Radius checking

[Perry Fisher]

I'm checking several appliance parts that have very large radii to the tune of 360 inches. I only have 4° of radius available. I know this is no where near enough for PC-DMIS to calculate accurately.

Suggestions are welcome on how to check these radii.

[dwade]

My first thought was get the drawing changed to a profile tolerance for the surface. Basics then scan it or individual points with T values.

Short arcs are never repeatable no matter what size radius is unless you can use full contact scan and then with 4° or arc length you are still way beyond the capabilities of most if not all CMMs

[Wes Cisco]

Well you could get some poster board and make a very large compass from some 2x4s. Set the compass with a scale or yo-yo (tape measure), then mark an arc on the poster board, cut on the the line carefully and wa-la one Southern Enginired Radii Template. From there you might have to use a casting and compare the casting to the template on a comparator.

Crude? Oh Yes, but no less ridiculous than trying to measure such a large radius on a CMM with such a small arc available.

Last resort is to always contact the customer/engineer and ask how they want this checked. HTH

[bw_bob]

I'm checking several appliance parts that have very large radii to the tune of 360 inches. I only have 4° of radius available. I know this is no where near enough for PC-DMIS to calculate accurately.

Suggestions are welcome on how to check these radii.

Get the C/L of the radius and move your coordinates there and check it that way.............

[#2#]

R360 huh? What is the tolerance on that? I'm with Wes, contact the customer, tell them you can't check it and see what they say.

[Ravenmaniac]

Are there any straight edges on the parts that you can align to that can get you back to the theoretical origin of the radial swing? We measure large arc parts here and I usually have to do it this way: After leveling and what not, I find some straight edge that can orient the machine axes. Hopefully for you, there may be one that cuts thru your radial arc on the same radius chord as the origin of the radius. I measure a line along that edge and then carefully take 5-6 points on the circle. Don't worry about where the **** circle is for now. The key is now to create a pierce point where the line cuts thru the arc and translate back theoretically along your line to the circle's origin using an exct radius offset. Then you set your x-y zero. Now, if you have a dcc machine, all the better, program your tip to take polar surface points along the arc in small degree increments. If doing this in dcc, the machine will even make the perfect vectors hits perpendicular to the point of tangency. Trying to evaluate a large subtended arc with only a few degrees to work with is almost impossible. If you take say points along the arc and evaluate it as a circle and one of those points is off from perfect nominal by about .0005, you could have a .06-.100 location and size error generated if your radius is around 30 inches. It's kinda like trying to sight in a new rifile and starting at 300 yards. You'll never hit the paper. The chances of trig error are too great. Hope this helps!

[Cumofo]

Use profile and get them to change the print.




G

[Perry Fisher]

dwade... Getting the print changed is next to impossible. The engineers for this customer are the smartest, most intelligent engineers on the face of the earth. They have told us that more than once. Checking as a profile or using the "T" values to calculate the radii might work.

Wes...I have a comparator large enough to get most of the radii. I have thought of making overlays if I had some mylar to plot on. If you have some send me a couple sheets. (kidding)

Bob... I have them programmed from the CL. Maybe use the CL combined with dwade's suggestion of "T" values?

Goodluck... ±0.02" tolerance. I could put something on the report and tell them the parts are to print because they will never check them. If the stove assembles is all they care about. Trouble is I will not supply data I can't back up.

[#2#]

dwade... Getting the print changed is next to impossible. The engineers for this customer are the smartest, most intelligent engineers on the face of the earth. They have told us that more than once. Checking as a profile or using the "T" values to calculate the radii might work.

Wes...I have a comparator large enough to get most of the radii. I have thought of making overlays if I had some mylar to plot on. If you have some send me a couple sheets. (kidding)

Bob... I have them programmed from the CL. Maybe use the CL combined with dwade's suggestion of "T" values?

Goodluck... ±0.02" tolerance. I could put something on the report and tell them the parts are to print because they will never check them. If the stove assembles is all they care about. Trouble is I will not supply data I can't back up.

±.02" on a 30 foot radius!!!??? For a stove!!!???

What, does this stove go on the space shuttle or something?

Wes's suggestion got me thinking. I know from time to time when dealing with gears here that the engineering department is able to scale the gear in CAD and print it out to full scale on a piece of film (like for overhead projectors). What I would do is see if you can get this done at the min. and max. Have them put a 1" square box on the film somewhere too. That way you can check the 1" square to 'calibrate' the film and make sure it printed actual size. Don't report the actual radius but rather just do a go-nogo check on the comparator that the radius is within tolerance and report it that way.

[bw_bob]

dwade... Getting the print changed is next to impossible. The engineers for this customer are the smartest, most intelligent engineers on the face of the earth. They have told us that more than once. Checking as a profile or using the "T" values to calculate the radii might work.

Wes...I have a comparator large enough to get most of the radii. I have thought of making overlays if I had some mylar to plot on. If you have some send me a couple sheets. (kidding)

Bob... I have them programmed from the CL. Maybe use the CL combined with dwade's suggestion of "T" values?

Goodluck... ±0.02" tolerance. I could put something on the report and tell them the parts are to print because they will never check them. If the stove assembles is all they care about. Trouble is I will not supply data I can't back up.

Are they basics to the C/L......if so just rotate about the axis and check points for the radius...

[Perry Fisher]

Not exactly for the space shuttle program. Look here. They are making these with large radiused fronts.

If I check as a profile what would I use for the deviation value with ±0.02 tolerance? 0.04?

[Wes Cisco]

dwade... Getting the print changed is next to impossible. The engineers for this customer are the smartest, most intelligent engineers on the face of the earth. They have told us that more than once. Checking as a profile or using the "T" values to calculate the radii might work.

Wes...I have a comparator large enough to get most of the radii. I have thought of making overlays if I had some mylar to plot on. If you have some send me a couple sheets. (kidding)

Bob... I have them programmed from the CL. Maybe use the CL combined with dwade's suggestion of "T" values?

Goodluck... ±0.02" tolerance. I could put something on the report and tell them the parts are to print because they will never check them. If the stove assembles is all they care about. Trouble is I will not supply data I can't back up.

Drat those ethics, always getting in the way of profit! I have the same problem Perry, but fortunately I am allowed some latitude by my companie's quality policy. I can report "visually ok" or just "ok" for something like this because it is not +/-.005" or less. I like Goodluck's idea of using CAD to print this out with a 1" background grid or cube for verification. This could even be done on plain paper and cut out, just as a visual comparison to make sure you are not a mile off.

The kid in me would still like to make the 2x4 compass just for the silliness of it!

[Darin Steiner]

Displace origin to c/l of radius then turn polar vector compensation on using the appropriate measuring plane. This will give the location of your hits relative to the origin of your cordinate system compensating the probe radius along a vector to the current origin. I use it all the time.

[Perry Fisher]

Displace origin to c/l of radius then turn polar vector compensation on using the appropriate measuring plane. This will give the location of your hits relative to the origin of your cordinate system compensating the probe radius along a vector to the current origin. I use it all the time.

OK now I'm as lost as last year's Easter egg. Can you explain with a little more detail?

Thanks

[Gregl]

OK now I'm as lost as last year's Easter egg. Can you explain with a little more detail?

Thanks

make the center of where the 30 ft radius your X0 Y0 and take hits along the radius and see how far off the hits are