Metric or Imperial, Fractions or Decimals

[Roderick JenkinsParticipant @roderickjenkins93242]

Hi guys,

I’m redrawing a small i.c, engine I designed and built a considerable time ago, with a view to possible publication. When I drew it (in DesignCAD 2) I was guided by Tubal Cain’s Workshop Practice 13. He advocated Imperial fractions for “rule” measurements and decimal inches to a thou or better where precision was needed. That system seems to me to be based on a method of working that largely involves marking out and centre popping. I tend to use co-ordinate methods on the lathe and mill these days. My new CAD package has the option to dimension with a second unit system, so what should it be: fractions and thous, thous and millimeters, fractions and millimeters or what? And what do the US prefer? I get the impression that fractions are still favoured over there. Advice please.

Rod

[Roderick JenkinsParticipant @roderickjenkins93242]

Dimensioning drawings

[Gray62Participant @gray62]

Personally, I have no preference, having been brought up with imperial measurements, I tend to use those predominantly however, if a drawing is dimensioned in metric measurements it would not deter me from using those drawings.

Produce the drawings in whatever units YOU are most comfortable with. If need be, most engineers can convert from one to another

regards

CB

[Keith LongParticipant @keithlong89920]

I’d second that, but with the proviso that if you’re happy with metric then these days that is what I think you should use.

It’s fairly clear that most model engineers build thing from plans produced long ago as well as new drawings, so looking to the future, our replacements 30 years up the line will be looking at your “old” drawings, but the ones doing the looking will have been brought up with, and be more familiar with the metric system.

In my working life over the last 40 years metric has been by far the most common system that I’ve used.

Keith

[JasonBModerator @jasonb]

I’m also happy to work in either but if you do decide in imperial I would prefer to see it all in decimals for the same reasons of marking out (or not).

As for the US most of what I have seen is imperial in decimal form, Model Engine Builder has all there plans done this way and the other recently drawn designs are the same.

J

[Eric CoxParticipant @ericcox50497]

What ever convention you choose stick with it and please do NOT mix them

[KWILParticipant @kwil]

Does that mean I have to use expensive and sometimes difficult to obtain Imperial fasteners when the near Metric equivalents are readily available (very small sizes excepted)?

[Roderick JenkinsParticipant @roderickjenkins93242]

Thanks for the input. Decimal inches with a mm option seems to be the way forward. KWIL, joking aside, I will probably offer an option of metric fasteners. There are few in the design but they are all BA.

Does any body have a feel for whether UK customers buy metric of imperial machines these days? If I get the chance I’ll ask Warco at Ally Pally this weekend (can I mention that on this forum?).

[TigermothParticipant @tigermoth]

BA or metric fasteners?

I recall going into a model shop in the 90’s and asking for 4 BA tee nuts.

Oh no, we don’t sell BA now, we’ve gone metric.

I then bought 2.5 and 3 mm metric tee nuts and went away happy.

Until about 3 / 4 months later when I returned to buy more metric tee nuts.

Oh no, we only sell BA here – there’s no demand for metric!

Nowadays I seem to have gone completely metric anyway but I still have a small stock of 4 BA which I have always found to be a very versatile size.

Bob

[SpurryParticipant @spurry]

Posted by Roderick Jenkins on 20/01/2011 12:41:17:

Does any body have a feel for whether UK customers buy metric of imperial machines these days? If I get the chance I’ll ask Warco at Ally Pally this weekend (can I mention that on this forum?).

I can only speak for myself, but I would never entertain a none metric machine.

You only have to look at a set of metric plans and compare to imperial ones. The metric ones are much simpler to read. For those brought up with imperial, they continue with what they know well, a human trait I suppose.

To asknewcomers to use imperial when they have only been taught metric, always seems rather pointless to me.

Pete

[mickParticipant @mick65121]

if your drawing plans for an I/C engine then I would suggest at least three places of decimals for imperial and two for metric. Fractions confuse the issue by giving the impression that the dimension is not important, fine for any castings in fresh air, but for mechanical components precision is always the best approach.

[BowberParticipant @bowber]

My personal preference is Metric but metric or imperial it’s always decimal, much easier to deal with when using a calculator or doing a quick paper calculation.

Steve

[Gordon WParticipant @gordonw]

My preference, as an ex draffy, is for metric on new work. Parts and materials in the dimns. they were made to, eg. 1″x1″ bar, not 25.4, 2″ dia bearing, not50.8mm, etc. Another thought is when making a scale model, especially a larger one, do you stick to original dimns. or metricate them ? I would stick with the original measuring system, or it could get complicated.

[Peter GainParticipant @petergain89847]

Metric or Imperial; could depend on how many reamers you have & how often they are used in the project. Using modern, affordable measuring equipment it is easy to turn or mill in either units. But replacing several good quality reamers at once can mount up to a considerable sum.

Perter Gain.

[Speedy Builder5Participant @speedybuilder5]

Have a look at the hotch potch plans for Southern Belle (Gauge 1 by Martin Evans).

Here the drawings are in milimeters, SWG, imperial sectionsand imperial fractions for diameters. Threads are TPI in Whit form and Ba. I guess this assumed that many model engineers are tooled up with imperial kit and that linear metric measurements were chosen to suit the scale of 10mm to 1 foot. Working on the basis that we are tooled up for imperial, at our imperial lathe, we turn a piece (Valve spindle) which is imperial on the diameters and metric for the lengths. Same would be true if you had the metric lathe where imperial fractions have been specified.

Surely if metrification is chosen for a plan, then metric thicknesses and threads should be specified.

Whilst criticising the plans, why are steam valves drawn with non captive spindles ??

Apologies to the publishing draughtsman, but in my opinion (for what that is worth) lets have metric OR imperial – I don’t think we readers mind which, but not a mixture of both.

[mike osParticipant @mikeos54812]

For me one or the other, dont care which but please be consistent.

mixing steel gauges ditto, most places only seem to keep metric now anyway & SWG seems to be metric equivelant.

fractions… always as decimals, I have never seen, any instrument capable of displaying 6″1/8 other than as 6.125………& converting is a PITA

[TerrydParticipant @terryd72465]

Posted by mick on 20/01/2011 16:53:49:

if your drawing plans for an I/C engine then I would suggest at least three places of decimals for imperial and two for metric. Fractions confuse the issue by giving the impression that the dimension is not important, fine for any castings in fresh air, but for mechanical components precision is always the best approach.

 

Hi Mick,

You are of course perfectly correct when you say that fractional measures are seen as less accurate. We used to use fractional measurements when designing and making large structural steelwork but decimal for machined components, depending of course on the tolerances required.

 

However it is a strange fact that mathematically, fractions are more accurate than decimals. Fractions represent exact amounts whereas decimals are often (not always) compromises. The classic example is of course pi which is exactly 22/7 but cannot be defined exactly in decimals. This is one of the main reasons why fractional calculation methods were taught in school mathematics. Unfortunately decimals are easier to calculate using electronic means than fractions although there are fractional calculators now I believe.

 

Best Regards

 

Terry

Edited By Terryd on 22/01/2011 13:39:26

[Billy MillsParticipant @billymills]

Terry

Have you gone mad? Pi has NEVER been exactly 22/7 since Archemedies time…it is more like 3.141592653589………. If you must use approximations 355/113 is a lot better. And you such a stickler for accuracy and precision?

Regards

Alan.

Now wait for the explosion………

[Nicholas FarrParticipant @nicholasfarr14254]

Hi Terry,

               I have a calculator that will do fractions, its a Casio fx-85ES.

 

   Dispite myself always believing that Pi is 22/7, when pressing the symbol button for Pi it does say it is 3.141592654, but unless we’re working for NASA or the like, I think we’re getting a little picky here.

 

Regards Nick.

Edited By Nicholas Farr on 22/01/2011 14:27:47

[TerrydParticipant @terryd72465]

Hi Nick and Alan,

 

of course it has a decimal equivalent but pi is non repeating and is believed to be infinitely long. At school we used to have a mathematical magazine called, believe it or not, Pi Magazine which I subscribed to for 5 years (and looked forward to in anticipation once a term). The decimal pi was printed along the top and bottom (header and footer if you prefer) on every page on 9 point font. This went on for the 5 years, once a term. It is interesting that despite calculating pi to several million places it has been found to be non repeating.  I was perhaps a mistake to use Pi as an example ,my real point was about fractions not pi, and I apologise for that if it led to confusion, (by the way 52163/1660 is even closer).

 

All I said was that mathematically, fractions are exact numbers unlike decimals which are usually quoted to ‘so many decimal places‘ or ‘so many significant figures’. I didn’t say that I used fractions in my machining. Of course some decimals are exact for example 1/8 is obviously 0.125, and 19/64 = 0.269875 but the latter would be generally rounded up to 0.270, so mathematically it is not exact. That’s all I said. I know that pi=3.141592654………, but mathematically it is not exact. Of course in engineering terms 3.142 is close enough, but to emphasise I was speaking in pure mathematical terms.

 

I’m not challenging our practices or being picky just stating a mathematical fact about fractions.  Forget pi.

 

Best regards

 

Terry

Edited By Terryd on 22/01/2011 14:53:16

[Nicholas FarrParticipant @nicholasfarr14254]

Hi Terry,

               agreeded. This calculator that I have is one I bought in a sale some time ago and has a lot of differant fuctions on it, many of which I don’t even pretend I understand totally, I just use it for curiosity more than anything.

 

I use run of the mill calculators most of the time, and 3.142 is what I’ve always used for Pi .in my work and have never had any problems with it.

 

Regards Nick.

 

P.S. I aws not challenging.

Edited By Nicholas Farr on 22/01/2011 15:08:56

[Bruce VoelkerdingParticipant @brucevoelkerding91659]

Hi, Rod.

I am a mechanical engineer in the States. A lot of work here is still done in the inch systemsince inch-sizedmaterial and fasteners are more readily available than metric sized stock or fasteners.

In machine design, one rarely sees fractions. More common is two place decimals for loose-toleranced parts (original 1/8″ -> .13″ with a +/-.01″ tolerance) and three place tight-toleranced parts ( original 1/8″ -> .125″ with a +/-.001″ tolerance). But, many working machinists will take a .13″ dimension and machine to .125″. Sort of goofy, eh ? We should have converted to metric in the 70’s.

I have a question: a year or two ago, a contributor in Model Engineer remarked that conversion from fraction to decimal is easy “if you know the system”. I have always wondered what he meant ? Theconversion is interesting – a deciaml equivalent is always similar to a “higher” order fraction such as:

19/64 = .296875 contains:

.96875 = 31/32

.6875 = 11/16

.875 = 7/8

.75 = 3/4

.5 = 1/2

This why the deciaml equivalents have a certain feel about them. If one sees19/64 =”.2986375″, one knows instantly it’s incorrect, since there is no “.6375” in the series. ( I am assuming one has at least the 16ths memorized.)

Therefore, if one can determine the first say three digits of an odd 64th fraction, filling in the remainding terms is easy as they have to fit the pattern above. Personally I work in what we call “decimal-fractions”, that is, if I am using 5/16″ stock I enter that as .3125″ in the CAD drawing, even though it may only be dimensioned two-place (.31″).

Any ideas ?

[John OlsenParticipant @johnolsen79199]

Well, I have always thought the Americans were sensible because they tended to mostly use decimal fractions on their drawings, and, even more important, used third angle projection. I alway though the arguments in vavour of first angle were pretty specious…yes, if I put a semi transparent version of the object in a box and shine a light through, I will get the top view at the bottom, the left view at the right, and so on. However, that is not actually how I use the drawing, and being human I tend to look for the view of the left hand side on the left, where it would be in real life. The only thing about American drawings I have not liked is those funny fasteners…how big is 10-32 anyway? But BA is open to the same objection.What I don’t like is having to add up a string of inch fractions to arrive at the actual figure to machine. You know the sort of thing, 5/8 +13/16 + 7/32. (Now was that supposed to thirteen sixteenths or one and three sixteenths…?)

On the Pi thing, may I (hopefully) clarify things a little. Pi is an irrational number. In mathematical lingo, that means that it cannot be expressed as a whole number fraction. There is no pair of whole numbers a/b that will equal Pi. The same is true for the square root of two, “e” and a whole lot of other interesting numbers. A decimal fraction is still a fraction, just written differently to save space. For example 0.123 = 123/1000. It follows that there is no decimal fraction, no matter how long, that exactly expresses Pi, or root 2 or e for that matter. As engineers this does not matter to us, since nothing we make is actually perfectly circular anyway, and we can rarely measure things to better than four digits accuracy anyway.

On the original topic, I think we are at the point now where metric should be the preferred system for all new work . Yes, the system is not without its idiosyncracies. I grew up with Imperial, have worked in both, and find I prefer metric.

regards

John

[TerrydParticipant @terryd72465]

hi John

 

Pi is a transcendental number as well as an irrational one. As an amateur mathematician I find it fascinating, Have you tried looking at is as a Gregory-Leibniz series, it’s most interesting

 

Best regards

 

Terry

Edited By Terryd on 23/01/2011 00:47:40

[John Stevenson 1Participant @johnstevenson1]

Fractions?

I have yet to see any machine tool with the dials graduated in fractions.

Wood working machines don’t count, we are talking accuracy here not axe strokes.

John S.

[Author]

Posts