Surface Roughness spec on a short length

[dgallup]

I have a turned part which gets partially over molded with plastic. Only about 3 mm are inside the plastic. We spec a surface roughness of Ra 3.2 to 20 with a cut off length of 0.25 mm for this area because we want the plastic to tightly grip the part. Outside of the plastic the surface finish has to be quite smooth. The supplier greatly increases the tool feed rate where we want it to be rough with a single point tool, the part almost looks like it's threaded. The problem is the 0.25 mm cut off length filters out all the waviness and I get a surface finish reading of slightly less than 1.0. If the surface was longer I could use a 0.8 mm cut off length and the parts would pass but the region is too short for that. What kind of measurement can I use on a short surface that won't filter out the waviness? These parts are good, you can run your thumb nail over them and feel the roughness, just can't measure it.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[IRstuff]

If what you have is acceptable and you can make a distinct measurement, why can you just specify the results of that measurement, with some level of guardbanding?

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[dgallup]

Well, the problem is the current measurement system throws out the part of the profile waveform that we actually want so if I just lowered the minimum acceptable Ra value we might accept parts that don't actually have the surface we are looking for. I don't want production to be dependent on "tribal knowledge" that gets lost the next time purchasing decides to switch suppliers.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[btrueblood]

Spec. the "threads per inch" of the finish you are getting, and a pitch diameter. Root and crest truncation or radii also if you think they are important. That should adequately define the actual finish?

 

[dgallup]

How would you put that into a standard surface finish symbol?

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[tbuelna]

With a turned surface I believe standard practice is to specify a cutoff length that includes at least 10 tool marks to produce a valid sample. The evaluation length should also include at least 5 cutoffs.

There are some online calculators that help determine the speed/feed rates for producing a desired surface roughness profile by turning. Achieving the specific surface texture profile you require in the location where the plastic material will be over-molded may require more than just a change in feed rate. Possibly a secondary operation using a cutting tool with different geometry.

If you need more information on how to define a specific surface texture condition on your engineering documentation, take a look at ASME B46.1-2009 and ASME Y14.36-1996.

 

[dgallup]

I don't have 10 tool marks to measure, only about 3 or 4. Therein lies the problem.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[btrueblood]

"How would you put that into a standard surface finish symbol?"

What I'm saying is don't control your finish with a surface finish symbol, control it as a threaded surface.

 

[IRstuff]

What exactly are you trying to control with the waviness requirement, and how would you have done with a standard machined part?

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm

 

[pylfrm]

dgallup,

If you only have 3 or 4 peaks and valleys, I don't think it will be very practical to specify or measure this in terms of surface roughness.

I'd recommend actually showing ridges on the drawing, defined with dimensions, tolerances, and notes as appropriate. Without knowing more details, my first thought would be to tolerance the crest diameter, radial depth, and axial spacing. Some additional control of the form (such as slope) might be justified as well, and could make the axial spacing tolerance unnecessary.

To provide some manufacturing flexibility, it may be helpful to allow either a single-start helix or multiple circular ridges.

Calculators similar to this for machined surface roughness and cusp height are widely available.


pylfrm

 

[dgallup]

We have found that if we begin the trace on the smoother surface inboard of the rough zone so that we have a long enough trace to use a 0.8 mm cutoff length then we are able to measure what we are looking for. This contradicts normal measurement practice but so we added a note to "trace sufficient length (including transition zone) for 0.8 cutoff".

Thanks for the tip on turning surface finish calculators.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[powerhound]

You could just forget about the filtering since the surface is so short. Consider specifying a Pa parameter. This is the average calculation of the primary profile.

John Acosta, GDTP Senior Level
Manufacturing Engineering Tech