Wear in a cylinder

[bherrin]

I'm having an issue that I hope has a really simple answer that I'm just overlooking.

I have an outside cylinder with a really bad localized worn area, and i need to determine what the diameter would be if the whole thing was ground down to match the low spot.
My first thought was to do a patch scan of the wear and do a minindices to find the low point, paste w/ pattern 180 deg out and viola, but its proving to not work out as neatly as i had hoped.

Please someone tell me there's an easier way of going about this.

[Douglas]

I would use circles but minimum circumscribed, wear on one spot would not skew the circle center as much as least sq , and any graphic display would evaluate it pretty good.

Report min/max on that minimum circumscribed circle and the min would be quite close to what it cleans up at on center... as long as the probe hits it

[Douglas]

that is what this was I did the other day, showed me how much was worn off the bottom of that outside diameter, .033 so this one would clean up at 4.9928 by those numbers

Capture.JPG

[Douglas]

I can only add... worn parts is very challenging at times especially if datums have wear it gets fun. A good portion of my job is checking wear and failure investigations... new parts are a cakewalk in comparison

[Matthew D. Hoedeman]

If it is a localized worn spot, like, in the middle ONLY of the cylinder, as in, full cylinder above & below, measure the good parts of the cylinder, align to it, then do a scan around the worn area, get the minimum distance from cylinder CL to the scan, that is the smallest radius you can grind it down to.

[gibsonridge]

i agree with Douglas but i would also consider multiple diameters to evaluate center point of each and construct a calendar for evaluation as well

[Douglas]

that pic was me determining where original part axis was before wear, I used a few circles like that then constructed a line to be the axis and had decent results.

[Douglas]

as long as 180° or more of the circle has little or no wear the min circumscribed should fit quite well to the original diameter.

[gibsonridge]

agreed , without knowing size of the part and probing capability's (more data needed)

[KIRBSTER269]

using an auto Cylinder Concentric Scan choose MIN_CIRCSC like Douglas says. You can choose how many levels you want and it will produce individual circles and You could report out each circle you can pretty much get your info there, my example is an ID with 5 levels. so Your Min would be what you would justify smallest outside diameter.

Capture1.JPGCapture4.JPGCapture.JPG

Capture2.JPG

[Douglas]

I did find it easier to hit the exact spots I wanted with circles... I am only touch probing though, scanning might have made the cylinder my choice


edit... yes thinking about it I would want to tornado the fk outta an auto cylinder and get a million data points in a helix if scanning

[KIRBSTER269]

Douglas tell your boss not to be so cheap, get you a scanner

[Douglas]

hey stop that sht I think he reads this stuff lol... I actually would spend it elsewhere first though, our product is not too complex for the most part and .001" is the tightest we get on tolerance. My first target is a computer so I can actually use the offline seat dongle... don't get me revved on that please

[bherrin]

More Info:
Cylinder is 6" tall and was previously measured at 5.8742" in diam before a worker bee decided to dress out a nick with extreme prejudice. It now has an area about 1.5" x 1.75" about 1/2 an inch down that reads 5.8562 on a micrometer. I work in an MRO shop and don't have access to any cad or drawings, only an overhaul manual list minimum diameters. 5.855" and up means an expensive and lengthy repair at an outside vendor (including an angry customer and a write up for aforementioned worker bee), below means we replace a much more expensive and incredibly hard to source part on out own dime ( and probably terminate the worker bee).

And apparently my version of PC_DMIS really like to freeze when i try to switch out the measurement strategy.

A min inscribed circ left with a diam of 5.8736" which is obviously not right, though i think that's more due to missing the low spots on the scan; trying distance to cyl now

[Douglas]

use maximum inscribed and it is the biggest circle that can fit inside data points, this case the opposite side of the part to the lowest wear. subtract that from the original diameter and you have the depth of flaw... as well as your data points capture it

edit... thinking about it that would move the center point a little... stick to maximum circumscribed to minimize that

[Peter Fuller]

If it is a localized worn spot, like, in the middle ONLY of the cylinder, as in, full cylinder above & below, measure the good parts of the cylinder, align to it, then do a scan around the worn area, get the minimum distance from cylinder CL to the scan, that is the smallest radius you can grind it down to.

This is the basic method I would use to address this issue. Except if the rest of the cylinder was in decent shape I would be taking my measurements from the ID for primary cylinder feature if ID/OD relationship is a critical factor but regardless distance to center line is going to be what is critical. I would expect applying fit math to circles would produce a number that wouldn't get me to the right spot the first time.

[Douglas]

my assumption was that the center point is not to move during a regrind or rework as if they will be dialing on that surface on the machine during rework. I think he does mean it is a freshly machined OD that got a spot of damage and somebody polished it out like a hack, so the remainder of the cylinder is good

[NinjaBadger]

Douglas has hit upon the question I was going to ask.

What's most important here....?

1) That the diameter is as large as possible (but the centre can move)

2) That the diameter is as large as possible (but the centre has to remain where it is)

[bherrin]

The minimum distance to cyl cl gave an answer that was consistent with everything else, so ended up using that. If I had more time with the part I would have loved to play around with some of the other options here, but the production people barely wait for inspection as is