A recent discussion at Reelsmithing caused me to think about the size of abrasive grits and the roughness of the surface that they would produce. My intuition was that an abrasive particle of 0.001 inch diameter might produce a surface roughness of 0.0005 inch or so. But there seemed to be a mismatch between the commonly available grits and the roughness (Ra) typically produced; i.e., the reported roughness is much less than than 0.5 * grain size.
So I googled around and found a chart that answered this question. It comes from Pace Technologies, a purveyor of supplies for the metallograph business. Clearly people who know their grits. The first four columns in the chart below are their data.
I added the last three columns. Columns 5 and 6 just convert their data to English (note micron = micrometer and mil = 1/1000 of an inch). Column 7 gets at the heart of the matter. It is just the ratio of the data in the previous two columns, and it shows that for a given abrasive particle size, you can get a surface finish with Ra of about 1/200 the particle diameter. So now I better understand what grit to consider.
This data is for finishing hard steel and it probably makes some difference if you are working with softer metals, but I have not yet discovered this piece of the puzzle.
In the Reelsmithing discussion, I steered the attention away from Brent’s very sophisticated reel and toward my own topic of interest. I apologize to Brent for this. When I get the chance to quiz a guy who knows more than me, I cannot resist.
Update 19 Dec 2014: I was trying to figure out sharpening stones, which are usually described as “fine, medium, coarse”. Here is a chart that I copied from the catalog of Woodcraft Supply LLC.
Update 19 Oct 2015: Today I found a stupid mistake in the spreadsheet, and so have replaced it with a new one.
If you try to read the conversation in Reelsmithing, it will not make sense because Brent’s comments have disappeared.
The wikipedia article “Surface finish” has good info on what surface roughness can be obtained with what process.