Help with surfacing.

[mcpgmr]

I'm designing a hydro form tool using surfaces that were done with a faro laser scanner. The faro guy left out some important veriable fillet surfaces and I'm trying to fix it with no luck. Any help would be greatly appreciated. I can upload the file or email it along with a graphic showing the missing surfaces. Thanks in advance!

[Hardmill]

I can look in the morning.

Send it.

 

 

PEACE

[Colin Gilchrist]

I'll look when I get home from work if you want.

 

colingilchrist (at) comcast (dot) net

[crazy^millman]

Should have used Verisurf and we could have made life a lot easier for you. You are in good hands though. Do you have the original point cloud if so send it to my Verisurf email ron(dot)branch(at)verisurf(dot)com and I will take a look at Monday.

[mcpgmr]

Thanks guys. This is a part that the customer could not supply us with a model only a mylar and sample part. We had faro in doing a demo anyway so he took a crack at it. He had some trouble finishing it so he took the scan data back to his office, did his voodoo and emailed it to us. I didn't realize the fillets were missing until I started designing the tool yesterday. I'll take a picture of the sample part and post it here and upload the .stp file to the ftp. Have to get some coffee running through my vanes first.

[MotorCityMinion]

Sounds like your looking for exact data. If not:

 

Are the variable rads functional or for relief?

If you can get away with close enough, put some radius gages on the sample and state some key transition locations and I'd like to take a shot at it.

 

On my second Cup o Joe now. I'll check back in a few.

[mcpgmr]

Here are some pics...

 

 

 

I checked with a radius gage and it looks like between 9/32 and 5/16. I have +/- .03 to work with. I'm uploading the convex side surfaces that I have and the concave side will hopefully just be an offset by the thickness of the material which is .040". I'll call the file top_bump.mcx and .stp and put them in the mcx_files folder. Thanks!

[mcpgmr]

Files are uploaded. Thanks in advance!

[MotorCityMinion]

Taking a look now.

[MotorCityMinion]

SO far no good. The bump isn't parallel or perpendicular to the plate, in any plane, and sits just above it. Best I could hope for is to rotate the bump surfaces till the four corners sit good and sqr, then recreate the intersects of the wedge shape, make a new extrusion from that, then fillet it. Sorry I couldn't,t be more help here as this ones a bit over my head. I'll keep messing with it anyway, if I come up with something close, I'll send it to you.

[MotorCityMinion]

After lining up the plate and bump, I extracted what data I could and made some surfaces. I have no experience with a faro laser scanner but it looks like somethings off here. Either the scanner is dead nuts accurate and the part is a POS, or vise verse. Nothing is symmetrical anywhere on this part but that's fine.

 

I had to take some liberties with the data and I couldn't get the rads as large as the model you posted. What I did create is on the FTP now.

 

mcx_files folder/MOTOR TOP BUMP.x_t.

 

Looks pretty good when viewed through the bottom of a rock glass.

 

Photobucket is running at a crawl now, I'll upload some images when it's moving faster.

[mcpgmr]

It looks great Motor. Thanks! I'll look at more closely in the morning. Thanks you so much. What methods did you use to fix the crap I had? Again Thank you for taking the time to help.

[mcpgmr]

Checked it out and it looks great but unfortunately out of tolerance. The base part itself was probably out of whack because I created it in the same wcs as the faro surfaces that were imported. What I'm going to try and do is cut thru the faro surfs at the top of the fillets you gave me and then try to recreate the fillets by sweeping a surface around the 3d cut that I project onto the faro surfs. Main goal being to maintain the surface definition that the laser scan created. Thanks again for giving it shot for me. See pic below that shows approximate dimension in the front view. Yellow being the faro surf and the green being your model.

[MotorCityMinion]

"What methods did you use to fix the crap I had? "

Remember, you asked for it.

 

Note: There are many ways that this could have been done, probably a few ways that are much easier as well. Some will read this and laugh or critique it, others might learn from it. This is how I did it, this time (Education: School of hard knocks and this forum.) If I had the part, and +/_ .03 to work with, I would have put it on a comparator, took some measurements, and started from scratch.

 

I aligned the model and moved the lower right corner to 0/0.

 

New level, create curve, all edges.

 

New level. Drew a box around the perimeter of the Bump, making it square to the four corners.

 

From the right view, I eye balled the outer most geometry of the wedge shape, then xform projected it. Edit, simplify to geo so that it can be trimmed and remove any geo that that lies inside the profile. (Break at intersection works good for this). For the splines that could not be simplified and trimmed, I started a new level, then drew new lines over the splines, looking close for deviations. Delete the splines and replace with the new geo. Now we can trim those lines to meet the surface of the plate, which was moved to Z0 in the top view. Don't worry about the rads at the bottom of the bump, just trim the angular sides to Z0., The rads at the bottom get put on after the extrusion gets made.

 

Now, still being in the right plane, xform the new geo to the center of the box we made. Close the chain with a line which should be at Z0., straight in Y axis, as viewed from the top plane. Turn on the level with the curves we created earlier and verify that the new geo matches with the extremes of the model. This is the profile used for extruding the new "Bump".

 

From the Front view, I eye balled the outer most geometry of the bump, width vise, then xform projected it. Edit, simplify to geo so that it can be trimmed and remove any geo that that lies inside the profile. You'll notice that the geo is not symmetrical so what I did was mirror both sides about the center of the box, then trimmed away then inside geo again. Looking at the largest arc, I drew a chord across it then broke that into 2 pieces, created a line perpendicular to the chord that ends at the arc. Use that data to calculate the rad size, then draw that circle at the chord ends, and another circle at the intersect of the first two circles. We should now have a decent rad to use to calculate the length of the Bump. Create a vertical line tangent to the outside of the rad. The distance of that line to the center of the box is equal to half the length of the extrusion.

 

Extrude the profile in both directions from the center of the box by that amount.

 

At this point, I saved the file as a .stp and brought it into Solidworks. Combine the bump and the plate. (Merge in MC), put a .3437 R. fillet along the sides of the bump first, then a .3125 R. fillet along the bottom, tangent to the plate.

 

Offset the surfaces with Zero thickness, hide the original body, and shell the surfaces to .04 thick. We now have the cavity we need. Done.

 

BTW, the solids package in MC may have been adequate for this but I needed the practice in SW.

 

I'll clean up the MC file that I used and upload that, MOTOR TOP BUMP.MCX, this way, you can change the location or length of the bump if need be. If you want a larger rad at the base, you'll have to make the plate wider, put on the rad, then trim the plate to its original size. Putting a variable rad along the sides of the bump is a nasty affair as there is too much geo that makes up the profile. If you had to, you could take the time to rework the profile so that only 1 rad and three lines exist, making a variable radius actually feasible.

[mcpgmr]

Thanks again for all the effort. I hope I return the favor some how. I did have a silly idea. I'm going to run some surface tool paths on the bump using the top surface of the base as a check surface then save as an stl file. I should get the resulting radius of the ball mill as my fillets around the bump. Sounds easy enough.

[MotorCityMinion]

Your welcome. With over 2800 posts under your belt, I'm sure you've had a tip or two to offer.

Anyway, I up loaded the .MCX, with the merged X_T. The rads in the new drawing are clean at .3125 and .3437. Mess with the levels, everything is there to recreate this if needed.

 

If you overlay the new solid, in the.MCX, over the old one (in the same file, different level), you'll see that I wasn't able to recreate the large rads in the original model.

 

"I'm going to run some surface tool paths on the bump using the top surface of the base as a check surface then save as an stl file. I should get the resulting radius of the ball mill as my fillets around the bump. Sounds easy enough. "

 

Not silly at all actually. The first thing I did after projecting the profile was machine from the right, using a ball, with comp set to off and depths set to zero, then save the toolpath as geo. It got rid of the splines no problem but still gave me way too many lines to be useful for data retrieval. This is probably more accurate than trying to measure the .stl. Give it a shot.

 

I'm not sure what your after here, but using a .25 ballnose will fit nice and give you the exact radius's created in the X_T., no need to mess with the .STL, other than to verify, IMO. The extruded profile was created from the splines, lines and arcs supplied, so all you'll get from that is same gibberish that came from the Faro laser.

 

I'm thinking a Blend or Parallel toolpath would work good here.

[MotorCityMinion]

Just Saw your earlier pic. One way to get those large rads on there would be to make a solid from those original bump surfaces, then do a boolean removal. Cut the new solid in thirds then toss out the center and right end, put some fillets on the edges of the new solid to smooth it out, then mirror the left end for a symetrical cut. I'll save it without the cavity and mess with it some and see what pops.

[mcpgmr]

Solid from surfaces? Can I do that with a surface or do I need to make a water tight volume to solidify?

 

When I mentioned using a .stl file I was thinking of using the stl2surf chook and then keeping those surfs for creating my hydro form tool. I have not had any time to mess with it today but I'm diving back into it now.

[MotorCityMinion]

Not a problem, I tried it with this one. Its thickening the solid, then trimming it to plate that's an issue. The four corners are almost flat, yet the sides have arcs in them. I was able to turn it into a solid and thicken it, then do a Boolean remove (combine) in SW with the model I created. It looked about 90% legit but where the the corners touched, at the base, from one model to the next it looked goofy. I'm thinking it may work though because a 1/4 ball cant get into the small notches. Perhaps one of those new HSM toolpaths in X4 would smooth it out some. If I get a chance, I'll e-mail ya the revised file. Check back in a hour or so.

[mcpgmr]

Well I'm half way there. I projected the curves from the top of your fillets,onto the original faro surfaces, trimmed the original bump surfs to the projected curves then created new .312 arcs between the projected curves and your curves at the top of the planer surface of the part. Then from there I created net surfaces at each segment around the bump and I now have a variable fillet radius surf around the bump. I think it's close enough. Took me quite some time to do this but next time I know to use net surfaces. Now I have to offset the whole thing by the thickness of the part and make it a solid so I can add it to the rest of the part. Thats another hurdle I have to jump. Here is a pick of the outside that I just completed. Thanks again for helping.

[MotorCityMinion]

Just for kicks, I went back to the stp you uploaded, rotated the bump in the front plane, using the lower right corner as a point of rotation. Three corners on the same plane now with a Z height of plus .015 from the plate. I then xformed the bump to Z0, surface finish blend using a 5/8 ball and it looked nice. Lil polishing here and there, but still decent.

 

Assuming that the bump surfaces were just a little out of wack, this would keep you within the +/_ .03 your looking for.

 

Maybe run a indicator across the top of the Mylar sample and verify the highest Z point then compare it to the stp. If ya have the wiggle room, the technique I used above was almost too easy. (Still not trusting the location of that faro data). The pros will be on this tomorrow assuming that the ftp is up. Good luck.

[mcpgmr]

Nice! Do you have a resultant surface from it that I can use? Can you upload that one. I still don't like the blend I have at the bottom of the bump where it transitions from the flat angle plane to the small fillet area. I didn't see this until I finished and thought I had it close enough.

[MotorCityMinion]

At lunch now. Up to my neck in work at the moment and the files on my Stick at home.

 

Here's what I tried:

 

Your file, create curves for the bump. From the front plane, draw a line from the right to left corners. Analyze that line. Should be aprx. .55 deg. off. Xform, rotate (fRONT pLANE) the curves and the bump surfaces, using the LOWER RIGHT CORNER as your point of rotation, by that amount. 3 of the four corners should now be sitting at plus .015 in Z, as viewed from the top. From the top xform translate the curves and the bump z-.015. You should be on top of the plate. If you check some of the curves along X axis, they are also much more parallel than before. Top of bump should also be around .804 from the top of the plate. Surface blend with a 5/8 ball and check your results. Did you get a chance to measure the Mylar? Hope this helps, sorry I couldn't upload the file.