What tool do I need? I need to measure the radius of a tiny fillet on a 90° edge.

PS Wait, I think this would be better quality
"Digital Chamfer Gauge - 1180" - made by Insize, for c. £90+VAT looks like better a build quality:

**LINK**
https://www.cutwel.co.uk/0-10mm-0-0-39-45-digital-chamfer-gauge-1180-series-insize

With an accuracy stated of 0.06mm you could be getting some readings that are not that accurate particularly if you factor in any errors in the maths converting to a radius due to rounding up/down. That is also assuming your deburring wheel is forming a perfect 1/4 circle in the first place.

Edited By JasonB on 04/01/2022 07:50:53

Yes, good spot – "0.06mm" IS pretty disappointing. That said, that is presumably the inaccuracy over the full 10mm, so if one has recently re-zeroed the device, hopefully it will be much more accurate at the smaller distances.

No I would say it is the reading will be +- 0.06 anywhere along the scale.

It's worse than most of the other callipers that they do but that is likely to be the fact the chamfer will be 1.404 times longer than the scale moves so any error in the scale and therefore reading is magnified by this factor

Yes, I agree that the chamfer will increase the inaccuracy compared to straight-forward digital calipers (with parallel measurement surfaces...  but beyond that I don't follow your logic.

Srely at least some of the stated inaccuracy must come be incurred during the 10mm of travel. 
Presumably the internal digital measurement tech to be standard, off-the-shelf parts.

I mean if I zero the device and move it just 0.10mm, what do you think the range of readings will be?
 

Edited By John Smith 47 on 04/01/2022 12:18:56

0.1+-0.06 most likely

So are you assuming that the measurement of the distance of travel is 100% accurate and that 100% of the errors are due to the play of 0.06mm?

i.e. So when freshly zero-ed, as soon as you move it, and move it back to zero that it would is still be reading somewhere between +0.06 and -0.06mm?

Seems a little excessive... 

Edited By John Smith 47 on 04/01/2022 12:33:59

To answer the question properly you need to do a gage r and r study.

Typically an accuracy spec which doesn’t include a %of length term is an absolute spec - I.e. it will be within the range no matter the length moved.

Accurcy and repeatability are not the same thing.

So zero, move 0.1mm and move back 0.1mm would I expect report 0, but absent any spec for repeatability it could be +-0.06 and not be out of spec - so you would have no grounds for complaining if it did show 0.05 for instance

It will read in the range 0.04 to 0.16

Probably inappropriate to measure a 0.1mm radius with instruments of this type because their error is a significant proportion of the radius. The instruments are more useful on larger diameters - ±0.06 is an error of 1.2% on a 5mm radius, rising to 60% on a 0.1mm radius.

How accurate does your measurement need to be?

Which of these examples are unacceptable and why? (Comments apply to both inner and outer chamfers)

For comparison, bottom left is a perfect radius. These are nearly impossible to make by hand, and although a machine will do better, cost rises steeply with increasing precision. There would have to be a very good reason for designing an accurately radiused chamfer, and most production engineers would look hard for an alternative.

Next up is a faulty radius; it's offset by 10%. At R0.1mm, the measuring instrument would have to be good for 0.01mm, which industry would normally say requires calibration ten times better, i.e. ±0.001mm. That's beyond any of the caliper instruments so far mentioned in this thread. And, is this faulty radius really a problem? For many purposes, it's near perfect. If so, we've spent a fortune on measuring the wrong thing.

How about the radius with variable error? My example is typical of a machine made radius, hand-made would be worse. It's close, but a suitable instrument would detect radial and axial imperfections. Again, is this curve unacceptable in the real world?

Lastly, is a radius approximated by 10 straight lines. Again, a comparator will detect it's wrong, and by how much, but why does it matter?

I don't measure radii often enough to justify buying a special caliper or gauges. And when I do measure radii, so far there's been no reason for the answer to be particularly accurate. As always with unusual requirements, it's best to question them rigorously. In engineering, the most expensive mistakes are caused by faulty requirements!

Dave

Edited By SillyOldDuffer on 04/01/2022 13:13:05

Simply by the amount of pressure you apply to the sliding head you will get a slightly different reading this is why callipers are not regarded as being as accurate as micrometers and even they will vary depending on your "feel" when using them.

Take a look through the Insize cataloge, I have the paper one that is about 1/2" thick all the standard shape callipers are +-0.03 to 0.04 accuracy, the ones with long reach legs, funny ends ext less accurate. Now look at the micrometers and they show accuracy in microns generally +- 2 or which is 0.002 or 0.003mm. You will see similar figures if looking at Mitutoyo. I'm sure these Insize ones would be ideal for checking weld prep chamfers and the like.

They have some very good optical measuring equipment in the catalogue that would be ideal if a bit out of price range.

Edited By JasonB on 04/01/2022 13:13:18

Dave S - I'm still not with this.

> I.e. it will be within the range no matter the length moved.
Yes, but in practice the accuracy of readings will entirely depend on where the inaccuracies are coming from.

In my experience, if I zero my current digital calipers, and move them to 10mm and back, I would still expect it to read either +0.01, 0.01 or -0.01. Yes, after a large number of measurements it is more likely to no longer still say 0.00, (albeit in my experience not by more than 0.01 or so).

i.e. In my experience opening and closing the calipers enough times will eventually create new errors that become obvious when the device closed and no longer reads "0.00mm".
And I have been assuming that the manufacturers would make their accuracy claims that assume a reasonable amount of use after each re-zeroing by the user.

But if opening and closing the calipers multiple times since the last re-zeroing does NOT create new errors, I find that surprising.

Either way, in this case are we seriously expecting that if I zero the device and move it by exactly 0.1mm that it will immediately read as much as "0.16mm" or as little as "0.04mm"?

(And by implication if that if I move it to 9.1mm that it would ready as much as 9.16mm or as little as 9.04mm?)

Personally my best guess is that error come from two effects
A) Errors that are proportional to the distance away from zero,
B) Errors that are proportional to the total distance moved up and down, up and down.

If either (or both) of the above are true, then the accuracy of measurement would be much increased by a recent zeroing followed by only a small amount of opening/closing travel.
 

Edited By John Smith 47 on 04/01/2022 13:20:31

Edited By John Smith 47 on 04/01/2022 13:24:03

@SillyOldDuffer

OK I think we've done this subject to death!

Speaking pragmatically, I currently have no way of ANY sort to measure any of my fillet radiuses. And there don't appear to be any tools that can help much other than a chamfer measuring tool which will of course be making the assumption that the fillet is extremely circular (which it may not be).

And so even if a reading of "0.04mm" might genuinely mean "0.16mm" maybe that is sill better that absolutely nothing.

'CIRCULARITY' of radius
Getting at least SOME sense of the cross-sectional shape of the fillets that I am grinding using a deburring wheel would be also nice... but maybe that is just too difficult.

I mean, it would be very nice to have some sort of device to help me see roughly what shape (i.e. roughly how circular) the hand-made fillets are, because when they are made by machine, presumably they will be made more accurately and certainly more consistently.

For that I was hoping to get some of those Starrett things I mentioned on page 1, and shine a light behind them to broadly see what sort of profiles my deburring wheels are creating.

OK I think I'll just buy that Insize 45 degree chamfer measuring gauge and be damned.
(But although it is only a working hypothesis, I still strongly suspect that in practice, it will be significantly more accurate than "+/- 0.06mm" when used for small chamfers. )

J

PS If at some point I can be bothered, I could use the other blades to measure a fillet at 30 degree and 60 degrees as well as at 45 degrees and run the trigonometry to see roughly how circular my fillets are.

although it may have said 10.00 when you moved it the jaws may have been a 9.94 to 10.06.

Same if you zeroed it and then put the jaws on an exact 5.000mm dia you may get a reading of 4.994 to 5.06

So lets look at your chamfer. zero and then take a measuerment and the screen shows 0.25 however the chamfer could actually be anywhere from 0.019 to 0.031.

Now convert that to a a radius, do you use 3.14, 3.142etc for pi

From the above two lines the actual reading you get could be either of your 0.2 or 0.3 radii increments or anything in between so if you want to be measuring and quoting by small amounts you need an apropriatly accurate tool. It's generally said your tools need to be 10times more accurate than the part.

I have a set of gauges that Peak4 linked to in his answer. They seem to be accurate enough for what I need. At 0.3 the radius is tiny and needs a set of magifiers along with my reading specs to be able to actually check the radius. They seem to be ok unless you are working to atomic standards.

regards

To quote John Smith 47

"OK I think we have done this subject to death"

I agree whole heartedly!

@JasonB - that's all find in theory and is obviously within spec. But what assumptions are you making about where and how the errors creep in?

J

PS Fwiw, I absolutely love my AccuMaster Fractional 6" Calipers. I think they are rated at "0.02mm" accuracy but in practice they seem accurate to +/- 0.01mm, which is easily good enough for me. They are so much quicker to use than analogue/Vernier callipers. The only time I have any problems is when I fail to re-zero them after a lot of opening-closing travel, at which point they don't reliably read "0.00" when closed.

They are exceedingly repeatable and the only time I have measured larger objects (e.g. steel parallels) they have been absolutely spot on (if recently re-zeroed).

Edited By John Smith 47 on 04/01/2022 15:13:43

PS I have placed an order for the 45° Digital Chamfer Gauge (1180-6) - made by Insize. Should arrive tomorrow.

One thing to be aware of is that Cutwel's website is lying to us. Each tool can ONLY measure one specific angle, whereas the Cultwel website implied that you could insert blades of 3 different angles, but I spoke to Cutwel and you can't!

I am still in the market for buying two or three simple super-small L-shaped fillet gauges i.e. with a design like the Starrett so that I can check the circularity of the profiles that my deburring wheels (I have 2 different types) are making. (??)

Edited By John Smith 47 on 04/01/2022 15:14:27

Not really a lie, it's a general description of the chamfer type of calliper, if you then click "view full range" it lists the three callipers

Edited By JasonB on 04/01/2022 15:24:56

Probably just me but do I discern a possible typo in all this?

Could tool be misspelt when troll was intended?

I’ll get my coat…….

Martin