creep at very low forces on a part that is not stress relieved. Of course you have the ability to stress relieve the part so this might be a cool way to measure the effect of stress relief.

I have seen a related effect like this when attempting to straighten shafts, it takes a relative high force to bend the shaft the first time but if you go a little too far and need to flip it over, the opposite force is much less. I suspect if I stress relieved it before each step the force required would be the same but I had no way to do it..
"The more classical way to do a 3 point bending experiment like this is to support the ends on knife edges that are perpendicular to the long axis."
"you may find things act squirrelly up to a certain force then everything is nice and linear."

I had considered knife edges, but was concerned about introducing error due to the rounded knife edges causing some local deformation (digging in). Knife edges do offer some advantage over columns as the part does not have to rock over a flat top surface as it would with a column.
Squirreliness due to the bearing points not being well seated because of light initial loading might be overcome if I place moderate loads balanced over the supports on each end of the test piece. That would cause the bearing point to "seat" well but would not induce flexion.
As my first attempt to get good bearing seating I am going to spend a little time lapping the flat surfaces of the pucks shown in post 37 above. They were made pretty casually and a bit of lapping of both the flat surfaces and of the center-drilled bearing surface is likely to help them behave better. Since the test piece is reatively long and narrow, it almost balances side-to-side on the one point on the single-point bearing and tiny forces could cause rocking. So any surface spicules could make them not seat well. And if that does not work, I'll try loading the bearing points. Finally, I'll switch to knife edges if needed.

"Point well taken about looking for creep at shop like storage conditions. If I remember correctly sometimes the creep behavior of materials is researched at a load that is some percentage of yield stress."
Yes, the grey iron creep studies I have seen discussed in the ASM Handbook were done at very high loads and at high temperature (700F). That is because, according to their reports grey iron does not creep appreciably at low load and room temperature. So, with conflicting reports from them vs those made here, it seems like a test is in order. (Some metals are very prone to creep---certain well-known motorcycles were known as "Butterheads" as changes in their motor aluminum alloys resulted in unexpectedly severe deformation under only moderate loads and temperature conditions.

"One thing I would be interested in is after you finished your creep testing you intentionally plastically deformed the part thus inducing an internal stress, flip it over and repeat the creep test,"
We'll see how much time and patience I have left after doing this experiment. Using the same setup, I do intend to look at deformation of straight edges due to both hand-contact heat and gravity. We all know that they deflect, but how much?

"Your deflection measuring device is under the load so you might want to protect it somehow from accidental overload by putting it on something that can move if overloaded or putting something solid next to it." That is a good point. As it turns out that probe holder shown in post 41 is pretty robust and the probe that will be in it (not the indicator shown in that photo) would be pushed down inside by a heavy force. Still putting some "safety blocks" on it to protect the probe more would be a good idea. [I may, in another thread, recount how I smashed the irreplaceable power supply of an expensive loaned-to-me mini-checker. It happened as I thought I was being extra careful and putting it "up for safe storage." I would not want to muddy up this thread with unrelated anecdotes.]

Thanks for your help on this. I hope others with relevant information will chime in.

Denis
 
Does anyone else remember a PM member who posted photos of a straight edge that he scraped flat, them torqued it for 2 weeks with one end clamped in a kurt vise and the other end with a 2 foot long arm and a weight? After 2 weeks it blued up on only 2 corners. He said he hammered on it and it went back flat.

Im thinking this was 2018 or 19 era.
 
Yes, the grey iron creep studies I have seen discussed in the ASM Handbook were done at very high loads and at high temperature (700F). That is because, according to their reports grey iron does not creep appreciably at low load and room temperature. So, with conflicting reports from them vs those made here, it seems like a test is in order.
Interesting point on the ASME study and thanks for providing info on how to find them. At one time I had access to a couple hundred pounds of those books. I suspect those studies used carefully prepared stress relieved samples or test coupons. I do believe in real life castings can move because a lot of reputable people have observed it. The interesting question in my mind is why. It may be something other than creep at least as defined by those studies. Perhaps the related phenomena of stress relaxation?

If you bend a bar until it just barely hits the elastic limit it will remain fairly flat but there will be a thin layer of material on the surface that is at the yield stress. In this case there will be high stresses inside the part and it may respond to a relatively small additional force like gravity. I am sure other things could result in high internal stress. The fully stress relieved part may see no effect.
 
Gard and the members, next time you scrape a new casting Straight-edge do me a favor and as your scraping it drill and tap a hole in one end, screw in a eyebolt and hang it up with a rope or nylon strap and hit it with a soft-blow hammer and Vibration Stress relieve it. I have called it "ringing it" for years but my friend Professor Alex Slocum of MIT said it is a common practice called vibration stress relieving. On raw stress relieved casting - new ones - will move or as you call it creep Or if you drop a good one and it gets hit real hard, do it after you stone it. On dropped SE's you may have to do it several times too. So once you get it flat and hinging at 30% - hang it up and hit it. It will be like a tuning fork and vibrate. Then check it again and I swear it will change the hinge points. When I scrape SE's I do this a lot until it doesn't creep/ move. . Here is what I mean. Look at minute 5:42. My friend Stefan Gottswinter did a You Tube show. It's a King-Way SE, but ALL SE's will move! (No scientific paper, just 50+ years of experience.) Stefan has a couple of other show on scraping.
 
Last edited:
It is interesting if you listen carefully to the dialogue above. Someone says "It worked" and then someone else says in a skeptical tone "weeeellll, there's too many blues" Unfortunately there were no before and after photos shown.

And then there was the class attended by one of my customers whom I only knew otherwise by seeing his name in the Seattle Metalheads online group who watched a similar demo by Rich and subsequently he and two other classmates reported to me that they could see no evidence of movement. Realizing that one of the folks was going to attend the class I asked them to be particularly attentive and see if it did or did not work and to report as he saw it. The other two I did not otherwise know.

So, if we are going to report observations like this, I feel it is appropriate to report them all. That whole business I am quoting was previously discussed here on PM.

Truth in advertising.

I really do hope other people will replicate this test to see if the evidence is convincing or not.

Denis
 
If you bend a bar until it just barely hits the elastic limit it will remain fairly flat but there will be a thin layer of material on the surface that is at the yield stress. In this case there will be high stresses inside the part and it may respond to a relatively small additional force like gravity
I am of the opinion that this elastic non elastic internal stress profile or curve is the underlying reason why some people believe that "bent shafts want to be straight" (i know i have seen people on this site say so, but no explanation is given.)
First time i read that was nearly 4 years ago. Didnt believe it, until i had a bent motor shaft to straighten, and increasingly hard hammer blows did nothing until i hit it hard enough and it went back very close to straight (no visible wobble by eye) in a single blow.

I have thought to bring the subject up but have not found an appropriate thread yet.
 
It doesn't seem outside the realm of possibility that non-stress-relieved materials could have this happen at all to me. (Movement with a good knock or vibration). Or even those that had strains introduced after stress relief. I've seen and heard people talking about stress and strain in materials often, many seeming to think of it as a singular thing in a material, when in reality there are of course countless regions under differing amounts of stress and strain throughout any given material. Some can be pulling one way, some another. It is a complex issue, and it's one reason that thermal stress relief is important.

And vibratory stress relief is a thing, and it is real. But I think to work well it has to move specific areas at the right frequency. Not every place, especially on a complex shaped part will vibrate the same, so it's not as good as thermal in every case, I think. I have seen those that cycle the vibratory frequency, which should help with that, but I guess they'd probably also need to cycle the amplitude. That would seem to get very complex and time consuming.
 
Last edited:
It doesn't seem outside the realm of possibility that non-stress-relieved materials could have this happen at all to me. (Movement with a good knock or vibration). Or even those that had strains introduced after stress relief. I've seen and heard people talking about stress and strain in materials often, many seeming to think of it as a singular thing in a material, when in reality there are of course countless regions under differing amounts of stress and strain throughout any given material. Some can be pulling one way, some another. It is a complex issue, and it's one reason that thermal stress relief is important.

And vibratory stress relief is a thing, and it is real. But I think to work well it has to move specific areas at the right frequency. Not every place, especially on a complex shaped part will vibrate the same, so it's not as good as thermal in every case, I think. I have seen those that cycle the vibratory frequency, which should help with that, but I guess they'd probably also need to cycle the amplitude. That would seem to get very complex and time consuming.
When I was making and scraping SE's to sell the process was to high heat 1170 F stress relief, cooked it 8+ hours and let it cool inside the furnace overnight, then rough machine it and then High heat stress relieve it again , then finish machine it and then scrape it - ringing it like Jan and Stefan did while they scraped it. Years ago we did the stress relieving differently. We heated to 1100 cooked for 2 hours, lowered the heat to 900 for 2 hours, lowered the heat 200 again and again after 2 hours until it reached 200 and shut off the furnace. The new way of cooking it for 8+ hours worked better in my opinion. We did both Ekretz. I am sure there are other ways too, but that process worked for me.
 
I can't remember when it was as my classes all seem to merge after all these years. Had to have been 30 years ago in Malden Missouri at a Federal Mogul plant that cast and machined aluminum auto pistons. They had centerless grinders that ground the pistons oblong a few thousands. The machine rocked the pistons in and out as it turned. They had me come in and work with 3 shifts so they could all be scraping the machines as a team. Prior to the classes they could not work together because they all had their own idea's and argued who was right, so each shift had their own machine they were rebuilding. It was a mess at first. In that part of the country in the morning the temp was in the high 60's and low humidity and by noon it was getting hot and humid. The plant did not have AC and we cooked after lunch, even with fans.

We were scraping the saddles of the Centerless Grinders and we used a 36" Brown and Sharp Straight-edge. If we touched up the SE in the morning, it would hinge perfect, but by 1 PM it would change and hinge on the ends. It was a pain as we had to scrape the SE in the afternoon to get it to check good on the Granite plate. I could not believe the SE would change with high humidity. We let it sit on the plate and never touched it and in the afternoon it changed. They found another SE in a cabinet the same size as the other one. We ended up using the one SE in the morning and the other one that we scraped in the afternoon. I went back a year later and they had built walls around the maintenance area and installed AC. It was nice and we only needed 1 SE.
 
I worked with my setup for a half hour or so today. The only modification I made was to put a lead-shot bags over each end of the SE to improve setaing and I gave a sharp downward tap over each swivel support. Those two things seems to eliminate any erratic (squirrelly) behavior and now test application of pressure and release results in an immediate return to baseline.

I tried heating the bow of the SE with my bare hand in the 43 degree F ambient workspace and found the 36 SE bowed upward a tenth in about 15 seconds when gripped in the center of the bow. I repeated this test a half dozen times during the morning and afternoon while taking brief breaks from working on a casting pattern in the heated portion of the shop. Grasping the bow ends should cause substantially less deviation due to the expected mechanics of expansion there. I'll try that tomorrow to see if that is true.

I also did some simple load tests and found that a ten pound load placed on the center of the bow caused a .00004 sag of the SE. Unloading resulted in a nice prompt return to baseline each time. I am starting to gain some confidence in this setup.

Denis
 
That's a term we use in the profession of scraping. But were finding the fulcrum points or pivot point . It is 30% from each end of the ends of the part. On a typical application you can eye ball it. If your want to get fussy you can measure it. You set a Straight-Edge on a granite plate and with your left hand you hold the SE and pull it toward you and push it away approx. 1" both ways from where it is sitting. Then do it with your right hand. If it hinges / pivots in the meddle of the part you scrape that area, if it hinges on the extreme ends it's high there.
Scroll to 5:30 minutes to this and he shows it. A typical way we do it. I am a bit silly in the beginning. lol
 
Last edited:
It doesn't seem outside the realm of possibility that non-stress-relieved materials could have this happen at all to me. (Movement with a good knock or vibration). Or even those that had strains introduced after stress relief. I've seen and heard people talking about stress and strain in materials often, many seeming to think of it as a singular thing in a material, when in reality there are of course countless regions under differing amounts of stress and strain throughout any given material. Some can be pulling one way, some another. It is a complex issue, and it's one reason that thermal stress relief is important.

And vibratory stress relief is a thing, and it is real. But I think to work well it has to move specific areas at the right frequency. Not every place, especially on a complex shaped part will vibrate the same, so it's not as good as thermal in every case, I think. I have seen those that cycle the vibratory frequency, which should help with that, but I guess they'd probably also need to cycle the amplitude. That would seem to get very complex and time consuming.

It definitely is a thing in steel weldments. That has been shown by reasonable controlled studies and is discussed by ASM, for example. Unfortunately, there are no confirmatory controlled tests for cast iron unless you know of such. If you do, please point them out. ASM stated that there were none for cast iron as opposed to steel though a number of years have passed since that statement was made. ASM:"Stress Relief. The relief of residual stress is accomplished by heating the iron to a temperature at which the stress is relieved by rapid creep. Of course, complex shapes must then be cooled uniformly so that stress is not reintroduced. Vibration has been promoted as a method for providing stress relief to iron castings. This procedure has not been demonstrated to be successful in a valid test."

It is noteworthy that they specifically speak to the issue and that this handbook is in the hands of engineers, scientists, and academics of all types plus those workers who want to read it. In addition, it seems likely that if some firm that sells vibratory stress relief devices for iron had a way to contest this statement they would have done so and would almost certainly sue for libel.

So, I respectifully very much doubt that vibratory stress releif for iron is a thing. But, I am willing to be convinced otherwise if decent careful studies can demonstrate it. In fact, I would then embrace it.

On the other hand, it is not far back in history that, for example, the "learned" men of medicine used and taught their proteges the benefits of blood letting and the foolishness of "germ theory." And they were quite disrespectful (and worse) of those with the gall to question them. They forcefully used all the usual illogical ad-hominems and appeals to authority in attempts to ridicule and discredit skeptics not to mention the occasional physical terror techniques. It was only by painstaking scientific research that such deeply entrenched beliefs were finally debunked though there are still a few holdouts here and there.

Denis
 
Last edited:








 
Back
Top