Challenge to received understanding of the Coeff of Static Friction

[Enthalpy]

Unfortunately, the model with plastically deformed contact points is too inaccurate to hold. It predicts a friction coefficient equal to the ratio of the shear flow stress to the compression flow stress, which is often near 0.6, and that's wrong even for most metals, whose friction coefficient is often between 0.12 and 0.4.

 

These two numbers don't even vary in a correlated way.

 

I know it's one historical model that is taught in books and courses, but there is no reason to believe it, use it, no teach it.

[studiot]

 

Unfortunately, the model with plastically deformed contact points is too inaccurate to hold. It predicts a friction coefficient equal to the ratio of the shear flow stress to the compression flow stress, which is often near 0.6, and that's wrong even for most metals, whose friction coefficient is often between 0.12 and 0.4.

 

These two numbers don't even vary in a correlated way.

 

I know it's one historical model that is taught in books and courses, but there is no reason to believe it, use it, no teach it.

 

 

Really?

 

Then I would say it is pretty good.

 

This table from the International Handbook of Mathematical, Scientific and Engineering Formulas, Tables, Funtions, Graphs, Transforms (Fogiel)

makes interesting reading

 

 

There are also extensive tables of special conditions that can reduce some of these towards 0.1

[Enthalpy]

The table excerpt provided is wrong, by much. For instance, we take a coefficient of friction of about 0.17 for hard steel against hard steel in bolts, without any grease.

[studiot]

 

The table excerpt provided is wrong, by much

 

Perhaps you should tell Mark that?

 

The table appears in 'Mark's Handbook for Mechancical Engineers ' as table 3.2.4

 

though I would have thought by the tenth edition someone would have already pointed that out if it were untrue

 

Mark's is, of course. the 'CPC Handbook' reference for this field.

 

There are some interesting notes and other tables, however, since conditions such as the presence of surface films make a big difference.

 

Table 3.3.3 lists

 

Friction for Steel on Steel

 

dry 0.78

Oxide film 0.27

Sulphide film 0.39

 

Friction for Copper on Copper

 

dry 1.21

Oxide film 0.76

Sulphide film 0.74

 

Engineers should learn to read the correct table correctly.

You may have been thinking of the data from the oxide tables.

 

Mark gives further data for specially ground harder steels that reduce these values even further.

Edited October 18, 2014 by studiot

[Enthalpy]

The table is obviously wrong, whoever wrote it many times.

Just take two metal parts, one being flat, and check at what angle they slip. You won't get angles like 45°.

 

Better take table for screws tightening. These are checked experimentally by many people - I did too. They give values between 0.1 and 0.2, not around 1.

 

Additionally, the friction coefficient varies much (although slowly) with the pressure, which the ratio of the shear to compressive elastic limits can't explain.

 

These are a few reasons why this attempt is abandoned for decades, along with the hills and valleys model.

[studiot]

 

The table is obviously wrong,

 

These are a few reasons why this attempt is abandoned for decades, along with the hills and valleys model.

 

 

 

This is an unsupported claim

 

citations and references needed

 

(I provided mine and they refer to more than one respected publication)

 

Further you have taken absolutely no notice of what I said about oxide layers, which must exist on your simple test parts.

 

 

Just take two metal parts, one being flat, and check at what angle they slip. You won't get angles like 45°.

Edited October 20, 2014 by studiot

[Enthalpy]

Engineering by thousands of people make a claim properly supported, far better than any publication. Have a look at the torque used to tighten screws, these tables work, I did the measurements.

 

Or put an object on a banked surface and observe the slipping angle. I did. The coefficient of friction is regularly around 0.2 to 0.3 for metals, even for this small pressure.

 

One other argument against the shear-to-compression ratio theory is that also the hardest material influences the coefficient of friction. Say, polished chromium and nickel coatings are way harder than bronze, and they show different coefficients.

[studiot]

You neither answered my points nor provided accessible references for your claims, both of which are against the rules of thsi forum.

 

I see no point discussing this further.

[Enthalpy]

For knowledge as accessible and well-spread as coefficients of friction, I feel it overkill to look for one source. The tightening torque for screws is in any book or handbook for mechanical engineering. Having used friction for years on a daily basis to design about everything in crash-test hardware, accumulated many paper and digital sources and tables on the topic, and having measured by myself because sources aren't so reliable, I don't really feel the need to point at one research paper or book. It's a bit like telling that bread is baked: one wouldn't ask for a peer-reviewed source about it.

[Nicholas Kang]

I agree that we mustn`t research on one source but many sources can help to verify the previous one to chech and see if any of them have a problem. Daily experience isn`t bad but thay are restricted by parameters like air resistance, gravitational force or even weight/mass etc. In laboratories, scientists are good at controlling those parameters to ensure nearly accurate results and they published their papers. Thay are highly technical but very accurate. You cannot assume daily practices are accurate becuase parameter controls are out of your hand. You don`t expect to check every corner of your tools or hardware to ensure no rust, no dirt or any variables that might alter your results whenever you use them, do you?

[Enthalpy]

But when every engineering book gives coefficients like 0.12 to 0.4, when I made measurements dozens of times and got results consistent with engineering books, when a simple inclined surface should have sufficed to everyone to see that metals glide far before 45° - then, a table claiming coefficients like 0.78 or 1 is obviously wrong and to be discarded.

 

Just for comparison, 1 is the kind of coefficient gotten from rubber tyres on a street.

[imatfaal]

Enthalpy

 

"obviously wrong" - needs to be backed up with the observations which cause it to be obvious when you are contradicting a well renowned source. If we learn anything from science it is that without repeatable measurements the obvious is quite often not true.

 

I will ferret around my store of junk to see if I can get a few pieces of material on Studiot's list that I can test. The coefficient of static friction is dead easy to measure - it is simply the tangent of the angle at which the object starts to slide under gravity

[studiot]

Imatfaal, if you are doing any tests you will need to create chemically pure surfaces for your samples.

 

Table 1 is just what is says pure aluminium on pure aluminium etc

not aluminium with an oxide layed on aluminium with an oxide layer.

 

For that you need table 2 and 3 and 4 etc.

 

I can only refer you back to Mark's, or any other standard Mechanical Data source.

Kay and Layby have some as do Machinery's.

[imatfaal]

Imatfaal, if you are doing any tests you will need to create chemically pure surfaces for your samples.

 

Table 1 is just what is says pure aluminium on pure aluminium etc

not aluminium with an oxide layed on aluminium with an oxide layer.

 

For that you need table 2 and 3 and 4 etc.

 

I can only refer you back to Mark's, or any other standard Mechanical Data source.

Kay and Layby have some as do Machinery's.

 

Yeah - I had begun to realise; do you have glass on glass? or is that too variable a material?

 

I have a nice disc (which can be cleaned up) of 950 silver (ie 950/1000 slightly better than stirling 925/1000) and a sheet of clean glass - would that make a valid approximation? And where could I find silver on glass?

 

I am sure I must have some nice copper somewhere and you list copper on glass; whilst I understand that the purity will affect the outcome, would it not provide a good approximation?

[studiot]

The point of metals is that they all react with atmospheric oxygen so a chemically pure surface is oxidised within less than a second to few minutes.

Some oxides are more stable, persistent and protective than others but all act as a sort of lubricant in thin films.

 

I am in Dundee this week so access to data is limited, however I will se what I can dig up.

[Enthalpy]

The coefficient of static friction is dead easy to measure - it is simply the tangent of the angle at which the object starts to slide under gravity

 

Good idea. Just remember that in such experiments, the pressure is usually very small, so one gets coefficients much larger than at the pressures useful in mechanical engineering. Try to have small contact surfaces, put loads on the parts.

 

If experimenting with glass, don't take flat parts (window glass is very flat), or you'll measure only atmospheric effects: air cushion or suction pad.

 

As a source of data, first check if it indicates at what pressures the coefficients of friction are measured, since the coefficient varies by a factor-of-three with the pressure, and discard the table if no pressure is given. That's a first, very efficient screening.

 

Mechanical design basing of friction, as I did for several years, uses tables that are pages long just for the most common alloys. A table like the extract given in this thread, with one single value per material pair without an indication of the pressure, would definitely not qualify as "reputable" - it would be discarded immediately. A query for a coefficient of friction usually ends with "I measure it by myself to have a credible value".

[Enthalpy]

OK, now the correct values of coefficient of friction are moved to speculations.

 

Just for info:

- I'm possibly the only one here who measured a coefficient of friction

- I used friction coefficients for several years in my job

- The figures I cite are in every book for mechanical design

- Such decent books cite the pressure used to measure the coefficient, to be meaningful

- Have a look at the fastening torques for screws. That's something observed more than once a year.

 

Sorry, but I give the proper figures here, which are the commonly accepted ones.