Alternate ways to measure surface roughness?

[Quietrabbit]

I am trying to compare the surface roughness of 3D printed specimens for a school project, but I am realizing I don't have access to any of the standard equipment that would be used to measure it. Are there alternative measurements I would be able to use that would yield similar results, using equipment commonly found in a high school?

 

[berkeman]

Paging @Ranger Mike

 

[Vanadium 50]

am realizing I don't have access to any of the standard equipmen

That's rough.

 

[Vanadium 50]

OK, seriously. Do you need to compare to an absolute standard, or is it sufficient to compare items you are making with each other? I am not an expert, but I do understand enough to recognize that these are two distinct problems.

 

[Quietrabbit]

OK, seriously. Do you need to compare to an absolute standard, or is it sufficient to compare items you are making with each other? I am not an expert, but I do understand enough to recognize that these are two distinct problems.

Just to each other should be sufficient. Thanks

 

[DaveE]

Can you print objects solely for the determination of local roughness, or do you need to measure a specific shape, or any arbitrary shape?

I'm thinking flow resistance in a tube or thin channel may be easy to do comparatively. i.e. how much low pressure water can flow through a small channel in a fixed time period.

You may also be able to do something with friction or stiction comparisons, but this stuff is notoriously variable.

If you have the time and inclination to build some electronic gizmo, then a laser or LED scattering off of the surface to a photodetector could be made to work. This is how Tencor of KLA-Tencor made lots of money testing silicon wafer surface quality. Spoiler, they did it a lot better than you (or I) ever could.

 

[Baluncore]

Are there alternative measurements I would be able to use that would yield similar results, using equipment commonly found in a high school?

Just to each other should be sufficient. Thanks

Gently stroke the surfaces of two samples with the wooden end or side of a pencil. Switch back and forth between the samples, until you can hear the difference in loudness and frequency for the same stroke.

The louder sounding item has the greater surface roughness.
The higher frequency item has the finer detail.

You may also feel the roughness, when you hold the pencil in the sensitive tips of your fingers.

 

[Ranger Mike]

A short course in Surface Texture​

Surface Texture Measurement has been around since 1938 when Dr. Ernest J. Abbott invented the Profilometer.

​

The American Society of Mechanical Engineers (ASME) developed a Measurement Standard that defines “Surface Texture” in the 1940s to govern “the geometric irregularities of surfaces.” B46.1 defines surface roughness, waviness, and lay, as well as parameters for specifying and controlling surface texture. The latest copy of ASME B46.1- 2019 is 130 pages. It is the Bible for Surface Texture.
Note the Normal Section Plane above. It is perpendicular to the Nominal Surface.Gear Inspection Machines use the Skid datum method with a displacement probe. Great for Roughness Parameters and in compliance with the ASME, DIN , ISO Standards. Note the bottom illustration of the Reference surface instrument. It used a very high precision surface internal datum. The linear movement of the stylus is well under 10 millionths of an inch Straightness accuracy.

​

In skidded gages, the sensitive, diamond-tipped stylus is contained within a probe, which has a skid that rests on the workpiece. Skidded gages (upper) measure roughness only. Skidless gages (lower) use an internal precision surface as a reference. This enables skidless gages to be used for measurements of waviness and form, in addition to roughness. Source: Mahr Federal Inc.

 

[Ranger Mike]

Now that we have a little back ground, the short answer is no, you are very limited on ability to analyze surface texture other that the fingernail approach. You still can gain Insite and tie a number to the surface texture by making a master specimen to use as the calibration master and several specimens. One specimen to be considered non acceptable or Bad and one that will provide proper function or Good and of course the master specimen or Best. I would even color these Red, Green and Gold. Now you can invite the people to scratch the specimens with finger nail and give opinions. This is still arbitrary. It would be nice to have a device to provide a number. More on this later.

 

[Ranger Mike]

You can fabricate a device to give a number by using compressed air. You have an air source, a device to indicate ( can be an air meter or air float) and a contact probe. The Probe would need a flat surface. It would have a small outlet hole thru which air flows. The idea is a flat surface with low surface texture will have a IN LINE high pressure flow number and yield less air flow because it would seal better. A bad surface would leak air a whole lot and have a low IN LINE flow number as the uneven surface provides many exits for the air. Think of your garden hose. When you put your thumb over the hose end water flow is reduced and you feel pressure on your thumb ( high IN LINE pressure). Now move the thumb over so half the hose end is uncovered. Less IN LINE pressure but more flow. There is a lot of info on pneumatic surface finish ( surface texture really) on line.

In the search bar at top of this page type in ideal gas law..will explain this law in detail so you may justify your device and its principles

 

[Quietrabbit]

Would it be possible to use an optical microscope and some sort of image analysis? I believe the ones we have go up to 400x.

 

[Tom.G]

Would it be possible to use an optical microscope and some sort of image analysis? I believe the ones we have go up to 400x.

How would you measure the depth of the irregularities with a microscope?

 

[hutchphd]

I would try to use a laser at pretty low angle and measure the resultant scattering with a light meter or calibrated photograph . See if that can reliably differentiate the roughness range you are interested in. Are you interested in the scale (size) of the roughness or the depth? (different wavelengths may indicate the size).
There is good software to do analysis of photos to count cells in a field and similar stuff. The free package from NIH called ImageJ might work well. You may have to be clever here! Good Luck.

 

[Ranger Mike]

now you know why the Profilometer took so long to develop. There is a lot of math required to evaluate a surface. You are literally dealing with a squiggly line. It has peaks and valleys. Do we measure peaks only? Valley depths? both? do we average them. There must be a process that is measurable, repeatable and accurate. That is why ANSI has over 120 parameters dealing with surface texture.

 

[Quietrabbit]

How would you measure the depth of the irregularities with a microscope?

Good point . If I kept the light the same would shadow color/darkness work. I probably could purchase a standard to compare to. I do need this to be replicable, but the main goal is to get a statistically significant difference that one pattern is better than the others. Or not. My grade doesn’t actually depend on if I find anything, just that my research and methods seem valid. Though using a proxy for measurement is totally fine as long as I can justify it.

 

[Quietrabbit]

Would it be possible to use an optical microscope and some sort of image analysis? I believe the ones we have go up to 400x.

I am assuming it would be difficult to get a good sample but if I cut the 3d printed part really thin and look at the edge in question I’m guessing it would look like a really sketchy version of the graph of surface roughness?

 

[Ranger Mike]

which begs the question, how do you define a squiggly line? How ca n you tie a number to good part versus bad part? Destructive testing is expensive.

 

[Lnewqban]

Just to each other should be sufficient. Thanks

Following the idea of post 7 above:
For a simple comparison, I would use a mechanical pencil with a soft lead.

https://www.officedepot.com/a/products/710186/Pentel-Super-Hi-Polymer-Leads-05/

If it is possible to write lines of similar length over each sample material using similar hand pressure, I would compare how much material each led has lost.

The difference is very evident when you draw or write on paper of different roughness.

 

[Quietrabbit]

I would try to use a laser at pretty low angle and measure the resultant scattering with a light meter or calibrated photograph . See if that can reliably differentiate the roughness range you are interested in. Are you interested in the scale (size) of the roughness or the depth? (different wavelengths may indicate the size).
There is good software to do analysis of photos to count cells in a field and similar stuff. The free package from NIH called ImageJ might work well. You may have to be clever here! Good Luck.

I think I am seeing a difference between the roughness when i briefly tested some random prints I had lying around. The resulting ”reflection” beam seems shorter for more pitted textures, is that right? If i point the laser parallel to the grain then i get a circular arc that does have some irregularities But i cant tell if thats just from the laser itself. The rougher surfaces also seem to result in a less beam. I am more interested in depth.

 

[Quietrabbit]

I think I am seeing a difference between the roughness when i briefly tested some random prints I had lying around. The resulting ”reflection” beam seems shorter for more pitted textures, is that right? If i point the laser parallel to the grain then i get a circular arc that does have some irregularities But I cant tell if that's just from the laser itself. The rougher surfaces also seem to result in a less beam. I am more interested in depth.

I got a rough laser speckle pattern with a laser pointer and a backwards camera zoom lens, but I think that is too fine because I cant see the speckle good over the visual surface texture in the print.

 

[hutchphd]

"Roughness" is not a well-defined term. How will this characterization be used? Do you want to be able to see differences in a process or what? If you just need to look for changes, then there are a variety of approaches. If you need to characterize the surface, you will need to be more comprehensive in the measurement. What are you trying to do?

 

[Quietrabbit]

"Roughness" is not a well-defined term. How will this characterization be used? Do you want to be able to see differences in a process or what? If you just need to look for changes, then there are a variety of approaches. If you need to characterize the surface, you will need to be more comprehensive in the measurement. What are you trying to do?

I suppose by roughness I mean flatness. I have 3D printed samples with varying parameters that I want to see how they effect the top surface. I am mainly looking for depth measurement because the visual differences seem to mostly be sagging in areas without as much infill pattern (intentional) and I want a quantitative way to justify that one set of parameters creates the smoothest top finish. Ideally I would also measure this on a rounded shape but I've about given up on that and will settle for horizontal.

 

[Quietrabbit]

which begs the question, how do you define a squiggly line? How ca n you tie a number to good part versus bad part? Destructive testing is expensive.

Would it be easier to do that if I create a rubber counter mold, more sliceable? I assume with some math I can correlate the microscope image to actual measurements based on the magnification of the lens, then with some image software get measurements of all the peaks and valleys.

If I interpreted it right, I believe this paper did that
https://link.springer.com/article/10.1007/s13632-013-0092-6

 

[hutchphd]

The local flatness can be determined by projecting a grid (either via a transparency or a hologram in a laser beam and taking a photo. You can also scan the sample like a bar code reader does.
The surface finish likely will be evident with a near-glancing incidence beam (maybe just the size and shape of the specular reflected spot would suffice)? Do you have a laser pointer to play with?

 

[hutchphd]

You want to measure whether part a is "better" than part b but you need to understand what "better" means! Probably before you try to measure it......process development is interesting.