reporting circumference

[kass_b]

Have a part 4 diam not realy round need to measure and report circumference of parts. Never had this one best way to measure and report. Parts have split in diam like a saw cut.

[#2#]

So do you need to report the actual surface distance or is reporting the circumference of the measured circle adequate. If you just need the circumference of the measured circle, insert a variable using c=pi*d. I'm not sure about how to use pi in a variable but I bet it can be done.

v1=pi*cir1.d

If you actually need surface distance, I don't know off the top of my head. I'll have to ponder it for a while.

[Matthew D. Hoedeman]

I do not think Pcdmis can give you 'length of line'. Sounds like something that would need to be scanned, continuous contact or laser, then length of line reported from a cad software.

[kass_b]

Thanks I actually need surface distance thought of pi times diam but parts are not really round. And yes Matt was wondering if there was a way for pcdemis to determine this. looking for a good way because there are sixty parts. Was thinking this will have to be done on solid works.

[Wingman]

How many points are you allowed to take?
This is probably going to suck as a solution BUT if you take enough points and distance 3D from point to point and using variables you capture a cumulative sum of the distances it may prove as accurate as solidworks.

[kass_b]

Can take as many as needed. tried constructing generic curve but need way to determine length

[#2#]

How many points are you allowed to take?
This is probably going to suck as a solution BUT if you take enough points and distance 3D from point to point and using variables you capture a cumulative sum of the distances it may prove as accurate as solidworks.

That's not a bad idea. Measure it as a circle, set the circle as the orgin and then create a bunch of polar vector points? Paste with pattern might also be usefull.

[Wingman]

I am talking cumulative distance. Point1 to point2 + point2 to point3 + point3 to point4 etc

[Wingman]

That's not a bad idea. Measure it as a circle, set the circle as the orgin and then create a bunch of polar vector points? Paste with pattern might also be usefull.

Didn't think of paste with pattern but it would make quick work of the points. Hmm

[kass_b]

I understand will try that, job won't be here untill Tuesday, Maybe something else will come up by then.

[Wingman]

Just had some thoughts and wanted to post them before I forgot. You could find 'center' and zero. Find corners of the 'split' and create start and end lines.

Using Variables
Assign Total Angle Angle between Start and end Lines.
Assign Step Angle as Total Angle / Num Hits

Align to Start Line

Count = 1
Circumference=0

If Count < numhits
Rotate Step Angle
Vector point(count) Radius,0
Circumference = circumference+dst(3d) point(count-1) to point(count)
count=count+1
endif

Circumference = circumference+dst(3d) point(numhits) to Point end

See attached pic
Starting point would need renamed to point0
Ending point would need renamed to pointend

enough hits it would work

I realize it isn't a perfect cure but it would get you there. Solid work might do better if it can fill in the spacing with a curve instead of a straight line.

Of course, if you knew the math formula for circumference of a small section of an arc you could compute every three points into an arc and calculate their distance. You can find the formula in Machinery Handbook I am pretty sure.

Would take some code up front but it is doable.

[Wingman]

Sorry, forgot to attach pic

[kass_b]

Wingman, thanks for all the input, Thanks everyone else, I will play around with this but if solid works creates better curve we will use solid works. Will try and remember to let you know how this one turns out.

[John Kingston]

Of course, if you knew the math formula for circumference of a small section of an arc you could compute every three points into an arc and calculate their distance. You can find the formula in Machinery Handbook I am pretty sure.

If your points are equally spaced, this would be easy. Let's say every 10 degrees. 3 hits would cover 20 degrees of arc.
1) Find diameter
2) Circumference = pi*rsquared
3) Circumference/360 = Distance for 1 degree of arc
4) 20*Distance from above = distance for 20 degrees of arc

Now I may be totally misunderstanding what you want, but I think that works.

[Wingman]

John,
That WOULD work. Quick too. Could be done in a simple loop.