PRECISION , ACCURACY and QUALITY

Advert

New Replies

Report Bug

Help/FAQs

PRECISION , ACCURACY and QUALITY

This topic has 20 replies, 15 voices, and was last updated 30 November 2010 at 21:57 by Sub Mandrel.

Viewing 21 posts - 1 through 21 (of 21 total)

Author

Posts
25 November 2010 at 14:17 #59331
MICHAEL WILLIAMS
Participant
@michaelwilliams41215
Three of the most misused terms in the engineers vocabulary are :

                   PRECISION , ACCURACY and QUALITY .

Whilst each term does have a strict meaning they are mostly used quite randomly and often quite wrongly .

So :

       Question (1) what does each term really mean ??

       Question (2) What interpretation of each or all defines a ‘good’ piece of

                              engineering ??
Advert

25 November 2010 at 14:17 #21755
MICHAEL WILLIAMS
Participant
@michaelwilliams41215
25 November 2010 at 15:29 #59337
Peter G. Shaw
Participant
@peterg-shaw75338
Ok, here’s my take on them:

Precision: that quality which determines whether or not a machine tool can cut to certain limits, or fineness.

Accuracy: that quality which determines the absoluteness of the quantity, eg is it 23.456mm or 23.457mm. Should really be used with either absolute (for measurements which may be used elsewhere and hence should be related to a standard, or relative which implies that the actual value may not be correct when compared to a standard, but is in comparison with something else. Also very closely tied up with repeatability which refers to whether or not the same measurement or whatever can be repeated and be shown to have the same value each and every time.

Quality: a very much abused word in my opinion. How often do you see something stated as being high quality, or, to put it another way, have you ever seen something stated as being low quality! I refer of course to advertising.

Ok, I’ll now stand back and await a caning! But no abusive language please.

Regards,

Peter G. Shaw
25 November 2010 at 16:33 #59342
NJH
Participant
@njh
Tush Peter – and you an ex-BT man too!

I seem to recall Quality being defined as “Fit for purpose” however it’s a while ago and I will go with your definitions for the time being. I guess this is another thread which could run and run.

Regards

Norman
25 November 2010 at 16:41 #59343
oilcan
Participant
@oilcan
precision=the tolerances to which i intend to machine.

accuracy = the size i actually machine.

quality = accuracy minus precision.

which is why i always claim that the axles i turn are not all different sizes but merely bespoke to there individual wheels.
25 November 2010 at 17:03 #59346
Martin W
Participant
@martinw
Hi Oilcan

I can see the ground moving above the grave of Sir Joseph Whitworth as he turns at the word ‘Bespoke’ . But yes I suspect most of us do the ‘Bespoke’ bit from time to time .

Cheers

Martin
25 November 2010 at 20:05 #59356
Nicholas Farr
Participant
@nicholasfarr14254
Hi,

PRECISION; pertaining to being precise. i.e. High precistion = absolute minimum deviation from spceified value.

ACCURACY; the closest practical possible value to that specified. i.e. Plus/minus 0.001mm tolerance = Very accurate.

QUALITY; pertaining to an ideal charecteristic. i.e. Stone age engineering = probaly very low quality.

Thats my view of it.

Regards Nick.

Edited By Nicholas Farr on 25/11/2010 20:07:18

Edited By Nicholas Farr on 25/11/2010 20:08:34
25 November 2010 at 21:04 #59360
Sub Mandrel
Participant
@submandrel
Precision is the degree to which a measurement is specified i.e. 0.010″ is ten times a precise as 0.01″, even though its the a same measurement.

Accuracy is the range of variation of a measurement. 0.010″ +/- 0.001″ is more accurate than 0.010″ +/- 0.002″

Another way of looking at it is to think of target shooting. Get a goud grouping way off the bull and your shooting is precise, but not accurate.

Get a spread of shots centred around the bull, your shog is accurate but not precise.

Obviously to hit the bull with every shot you need both precision and accuracy to a suitable standard; do that and you have achived an appropriate level of quality.

’nuff said?

Neil
25 November 2010 at 22:32 #59376
Nicholas Farr
Participant
@nicholasfarr14254
Hi, Posted by Stub Mandrel on 25/11/2010 21:04:40

Another way of looking at it is to think of target shooting. Get a goud grouping way off the bull and your shooting is precise, but not accurate.

Not very precise if the specified target is the bull. I’d call it consistant.

I’ll put it this way; accuracy is the tolerance of precision.

Regards Nick.
25 November 2010 at 22:44 #59377
John Olsen
Participant
@johnolsen79199
Nick, You are a bit rough on the stone age guys, some of the things they built are still standing, which suggests that the quality was good. Some of the items were aligned to quite high precision too. There are stone structures built before metal tools were available where you cannot slip a knife blade into the gaps between the stones, so they were working to quite good accuracy too.

I’d agree with Stub mandrel about the shooting, it is precise, but not accurate if they are all in the same wrong place. This implies a different sort of error to correct than if they are all over the place.

regards

John
26 November 2010 at 00:28 #59386
Nicholas Farr
Participant
@nicholasfarr14254
Hi john, I’ll take your point on the stone age guys. Probaly not a good comparison, but I don’t think thier fashion would make it on the catwalk.

Some dictionary defenitions of precise: clearly expressed or delinated; distinct; definate; eactly corresponding to what is indecated; correct….. Or in other words not deviating from the specified target

Imagine two precision levels, one having a resolution of 0.1mm/10 and one of 0.001mm/10 but both claiming an accuracy of 1%. Thier accuracy would be the same but the precision will be much different.

Imagine the old English Longbow men one fires all his arrows precisely on the alliance with great accuracy. Bit of a job correcting that sort of error.

Regards Nick.

Edited By Nicholas Farr on 26/11/2010 00:29:38

Edited By Nicholas Farr on 26/11/2010 00:37:01
26 November 2010 at 08:55 #59396
Terryd
Participant
@terryd72465
Hi Michael,

Getting into semantics is a dangerous place to go, it can be Tiger country! Unfortunately the English language itself is, by it’s nature not very precise. It is this quality which, in many ways, makes it so useful and is one of the reasons it is more or less the lingua Franca of the internet. It is also the reason for a lot of argument especially on Forums like this. One of the reasons for this lack of precision is that many of the individual words have more than one meaning, and that attribute of the language is a major cause of differences of opinion.

For example QUALITY has two meanings and is also subjective, in the first definition it means an attribute. We say that individual metals have different qualities. This does not imply that one is intrinsically better than another. Steel has certain qualities that make it useful in a very wide range of applications while brass is better for others. However a second meaning of quality is how good something is in comparison with similar. My quality of life is probably better than someone similar 150 years ago. A Ford Focus is arguably of better quality than a Ford Anglia, while it would be unfair to compare the Focus with a Bentley.

PRECISION also has at least two definitions, but this time more objective. In the first definition, precision implies accuracy, and that accuracy is a quality (attribute) of precision in this meaning. I can make a one off model which is very precise, in this context. The second meaning is of repeatability, as in production engineering. in colloquial terms we say, for example, that ‘all of those mobile phones are made precisely the same’.

ACCURACY, on the other hand is less ambiguous. It is the degree of closeness to which a target or goal can be achieved. Be that target a dimension or the bullseye of an archery target. We can say of a darts player that his first arrow was precisely on target when he hits the treble 20 he wants. when he gets the other two in the bed we say that he threw his darts with precision. As I said Accuracy is a quality (attribute) of precision.

As I said, Semantics is a dangerous practice and is not very accurate at times

Terry
26 November 2010 at 09:17 #59398
Hugh Gilhespie
Participant
@hughgilhespie56163
Just my tuppence worth.

One thing that hasn’t been mentioned so far is repeatability. For us amateurs, this is often more important than accuracy. For instance, if we are turning a taper to fit a cone socket, we don’t really care if its 9 degrees or 10 degrees but by using the same set up to cut both we do get accurate repeatabilty and hopefully a good fit.

A lot of the cheaper measuring devices have a precision greater than their inherent accuracy, the digital angle gauge I use claims an accuracy of 0.2 degrees but reads to 0.1 degrees. However, provided its repeatability is at least as good as its precision this doesn’t matter as our normal type of measurment is a comparison, not an absolute measurement.

Regards

Hugh
26 November 2010 at 10:03 #59404
Gordon W
Participant
@gordonw
I like the old meaning of “nice”. Hate the modern meaning.
27 November 2010 at 11:09 #59469
MICHAEL WILLIAMS
Participant
@michaelwilliams41215
GW : I couldn’t agree more !

” all parts exactly suited to their purpose with each fitting another with perfect nicety and the whole machinery a wonder of function and beauty “

TD: On semantics (apart from not being entirely sure of the meaning of the word …. ) I’ve often thought that scientific principle should be applied to language generally . The only problem is that all the common principles of physics and mathematics get one into difficulty straightaway . For example a Heisenberg defined word for cat would mean cat (most of the time – probably ) but mean nothing at all for very small cats .

Oilcan: tongue in cheek I know but there are plenty of jobbing workshops that don’t do much better in real life !

To get back to business – what I really want to get at is whether ‘ Quality ‘ is something that can actually be defined or whether it will always be a subjective property .
27 November 2010 at 11:45 #59472
blowlamp
Participant
@blowlamp
Do you feel the question has been asked with sufficient PRECISION to allow me to provide an answer of acceptable ACCURACY that would be of a suitable QUALITY?

Edited By blowlamp on 27/11/2010 11:47:47
27 November 2010 at 12:06 #59476
KWIL
Participant
@kwil
I am with Terryd, Quality use of the language, precise in its construction and accurate in its spelling, what more could you ask for?
27 November 2010 at 14:30 #59480
Steve Garnett
Participant
@stevegarnett62550
Posted by MICHAEL WILLIAMS on 27/11/2010 11:09:13:

To get back to business – what I really want to get at is whether ‘ Quality ‘ is something that can actually be defined or whether it will always be a subjective property .

All the time it is being defined by subjective terms it will remain subjective, won’t it? Every time something is built to a defined standard that takes care of the technical aspects of its construction, you could say that it met that quality standard – as far as it’s defined. But most if not all of these standards don’t include any subjective guidelines, even – for the simple reason that it’s personal. It has been observed before that good engineering has a beauty of its own – and that is very hard to define in ‘quality’ terms. I suppose it could be argued that the Chinese may have taken advantage of this in the past to produce machines that look very much like quality engineering…

So you could say that quality in engineering terms can at least in part be defined, but whether this would ever amount to a ‘holistic’ and absolute definition – well I rather doubt it. You could take almost any two items, and build one like a brick outhouse, out of well-chosen materials, etc, to an accuracy specification blah blah, etc and give it an excellent paint job, and many people would say ‘yes, that’s a quality piece of engineering’. Alternatively, you could analyse exactly what was actually required of the item, and design a construction that met all of the specification using rather less material, didn’t look as good because you didn’t paint it, but technically might out-perform the brick outhouse version. They’re both ‘quality’ – but which approach is ‘best’?
27 November 2010 at 14:57 #59483
Terryd
Participant
@terryd72465
Thanks Kwil…. I think

regards

Terry
27 November 2010 at 19:24 #59497
Chris Trice
Participant
@christrice43267
Precision and accuracy are very similar but I’d class precision as meaning how fine a tolerance the machine is capable of as far as finishing a workpiece to size using any means whereas accuracy pertains more to how much you can rely on the dials to place any tooling in a position that meets your desired level of precision as a result of the standard to which the machine has been manufactured. Quality used to suggest that something was made to a high standard but the word doesn’t mean anything nowadays without an adjective in front of it.
30 November 2010 at 21:57 #59674
Sub Mandrel
Participant
@submandrel
My scientific, rather than engineering training is quite clear.

Precision is the degree to which a quantity is defined or measured – the number of digits after the decimal point, so to speak. It says nothing of accuracy, which is (for example) how closely the mean of a series of results approaches the true figure.

Opens some very interesting ideas. If you measure enough peoples’ height to the nearest inch, you can get an idea of average hight to a precision of a fraction of an inch. But to get the most accurate result you need to add random error (dither) to the measurements!

Neil
Author

Posts