Thread: A question for the metrologicly (??) minded...

I wish to bore out a part section (target diameter is around 5"/ 127mm) but can't think how to best measure what the diameter/ radius is. This trough will not be uniform but will have some diameter changes so full form gauging would be problematic - although not impossible
This is not it, but an example of the type of profile I'm looking to measure. Measurement will need to be in situ so that I can adjust the boring head to be close to target (less than a mm)
P1020702 (Medium).JPG

The best method I've thought of is a couple of knife edges on a known length of beam with a dial gauge in the middle measuring the amount the surface deviates from flat. A bit of maths and I could calculate the radius
radius measure.jpg
The downside is that to get reasonable accuracy the indication must be good to micron levels (I have one to 2 micron), which with a machined surface may be asking a bit much. (0.002mm steps in d where 2l= 50mm gives steps in radius of 0.045mm. An error then of 0.01mm means the radius could be around 1mm off. Reducing 2l makes things better but the gain on any error is still large)
Any other ideas out there?

Michael

Heh, well this amateur machinist will let the experts have a say before I offer some ideas.

If yer looking for an excuse to buy some new gages I see in the Insize cataloge they have a Digital Radius Gage able to read internal and external arcs of a radius .2 to 27.5", and with a resolution of 0.00005"

If you have a DRO on your mill you could touch probe the arc at 3 points and calculate the radius. I'll check some of my metrology books tomorrow but i don't think i will find much.

Originally Posted by YBAF

If yer looking for an excuse to buy some new gages I see in the Insize cataloge they have a Digital Radius Gage able to read internal and external arcs of a radius .2 to 27.5", and with a resolution of 0.00005"

Finding a price for one of these was tricky, but at 280 pounds (that is, over A$500), a bit of a stretch for one job. Interesting to see they are using the same principle though - I may have to have a closer look.

Originally Posted by snapatap

If you have a DRO on your mill you could touch probe the arc at 3 points and calculate the radius.

No touch probe (a Taster is the closest I come) but it would mean disturbing the set up to measure which could be a problem.

Michael

does it have to be bored with a boring head?

Could it be line bored in a horizontal mill?

When you do line boring it is easy to measure the diameter of your cut from the tool stickout. The cutting tool is held in the tool holder which is just round bar which rotates around centre. So you know the diameter of the bar, and you measure from the tool tip to the other side of the bar, minus off half the bar diameter and that leaves you with the radius of the cut. That is how

Originally Posted by Michael G

The downside is that to get reasonable accuracy the indication must be good to micron levels (I have one to 2 micron), which with a machined surface may be asking a bit much. (0.002mm steps in d where 2l= 50mm gives steps in radius of 0.045mm. An error then of 0.01mm means the radius could be around 1mm off. Reducing 2l makes things better but the gain on any error is still large)

Michael

G'day Michael, I'm having a bit of trouble following both your math, and also your conclusion regarding chord length. What you're proposing is a standard way to measure telescope mirrors (except it's normally done with a micrometer instead of an indicator, but that makes absolutely no difference). Look at a "Spherometer" for more information. In this case it's much easier as it's not a sphere. I believe you will get increasing accuracy with greater chord length, to the point where once it becomes 2R, you are directly measuring the radius and there will be a 1:1 relationship to what you've labelled d. So in other words I would increase the distance between your two pins/balls/blades/whatever contact points you choose, to the maximum length available given the length of your arc as the error will be a square root of the difference between your chosen lengths.

There's also no need to for chord to be a straight line, and you could have a smaller arc for your base. The only critical factor being that it bears on the contact points separated by a known distance.

Edit: Incidentally, you have me a little curious now as to why you're making this measurement in the first place. Are you wanting to measure and replicate an existing arc? As I understand it, you will be boring this yourself, and, while I don't want to tell you how to suck eggs, that's not the way a boring head would normally be set to bore an arc, so I'd be keen to learn more.

Had a bit more of a think and wondered whether rollers would help -
roller rad.jpg
Pete, I don't know what my maths is doing either it seems. I put some values into a spreadsheet and the numbers I gave above are what came out. Same scenario, looks much better in terms of discrimination.
roller rad2.jpg

I only have around a third of an arc, so can't use a chord of 2R.

Michael

Originally Posted by YBAF

Heh, well this amateur machinist will let the experts have a say before I offer some ideas.

me too...i cant even add up sometimes.......let alone do trig

Originally Posted by .RC.

does it have to be bored with a boring head?

Could it be line bored in a horizontal mill?

Funny, I was thinking how doing that seems to be a piece of pyss.....

..... if you have a HBM!

PDW

G'day Michael, no I didn't mean to imply that you need 2R, instead what I meant is that the closer you get to that, the less the proportional error. You suggested a very short chord, and if I understood correctly, implied a shorter chord was better. In fact I'd suggest the opposite is the case, and you want your measuring chord to be as long as possible. I ran some numbers on even a 100 mm chord last night (which I think is ver short), I was very tired but they looked ok to me in terms of resolution.

I bore arcs relatively often I guess, indeed there was at least one that I remember doing in that little rod bender I put up here a couple of weeks ago. I did some others about the same time, but can't recall now what they were in. There's probably various ways to do it, but the way I do it is to pick up the edge and zero my DRO on that edge, move the X over 127 mm, insert the boring head and swing the tip past the edge while adjusting out. When the tip of the boring head just touches the work, the R is set to exactly 127 mm. You then move the head back toward the work to set your centre distance. I do recall with that rod bender both the X and Y were offset a specific amount, so you need to pick up both datums before commencing. You don't need to pick up on the work itself to set the R, and anything will do, that is purely just to set the boring head so you can set that on any scrap and them remove both for final fixturing of the part, pick up locations as required and then replace your pre-set head. Conventional bores can be set the same way, and I'd guess the accuracy would be down to hundreds of a millimetre if set with care.

To change radius, you don't need to go through the whole process again. Instead mount and swing the boring head past an elephant's foot tip on a dial indicator (ie so the indicator is laying horizontal to the table). Move the boring head in and out the required amount as read off the dial indicator. This saves having to reply on the markings on the boring head, backlash in the adjustment etc etc.

Hope that makes sense and helps.

Hi Michael.

Originally Posted by Michael G

Had a bit more of a think and wondered whether rollers would help

I'll leave the maths for this idea up to you, but would having the center roller bigger so you end up flat across all three rollers when you are on size help? then you arent trying to measure a "size" you just want "0"(of course go to far and you're on your own )

Originally Posted by Michael G

I only have around a third of an arc, so can't use a chord of 2R.

I think D*0.71 is the chord sweet spot. I make that about midway between your two sources of error.

Stuart

Originally Posted by Michael G

Had a bit more of a think and wondered whether rollers would help -

Hmmm... would you get by with just one pin gauge and a slip block across the top. Thats two sides of the triangle measured...

Gauge blocks do not have sharp ends, they are chamfered, so you will not get an accurate measurement.

Originally Posted by .RC.

Gauge blocks do not have sharp ends, they are chamfered, so you will not get an accurate measurement.

And heres me with a set siting on the table in front of me - Powers of observation, fail.
...and hense the Michael G three pin working.

Originally Posted by Pete F

...but the way I do it is to pick up the edge and zero my DRO on that edge, move the X over 127 mm, insert the boring head and swing the tip past the edge while adjusting out. When the tip of the boring head just touches the work, the R is set to exactly 127 mm.

Makes sense and I'll probably use a variation of that method but this will be a casting so I will be battling draft (DOC increases) and machining allowances as well as uncertainties from the casting process. Being able to measure the radius provides a warm fuzzy - shows me I haven't gone too deep or over size. What would be nice would be to be able to do a clean up cut to get a uniform surface and then knowing the radius of that, offset from there

Originally Posted by Stustoys

... would having the center roller bigger so you end up flat across all three rollers when you are on size help?

Don't think so (although happy to be shown otherwise). With three rollers the same diameter it is a general method. If you tried to get the tops 'flat' you would have to have a special roller for each size you were aiming for. Need to do the trig...

Originally Posted by PDW

Funny, I was thinking how doing that seems to be a piece of pyss.....

..... if you have a HBM!

Yes, the boring part would be...

... but I would still need to measure it!

Michael