Laser Tracker - Measuring a Cylinder

Jake E CEI · Oct 12, 2018

Hi All!

I recently started a new position where I will be using a Laser Tracker to measure parts for dimensional conformity. Most of my previous experience is with CMMs and other types of layout gages (i.e. Optical comparators, Contour tracers, Profilometers, and most common hand gages) as well as I am moving from measuring car parts and small structural pieces of steel, to measuring a box a little larger than a shipping container.

I need to measure a set of Cylinders (3.5" Diameter, and roughly 3' tall) that will be set in concrete to anchor the large boxes we will be installing, the concrete floor would be equivalent to Datum A and I would like to inspect them for perpendicularity as all previous measurements performed on the cylinders prior to myself coming on board to this new company have been an X,Y,Z coordinates from the top center of the cylinders. While this info has been beneficial to my company for leveling purposes we are still having issues with bolt holes not lining up and my first thought was Perpendicularity being off.

My concern is trying to measure the posts with a laser tracker and trying to collect as many valid points as possible with out breaking contact between the SMR and Tracker unit. I do have one of the adapter for the SMRs with a center post on the bottom of the adapter and I will probably start there.

Just curious if anyone else has had to measure anything like this in the past and if you have how do you go about it. I am thinking I may need to get a custom adapter for the SMR.

Milland · Oct 12, 2018

Not to dissuade you, but it seems that if you can use a long steel tape measure with equal tension on the end (a fish scale will do it), then you should be able to get bolt centerlines measured quickly and easily. At worst you might need to calculate the droop and remove that catenary error.

The four rectangular centers and two diagonal should be enough. If needed, a precision level on top of the cylinder gives you angular errors, which if excessive can damage anchoring bolts.

Jake E CEI · Oct 12, 2018

Milland said:
Not to dissuade you, but it seems that if you can use a long steel tape measure with equal tension on the end (a fish scale will do it), then you should be able to get bolt centerlines measured quickly and easily. At worst you might need to calculate the droop and remove that catenary error.

The four rectangular centers and two diagonal should be enough. If needed, a precision level on top of the cylinder gives you angular errors, which if excessive can damage anchoring bolts.

I was actually thinking about trying a method like that yesterday, glad to see I wasn't crazy in thinking that would at least get me in the ball park.

I will have to measure the positions of the posts with the tracker anyway, so at the very least I can do both methods and see what kind of correlation I get and go from there.

Thanks again!

homanfab · Oct 12, 2018

I would suggest measuring several points on the floor (a few next to each column) to create your datum. Then sweep the base of each column at the floor, you should be able to sweep 90-180° of each column. Next I would trace the top edge of each column.

The laser tracker software then should be able to fit a plane to the points collected on the floor and then fit circles to the sweeps at the top and bottom of the columns. Fitting a line between the center of the bottom and top circle of each column would then allow you to determine perpendicularity to the floor and parallelism to each other.

To sweep the floor I would use one of the flat round adapters and to sweep the top edge us the adapter designed to center the target on an edge. The diameters of the circles traced at the top and bottom of the circles do not need to be the same diameter so you changing the adapter won't affect the measurement. All of the lasers I have used have included several adapters.

You can typically select points either using a continuous scan where it collects points when the target moves a selected distance or discrete points. I believe you reduce the accuracy using the continuous scan but it makes it very easy to trace edges.

Another good idea is mark the location of the points measured on the floor. If you need to move the laser or set it back up, you can remeasure those points to transform any new measurements back to the original measurement location. That eliminates the need to remeasure everything and provides a way to recover your setup if someone bumps the laser.

Something like this for the floor:

Edge finder:

adama · Oct 12, 2018

Can you simply securly clamp a V block to the bottom of a SMR? Like this a few hits around each post at the bottom it should then be dead easy to calculate center position to what ever accuracy your systems good too.

Jake E CEI · Oct 12, 2018

homanfab said:
I would suggest measuring several points on the floor (a few next to each column) to create your datum. Then sweep the base of each column at the floor, you should be able to sweep 90-180° of each column. Next I would trace the top edge of each column.

The laser tracker software then should be able to fit a plane to the points collected on the floor and then fit circles to the sweeps at the top and bottom of the columns. Fitting a line between the center of the bottom and top circle of each column would then allow you to determine perpendicularity to the floor and parallelism to each other.

To sweep the floor I would use one of the flat round adapters and to sweep the top edge us the adapter designed to center the target on an edge. The diameters of the circles traced at the top and bottom of the circles do not need to be the same diameter so you changing the adapter won't affect the measurement. All of the lasers I have used have included several adapters.

You can typically select points either using a continuous scan where it collects points when the target moves a selected distance or discrete points. I believe you reduce the accuracy using the continuous scan but it makes it very easy to trace edges.

Another good idea is mark the location of the points measured on the floor. If you need to move the laser or set it back up, you can remeasure those points to transform any new measurements back to the original measurement location. That eliminates the need to remeasure everything and provides a way to recover your setup if someone bumps the laser.

Something like this for the floor:
View attachment 239906

Edge finder:
View attachment 239907

This was my original game plan, my only concern was only being able to capture around 180 degrees at a time without moving the tracker, making the time to capture a complete circle quite large. I am not a fan of using partial arc measurements to build complete circles, I have had horrible luck with repeatability and accuracy in the past.

Have you used the sphere measurement tool? It looks like a long dowel pin with a conical end, I was thinking of using that to give me a little clearance above the top of the cylinder and doing a free hand helical scan of the cylinder.

Jake E CEI · Oct 12, 2018

adama said:
Can you simply securly clamp a V block to the bottom of a SMR? Like this a few hits around each post at the bottom it should then be dead easy to calculate center position to what ever accuracy your systems good too.

This might be the route I end up going if I can't correlate another method, I was curious to see if anyone had used an extension or adapter of some kind to give the SMR some clearance and be able to try and capture a full 360 degrees of the cylinder without moving the tracker location. After reading through some of the replies and thinking about it, I am not sure it is possible to do without moving the tracker itself. But using a V block would give me a consistent reading around the cylinder.

adama · Oct 12, 2018

I have never got to play with a laser tracker, so i don't know how they correlate, but i do have experiance with a faro arm, to a degree the both do the same, record points in 3d space, if you can have 3 or more points that are all the same distance of the middle of the object you can easily calculate or just draw in if you have it as a cad file the circle that fits best and hence you have your accurate centers.

Yes presumably you can not reach behind the object, but if you can get points more than 180 degrees around it on a known radius from it that really should not matter. That said, i would have though the laser has a way of jumping places, with a faro arm, you simply hit 3 or more known points, move and then rehit them and hay presto you know your new position from the known objects and your off again. Think the most jumps i did with the 6' faro arm on one job was over 10, not so much reach, but to get around inside the cabin of a piece of machinery.

John Garner · Oct 12, 2018

Jake E CEI --

Where are the won't-line-up bolt holes, how many are there, and what are the nominal clearances? How constant is the temperature?

Are all cylindrical posts set in the same several-feet-thick concrete monolith?

How well controlled is the geometry of the mating hole-pattern in the to-be-attached pieces?

John

Jake E CEI · Oct 15, 2018

John Garner said:
Jake E CEI --

Where are the won't-line-up bolt holes, how many are there, and what are the nominal clearances? How constant is the temperature?

Are all cylindrical posts set in the same several-feet-thick concrete monolith?

How well controlled is the geometry of the mating hole-pattern in the to-be-attached pieces?

John

The bolt holes are located about 2 and half feet above the concrete slab, the cylinders are spaced apart by 14" in a grid of 5 rows of 19 cylinders. Bolt holes themselves are controlled with a -.01" to +0.0 tolerance on diametrical size. All cylinders are set in the same concrete slab. Temperature is not at all constant as the posts are outside, unfortunately. And the mating piece has the same diametrical size control with only a true position call out of +/- .03" to datums C,A, and B of the mating part. And as far as I can tell we have not been measuring the true position of these holes before I came on board. only the reference dimensions (linear distances) that would control the true position of the bolt holes.

My thoughts are that we are not at all consistent on the perpendicularity of the columns or the bolt holes themselves on either the cylinders or the mating part. and also we vary to much on the linear distances from bolt hole to bolt hole on both parts, which would only exacerbate the issue.

I have been brought on board to increase our measurement capabilities and I can see the issues we are having, I was just curious as to what route may be faster and or more reliable to measure the cylinders with, due to being in such an uncontrolled environment and using a laser tracker for the first time vs a traditional CMM as I am more accustomed too.

hanermo · Oct 15, 2018

Forget the laser. Imo.

Suggest:
Make basic steel gages, with one adjustable length (short) using a very fine pitch screw.
Using gage blocks, a gage rod, and a surface plate adjust the gages to match exact lengths needed.

2 pieces of 15 mm D drill rod, glued to cross members (epoxy), can make very rigid light cheap easy bodies.
Repeatability should be well under 0.001" or 0.02 mm.
Epoxy ball bearing balls into holes slightly undersize for single point contacts as needed.

Use gage pins or alignment pins, pack of 50, as needed for locations desired.
Adjust on gage block set, into oversize holes, epoxy pins in place.
Then measure and write any errors onto a tag next to pin.

All above costs 200$ or less for 3-4 gages and takes maybe 3 hours to do.
For 200$ more you can use linear ball-bearing guides, a tripod-parallel-rod design, and 2 each fixed-fixed endplates at both ends for adjustable 3 foot micrometers.

For 1000$ more I would make them for You, set of 3.
For 500$ more each you can get a glass scale with 5 micron or 1 micron resolution accurate to a few 2-4 microns, and a readout.

John Garner · Oct 15, 2018

Jake E CEI --

I offer my sympathy, having spent a lot of my career trying to make precise assemblies out of poorly-thought-out loosey-goosey parts.

At this point, I don't see any particular value in making Laser Tracker measurements of the as-is configuration. You already know that the should-fit parts don't fit correctly, making "What now?" a legitimate question.

The answer to that question could be a simple as opening the holes, or as complex as making every post into a two-or-more piece assembly with adjust-as-necessary pieces at their tops.

When it comes time to adjust those tops, the Laser Tracker and its software become a very useful measuring tool.

John

Jake E CEI · Oct 15, 2018

John Garner said:
Jake E CEI --

I offer my sympathy, having spent a lot of my career trying to make precise assemblies out of poorly-thought-out loosey-goosey parts.

At this point, I don't see any particular value in making Laser Tracker measurements of the as-is configuration. You already know that the should-fit parts don't fit correctly, making "What now?" a legitimate question.

The answer to that question could be a simple as opening the holes, or as complex as making every post into a two-or-more piece assembly with adjust-as-necessary pieces at their tops.

When it comes time to adjust those tops, the Laser Tracker and its software become a very useful measuring tool.

John

My company at least seems to understand the situation they are in and more than likely we will be using the tracker variable data do dial things in as close as we can for fit. The current fix has been to use a smaller bolt on the ones that are not lining up properly, but we would like to be better if we can.

Thankfully the company that I am with is just starting to get into manufacturing (new division of a electrical contracting firm) and understand that there are going to be lots of growing pains. And I look forward to the challenge of trying to get us into the position we need to be as far as Quality and Metrology are concerned.

I have seen some new videos of Faro's newest add on for the laser tracker, the Vantage S6 with the 6Probe. This looks to combine the mobility and range of a traditional Laser Tracker with the probing abilities of a Arm. I may have to try and schedule a demo to see if this will be a better fit for us, although I may get some pushback from management as our current tracker is less than a year old.

Primo1 · Nov 17, 2018

I have measured many cylinders with a laser tracker, most holes have a depth and therefore a cylinder by default. The most reliable and repeatable way to gather information is to take measurements equality spaced around the circumference as well as the top, middle and bottom of the cyl. Also, a least squares fitting of the cylinder will give the best position as well as perpendicularity, not dia. Hope this helps

Search

Largest Manufacturing Technology Community on the Web

Search

Laser Tracker - Measuring a Cylinder

Jake E CEI

Plastic

Milland

Diamond

Jake E CEI

Plastic

homanfab

Aluminum

adama

Diamond

Jake E CEI

Plastic

Jake E CEI

Plastic

adama

Diamond

John Garner

Titanium

Jake E CEI

Plastic

hanermo

Titanium

John Garner

Titanium

Jake E CEI

Plastic

Primo1

Plastic

Similar threads