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

In which case the answer is to eyeball it, and seeing R0.1mm would need the aid of a loupe. The human eye is good at spotting malformations, especially when the radius can be compared with a gauge with a bright light behind, and then in front of it. Finger tips are remarkably sensitive to tiny irregularities as well.

Human sensory methods don't measure anything in the sense that a machine can be told to remove 0.05mm or whatever to correct a dimension. But they do allow the workman to progressively refine his work. From what you've said, your chamfer can be made as well as is needed without measuring it at all.

Before science and technology merged everything was made by comparing rather than measuring. Because it takes so long to do well, most early products were crudely made, but craftsmen have been capable of making fine mechanisms like clocks for centuries. Also high-art, decoration, jewellery and other posh stuff. Apart from the cost, other serious problems are training craftsmen, and their inability to make parts identical within tolerances. Whilst craft methods are no good for mass-production, they're often excellent in a home workshop.

Likely none of the disadvantages apply to you unless the radius is part of a delicate bearing. Otherwise, the radius can be formed by several strokes of a fine toothed file angled to produce a reasonable curve. If the result looks OK it probably is.

If the subject has been done to death, why are you still asking the same sort of question? From my side of the internet I can't tell if you need a gauge and eye-glass costing £10 or a Shadowgraph like this one, £4,388.60 on ebay.

Dave

Wrong. If you click on the link to the individual product for Sku:"1180-6"
(https://www.cutwel.co.uk/0-10mm-0-0-39-45-digital-chamfer-gauge-1180-series-insize )

You get this:

==> It says "Product Information" and the very first of the bullet says "Measure 30 degree/45 degree/60 degree chamfers".
Note that it it says "Product Information". It does NOT say "Product Range Information"

If you read all of it carefully this is self-evidently a mistake and if you are moving at speed it is easy to be fooled, as I TBH, I very nearly was.

The Cutwel staff agreed with me, were hugely apologetic and said that they will change the text ASAP.

> In which case the answer is to eyeball it, and seeing R0.1mm would need the aid of a loupe.
Yes.

> Finger tips are remarkably sensitive to tiny irregularities as well.
Yes, and that's the whole point.
Small deviations that the human eye can't really see unaided can make or break the user-interface of a product.

> From what you've said, your chamfer can be made as well as is needed without measuring it at all.
A prototype yes, absolutely.
But when specifying a product for someone else to manufacture... absolutely NOT.
If even in small scale production which could initially be done by hand, a loose specification could make or break the entire viability of the product.
Worse, the eventual larger-scale production would of course be done by robots, die pressing, tumbler machines and who knows what.

And really, that's the point, I need to specify the required output not the processes.

Re the Shadowgraph, as I have now said more than one, my some of my fillets are quite long and I have assumed that it would be extremely difficult measure the size of a fillet along the length of a longer fillet on a part. Obviously it would also be well beyond the budget of £100 or so, which I have also mentioned more than once.

Look, I know that all this is rapidly becoming TEDIOUS for most of the community here, but through all this endless questioning of my motives, you are compelling me to explain details that I had no intention of going into. These are details that are of no interest to the majority of this community, for which I get flamed if I ignore any contributor's questionings.

So back to my original question, I was just asking if anyone knew of any tools that would help me measure the radius of very small fillets, at a half-sensible (ideally sub £100) cost.

If you think that the answer is "no", then let's move on.

NO

Robert Butler

Whilst that may (or may not) be true without detailed knowledge of exactly how the thing works it is impossible to say where any inaccuracies start.

Assumptions are dangerous things in metrology.

Not exactly. I am saying that the range of possible readings it could give and be within the spec is 0.04mm to 0.16mm.

Decent calipers almost never need their zero resetting. It more likely to be reset because you are using it for some other purpose than absolute measurement - for instance I have a target sized shaft so I zero on that and now my digital caliper means I dont have to think about how much to take off (apart from the should I divide it by 2 for this lathe cross slide )

Guessing is a sure way to be wrong on average 50% of the time... It is very unlikely that total travel distance makes any difference to the accuracy.

Split post because apparently I typed to much...

Edited By Dave S on 04/01/2022 17:46:56

Second part to my excessivle long post;

The way *most* of these sort of devices work is based on a pair of printed circuit boards and some sensing electronics. The PCBs form a variable capacior, which is 'driven' to give an output that is proportional to the distance moved. There are many sources of errors in that system. Firstly if there is a clock used to drive things then it is probably a quartz crystal, and temperature senstive. There may be step and repeat errors in the PCB manufacture, as the battery discharges there may be an effect in the sensitivity.
All of these things (and many more I haven't thought of) matter if you are truely wanting to measure an very small thing.

usually for an electronic linear measurement device (which the scale in a caliper is) there is a spec like accuracy 100 to 3000 mm -(5+5L/1000) µm
so for a 100mm scale that would be 5um + 5*100/1000um -> overall accuracy 5.5um, and a 1000mm scale it would be 5um + 5*1000/1000 -> 10um.
Because the calipers are short this is not usually listed - it applies more when you are looking at several different lengths of the same thing.
If you look at mitutoyos calipers as they get bigger their accuracy gets less. They use the same technology, so (for instance) the 200mm caliper is 0.01mm resolution, +-0.02mm accuracy, their 300mm caliper is 0.01mm resolution, +-0.03mm accuracy, and their 1000mm caliper is 0.01mm and +-0.07mm.

There should also be a repeatibility spec - how well the scale shows the same number when moved away from and back to the same place (in this discussion the same place means *exactly the same place*,not I moved it there and back by hand). For Mitutoyo calipers this is typically 0.01mm. so with a 200mm caliper you could be out by 0.02mm, but also get a different number by 0.01mm on measuring the exact same thing multiple times.
Quantifying this is part of when it matters it matters. For most things in hobby land it actually never matters

Your anacdotal assertion that your calipers, which are spec'd as +-0.02mm accurate 'seem accurate to +/- 0.01mm' is just that - a seems - unless you have actually checked it properly.
With a set of calipers it is easy to be 0.1mm out just by holding them "wrong". You might always hold them "wrong", and so repeat to 0.01mm (that is likely the repeatibility spec based on the technology used).
Without a traceable calibration and a temp controlled lab and a whole load of other minor but actually matters things you are probably not making measurements to better than 0.05mm with a set of calipers.

I use my traditional Verynear calipers mostly - when it matters I use a micrometer or other method (gage blocks and slip joints are a good one here).

People get fooled by digital - its not any more accurate just becase the computer shows you a certain number of digits...

Dave

Who used to work in the field of very high precision metrology, and occasionally makes things in is garage

A high powered inspection microscope has a very thin field of focus.

Mine is 40x IIRC, and you can 'zoom' up and down an edge and examine it in great detail by ignoring the out of focus parts.

I got mine very cheap, mostly because the guy was moving to the USA and clearing out, but a similar one on ebay would probably not be that much.

Ill grab soem pics later when I head out to the garage.

Dave

for 20 euro, I am tempted to buy this set radius gauge , drill the rivet out, send John the 'leaf' he wants.

All those in favour, send me an un franked UK 1st class stamp.

Bob

I would repeat in response to the last question by John NO

Robert Butler

+1 on NO

My added bit - Hope this thread goes away and people stop wasting their time with this troll OP.

Save yourself 19euro and just drill a 1mm hole in a square of scrap, saw and file away one corner and you have an L shaped gauge with 0.5mm internal radius to send to john. I won't suggest he makes one as that could lead to another long thread

@Dave 2

> They use the same technology, so (for instance) the 200mm caliper is 0.01mm resolution,
> +-0.02mm accuracy, their 300mm caliper is 0.01mm resolution, +-0.03mm accuracy, and
> their 1000mm caliper is 0.01mm and +-0.07mm.

Yes, interesting stuff... however to my point, is that not exactly what you would expect if errors are proportional to the distance travelled by the sensor, no?

And if that is the case, then surely a smaller movement would imply smaller errors.


I know that my digital callipers have their theoretical limitations and are not doubt loathed by purists, BUT what I have been finding in practise (which for me is what counts!) is that reading my Vernier calipers, particularly when in read in sub-optimal lighting conditions, was either:
A) seriously slowing me up or
B) introducing serious reading errors or
C) sometimes both !

Yes you are absolutely correct about there being quite a lot of skill required in order to make accurate measurements. And my digital calipers are quick to scream at me in this regard! But readings are so quick that I generally measure each measurement at least 3 times to 'sanity check' each of my measurements.


Fwiw, yes I have an old x10 to x20 binocular dissection microscope which I sometimes use to see what's really going on! And occasionally this is extremely useful. The downside is that it can be rather difficult to get it into position.

No. As I said on Page #1, that style of radius gauge is of no use to me because it does not have these features:

Nobody is compelling you to visit this thread.
If you have nothing constructive to add, please leave.

OP promoted to Moderator as well now! And after 4 pages of posts we are are back to where we started from as with other Forum Topics posted by OP who also lacks grace.

If this "invention" ever gets beyond prototype stage I pity the poor bugger who draws the short straw to put it into production.

Robert Butler

I'd love to see the inventor's face when he finds all these lovingly crafted high precision features that have cost him so much time and effort immediately deleted in favour of simple to produce parts, that can be assembled by unskilled labour in a sweatshop somewhere none of us have ever heard of.

Implicit maybe, but in the spec - not at all. I can legitimately sell you a long scale that is off by 0.05mm after the same move that a short scale would be out of spec at...

Distance travelled is only one of the variables that a longer scale introduces. A longer scale is of course the most obvious one (thermal expansion). Time to move is another. There are others as well if you think about it.

[rhetorical question] Would a slow move over a short distance be less accurate than a fast move over the same distance? [/rhetorical question]

The main reason I tend to use my traditional Verynears is that, if Im honest, the digital ones nearly always have a flat battery

However I also use them because Im not fooling myself that they are super accurate. 0.5mm is plenty for verynears, even if they do read to 0.1mm. Like most here I will often make one part and then make the mating part 'to fit'. For that as long as the first part is within 0.5mm of its nominal it very rarely matters....

In the interests of this thread I have wounded a Stanley Knife blade:

I just boshed a 45ish degree 'radius' on the back edge with a diamond stone.

This is the (quite poor) photo through an inspecion microscope:

measurment of a radius on an edge is potentially doable in this manner, with the bonus that you can inspect the finish as well.

Inspection microscopes, unlike dissection ones, have an x-y stage under them. This one is imperial - it does 1/10000" on its dials. I also have one with 0.001mm micrometers.

Fixturing is not hard - there is no force. This would be adequate for instance (took it after, when I though a fixture photo would be useful, hence the upside downness of the blade):

As you are planning on subbing out this work once a spec is figured out you need to also think about how you are going to inspect and reject sub standard parts.

If a spec came to me saying 0.25mm radius I would look at it and think about:

A: Quote high enough that the work buggered off, or paid a hansome profit for actually doing the job properly

or

B: just break the edge like we always do and hope no one notices...

If its important enough to spec then its important enough to check.

Dave

Edited By Dave S on 04/01/2022 20:32:21

Edited By Dave S on 04/01/2022 20:33:10

Edited By Dave S on 04/01/2022 20:33:51

The OP might care to read this article:

https://en.wikipedia.org/wiki/Accuracy_and_precision

[This posting has been removed]

I just remembered that out good friends at ARC sell similar "L" shaped gauge sets. Metric set goes down to 0.5mm but the imperial one may be of more use as that gives 0.010", 0.015" and 0.020 (approx 0.25mm, 0.38mm and 0.5mm) so would cover a few of your smaller radii and not a bad price compered with a Starret set link