ISO best practice for measuring rough surfaces

[[Za...]]

Our company uses metal 3D printing to manufacture parts. After printing, the parts get machined to bring them into tolerance wrt our drawings, but a couple of the features are left as-is and are used as datums. The surface of these features is very rough (think like a metal file, or sandpaper), and anything you measure wrt them tends to be a total crapshoot; Sometimes I will end up with, for example, a position that is OOT by 1-3mm and my bosses aren't happy.

What are some best practices I can follow for measuring rough-textured surfaces? Obviously I don't expect them to fall into tolerance, I just want to be able to say I am doing the best I can and the outages aren't anything on my end. Ex. polylines vs. single points, filter settings, etc.

Thanks!

[[Jo...]]

I imagine it is like measuring cast features. If you are scanning with a Ruby probe you may want to consider single touch points, so you do not wear your probe out scanning. This in turn will throw your data out even worst.

You might be able to use a different probe material that are more expensive that will be able to scan without premature wear. Silicon nitride probes are what you would be looking for.  I haven't tried this personally; in the past I could get away with single touch points. 

With the ladder option you may have luck with the filter/outlier elimination settings to filter out crap data points. It is hard to filter everything though.

Hopefully other users can provide more input.

[[Ch...]]

I've measured on some metal 3d printed parts and I know exactly the type of surface you are talking about.  

Joseph pretty much nailed it.  There are two types of strategies I choose to work with.   You either scan the crap out of it using a very heavy point density.   Or you can use the minimum number of points required for the feature.  It'd be a good idea to coordinate with engineering to decide exactly where the points should be taken.  Kinda like datum locators on castings.  

[[Ze...]]

You can also use the largest probe diameter that will accurately measure the surface, as the high points are usually the mating surface and you'll generally get a smoother probed surface to work with.