Imperial fractions on drawings.

[Clive FosterParticipant @clivefoster55965]

Strictly the common 1/64 and ± 5 thou if no specific instructions are "make it to within" allowances rather than true tolerances. Fine for individually made items and simple enough that we all know what it means in practice but the real thing, as needed when making multitudes of things which have to fit together, usually has to be rather more sophisticated.

As an aside in the classroom I was originally told 10 thou if no instructions which was swiftly corrected to ± 5 when I got back to the shop.

When the drawing office of Mighty Great Big Industries gets involved they need to make sure that the actual dimension and tolerances concerned are unambiguously given and the permitted variation is such that the size variations for mating components don't actually overlap. Its embarrassing when multitudes of holes are too small for the matching shaft to go in.

Where fractions are concerned if a dimension is in 32 nds ± 1/64 takes you midway towards next one up or down so there is the potential for overlap. Taking measurement and calibration errors into account some percentage of components will overlap. When you are close to the middle its easy to estimate wrong way and tempting to do so with a big batch of components close to the (invisible) line! ± 1/128 has a clear gap between the middle of the interval to next 32 nd up or down and the right dimension so overlap is unlikely unless the measuring gear is way out. If you let the statistics guys in on the calibrations they will insist on a 95 % confidence limit that your measuring equipment is within ± 1/4 of the interval so all in all you should be safe and things will fit.

Same principle as a digital instruments least significant digit.

Clive

[KWILParticipant @kwil]

Just become familiar with both systems, if its drawn in Imperial, make it in Imperial and vice versa.

[colin hawesParticipant @colinhawes85982]

Here's my tuppenyworth …… In the early days of model engineering the average amateur would not be able to justify the cost of a micrometer so fractions would be preferable. Important parts would be made to fit together. As precision instruments became more affordable precision decimal dimensions became more prevalent. In my early days in industry it was common for dimensions to be drawn as follows:-

Fractional +/- 1/64"

Decimal +/- 1  on last decimal place as shown on drawing dimension

All dimensions in inches

It seems an unambiguous way of dimensioning a drawing to me! Colin

Edited By colin hawes on 03/04/2014 12:03:53

[Nigel BennettParticipant @nigelbennett69913]

What you have as tolerances on untoleranced dimensions rather depends on what you are building. What system you use for building bridges won't be much good for building watches!

BS EN 22768-1:1993 or ISO 2768-1:1989 should tell you all you need to know.

[Speedy Builder5Participant @speedybuilder5]

I had some of Martin Evans's drawings, all in fractions and for fit tolerances, used the terms 'FULL or BARE'. I also have a set of fractional verniers which measure to the nearest 1/128" on one side and 0.01mm on the other. My plans for SPEEDY don't have any tolerances on them at all. It is left to the "Engineer" to work out his own. Now it so happens that " I are an Ingineer"

My biggest problem is when turning something to metric dimensions on an imperial lathe. Do I re-dimension in inches or keep on measuring with metric mike etc.?? As above – It depends upon the tolerances required

BobH (AMI G Tech E; Tech Eng CEI – unsubscribed)

[John Stevenson 1Participant @johnstevenson1]

Seen painted on a lorry chassis some years ago.

Unladen weight:-

7 tons, 5cwt, 56 kilo

[jason udallParticipant @jasonudall57142]

Fractions. ..measure with ruler or tape measure..
.decimals..measure with micrometer or caliper.
…you guess which to use and how accurately to machine..

Best of luck. ..
Btw how much deviation from 1/4″ does it have to be before it is say 0.248″ Not 1/4″..?..
…I hate fractions used in measuring…it gives me that do they mean 0.250 or 0.25000..? Moment

Indudtrial drawings still exist in fractons but typically they use decimals where it matters and fractions for nomimal ie stock size..

[IanTParticipant @iant]

Bob,

Just about all my machines and measuring devices are Imperial but my favourite Vernier (the non-digital one!) has both Inch & mm scales. For some reason I find it much easier to read just the 'mils' off it these days. I also tend to draw (CAD) any new designs in metric/mm as I'm increasingly using metric fasteners and materials.

However, when turning (or milling) my machine dials are imperial, so of course it's natural to use Thou's but it runs a bit deeper than that. I still "think" in Thou's (probably always will do so) but it's not really a problem in practice. For most purposes (e.g. small distances) 0.1 mm is the same as 4 thou and I can do simple conversions in my head (so I can't be completely daft just yet!).

Whilst I do have full sets of Imperial drills (fractional, letter, No.) that will probably last me my lifetime, I now also have a full set of metric drills in 0.1mm steps up to 10mm that I find myself using more and more. And of course for larger dimensions they have these wonderful things called 'Calculators' to let you move quickly & accurately between the two systems.

As for tolerances – sorry but I don't do them. Generally I make a hole to roughly the required size and then make another bit to fit into it. If I want a very good fit (VG Fit?) – then I might ream or D-bit the hole first or if I want a really, really good fit (RRG Fit?), then a bit of lapping might occur (but I'm normally much too idle!). Fortunately, I don't work for Rolls Royce and it takes me quite long enough to build things as it is! But as usual, different horses for different courses – and everyone to their own..

Regards,

IanT

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Dimensioned drawings as such are rapidly going out of use in industry . Another 10 years and they will just be a memory in some places .

Already in quite a lot of design and manufacturing companies the only definition of a component that ever exists is a 3D CAD model .

This model is used for all design work and evaluation including finite element analysis et al .

From the optimised model then :

Tooling sets are ordered .
Material is ordered pre cut to billet size .
CNC machine control program is derived .
CMM validation program is derived .
Billet and tooingl sets automatically loaded into machine(s) by pallet or robotic system .
Machining done .

In a few cases with no human intervention . In many cases with human monitoring but minimal human input . At no stage are ‘drawings’ used . Whilst all dimensions of component are defined accurately – usually to high accuracy – no human operator needs to know in any detail what many of them are .

In cases where actual drawings are needed for problem solving etc the drawings are extracted from the 3D model .

The 3D models used are accurate at any plausible level of zoom . So enlarging more and more reveals more and more detail – for instance precision clearances are actually drawn in and at suitable magnification can be seen and if nescessary measured off .

Relevant to Model Engineering ? :

I’ve always thought that 2D assembly drawings with a few scrap views to show important details and plenty of notes were a better way to present construction projects than fully detailed parts drawings .

So moving into modern times a quality 3D CAD model of the construction project with all detail drawn in and which can be explored at will by the constructor does the same thing but in an even better way .

Individual constructors can extract individual component details for making by tradititional means or modern machining enthusiasts can fast track to CNC .

‘Dimensions’ as such can be extracted as needed and in different forms according to choice and to suit different constructors and processes .

Just a few thoughts .

Regards ,

Michael Williams .

[jason udallParticipant @jasonudall57142]

Michael..great ..as long as 3d model is well checked first..other wise..an error would propagate un noticed..all the way to delivery..
Seen it happen..
Any way engineer generated drawings are becoming the norm..and until the simulator/cad/cam system also defines the machining technique you will ..as I have..still get requests to quote on parts that the designer has no idea how to make.. ( you might argue that they might not need to ..but)..
I always found dimentions/ tolerances on drawings to seem to take almost as long as the rest of drawing..

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Hi Jason U ,

In a 3D model the ‘dimensions’ of parts are intrinsically locked together .

Errors are certainly still possible but the general experience is that most errors get spotted ‘pictorially’- without having to consider dimension details at all in some cases .

Hard to actually prove but as far as I am concerned design using a 3D CAD solid model is intrinsically more sound and less error prone than using 2D multiple flat view methods .

Even in more traditional design at the conceptual stage doing assembly sketches in isometric is often better than doing multiple flat views .

As for designers – if they don’t know how to make the parts being drawn by them then not only are they not proper designers at all – they are total liabilities .

Regards ,

Michael Williams .

[Nigel McBurney 1Participant @nigelmcburney1]

I started my apprenticeship in 1958 as a scientific instrument maker,the drawings were all imperial,but our boss was a firm believer in metrication,he felt that continental optical instruments were finer as the use of one mm and half a mm was better than one thirty second and three sixty fourths,particularly for the widths of small steps and widths of knurls on items like eyepieces,when I completed my apprenticeship I was presented with a Roche Etalon vernier and I had the choice of separate imperial scales for internal/external readings or two scales metric and imperial for the external measurement and subtract the jaw width for internal measurement,the boss recomended that i have the imp /metric version as he declared that metrication would be with us within ten years,also he said never lend your vernier to anyone they are all to easily damaged, I still have it and it is superb no batterys to run down or false readings though its initial cost was a weeks wages for a skilled man, within a couple of years I got a desk job so all my tools went home for model engineering and how I wish that my vernier had the dual imperial scales. Most of my time was spent with metric drawings as a desk bound engineer yet I still prefer imperial , my home workshop has an all metric triumph 2000 and a master 2500 with dual dials and an imperial super 7plus two mills one of each measurement, Though I am familiar with both systems I am far quicker when using the imperial machines. The one item that more use should have been made of was metric drills in imperial days ,when working imperial the 64th step was often too large when reaming or requiring a small amount of clearance on a drilled hole,its a lot easier to use a metric set in regular 0.1 steps than the irregular steps in number and letter drills.

Science labs in scools and universities in the fifties and earlier were metric ,98% of our instruments which were mainly for school labs had metric scales, I learnt in all metric labs at school,imperial units at tech,and then when we as a country turned metric,the metric units were all different to the ones I had used at school,crazy.

[John BromleyParticipant @johnbromley78794]

I thank you all for your replies, I never expected such detailed answers and opinions.

Kwil, you may well be on the money. It sounds very much like I will have to commit to a project using the designers chosen system of measure. It's not going to be easy, but I will learn much from it and it will make many future tasks easier.

John

[Neil WyattModerator @neilwyatt]

May I just add that the E-series preferred values, as used for electronic components, would be a great boon for designing in metric if used for material and fixing sizes. Even better than imperial, with its useful ratios.

Sadly only the BA system comes close to this ideal.

Neil

[Michael GilliganParticipant @michaelgilligan61133]

Posted by John Bromley on 04/04/2014 09:28:02:

Kwil, you may well be on the money. It sounds very much like I will have to commit to a project using the designers chosen system of measure. It's not going to be easy, but I will learn much from it and it will make many future tasks easier

.

John,

You mentioned that you have a Metric Lathe, and that you want to build some of Tubal Cain's designs that were drawn in Imperial Fractions.

I respect your conclusion, and KWIL's opinion, but please don't forget the suggestion that I made earlier:

By far the simplest, and most elegant, solution is to pretend that there are 25.6mm to the inch, and re-dimension the drawings accordingly. The finished model will be very slightly larger than the Imperial one, and your parts would not be interchangeable with those made by an Imperialist, but does that matter?

The beauty is that you would be making a proper Metric Version of the thing, instead of faffing about with converted numbers.

Reminder: In this alternate universe, there are 25.6mm to the inch, so

1/256" is 0.1mm

1/64" is 0.4mm

1/16" is 1.6mm

1/4" is 6.4mm

1" is 25.6mm

You can fill in the gaps for yourself, and it should take very little time to convert a drawing from Imperial Fractions.

The machining will be so much easier !!

MichaelG.

[Mark CParticipant @markc]

Jason, i agree with you on the 3D front – most of the firms I sub out work have only recently made the leap from faxing drawings to PDF! The ones that can take models generally don't want them as they always want drawings to work from and generally want me to do the drawing.

Michael, the idea that errors are vastly reduced might be true but when you do cock up it can often be monumental – I hold my hand up to having just opened up some M8 holes in a mating part as I discovered that the threaded holes receiving them are done at M10 and although the hole positions are "in context" or position linked the size is not – fortunately the change was easily accommodated otherwise it would have been an expensive afternoon as the plate with the threaded holes runs to almost £4K…. The other problem with solid models is that most people model at nominal size and unless you want to start doing the tolerance stack thing and multiple models in various stages of production (as cast, as turned, as milled etc) the whole system falls apart – and I am an avid supporter of 3D but you just can't beat a paper drawing with real information on it. The best option I have found is to send the model and a drawing and some shops like this arrangement as they can drag the geometry in from the model and then tweak the program from the drawing.

Mark

PS, does everyone know that the machinery show is on next week at the NEC (MACH2014), it's a real eye opener if you have never been (or if you have been) and you can register for free if you are near

 

Edited By Mark C on 04/04/2014 23:30:11

[GaryMParticipant @garym]

Hi John,

This **LINK** might be useful. Lots of other tables on his site as well. I find the thread tables here very handy.

Gary

[KWILParticipant @kwil]

Micheal,

i like your alternative universe but as I have DROs on everything, I probably will not change.

K

[Gary WoodingParticipant @garywooding25363]

Posted by MICHAEL WILLIAMS on 03/04/2014 19:27:05:

Errors are certainly still possible but the general experience is that most errors get spotted 'pictorially'- without having to consider dimension details at all in some cases .

Hard to actually prove but as far as I am concerned design using a 3D CAD solid model is intrinsically more sound and less error prone than using 2D multiple flat view methods .

 

Basically true, but errors can still occur.

A few years ago, I was a member of a small team who had to design and build a special wheelchair. One member, who was very experienced with Autocad, accepted the job of creating the drawings. We should have heeded the warning signals when the first drawings appeared with dimensions in 1/10000", but we continued until one day we got a drawing where two important parts simply wouldn't fit. A dimension was exactly 1" out, to 1/10000"!. He had committed the cardinal sin of modifying a part without updating the dimensions.

Gary

Edited By Gary Wooding on 06/04/2014 17:55:00

[steveteeParticipant @stevetee]

I have expressed my thoughts about the metric system elsewhere on this forum. In 1841 Joseph Whitworth developed the worlds first standardised screwthread system, it uses fractional sizes. In 1969 the Americans went to the moon, in imperial measurements. That'll do me. Like the earlier poster I was expected to be able to convert fractions to decimal sizes from memory, it's a useful skill. Oh and by the way .250 and .250000000000000000 are exactly the same thing ……… it's like saying " Ooh look I have 25.00000p in my pocket". Nought is nought to however many decimal places you like.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by stevetee on 08/04/2014 01:12:12:

Oh and by the way .250 and .250000000000000000 are exactly the same thing

………

Nought is nought to however many decimal places you like.

.

But 0.0001 is not the same thing as 0.0000

… even if using three decimal places would appear to show that it is.

MichaelG.

Edited By Michael Gilligan on 08/04/2014 07:40:19

[JoParticipant @jo]

Posted by stevetee on 08/04/2014 01:12:12:

I have expressed my thoughts about the metric system elsewhere on this forum. …….. In 1969 the Americans went to the moon, in imperial measurements.

But I think you will find when they went to the moon they did not use fractional measurements.

Today even the Americans are modernising as many of their bigger car manufacturers are moving over to use metric, so it can't be long now before they catch up with the rest of the world

Jo

[steveteeParticipant @stevetee]

One tenthousandth and zero are never going to be the same , how could they?

Here is a link to New Scientist which carries alink criticising NASA for still using imperial measurements in 2009. One sentence reads "But NASA still can't make the jump to metric.".

This would tend to suggest that it would be unlikely that the Apollo program 40 years earlier used the metric system.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by stevetee on 08/04/2014 16:26:07:

One tenthousandth and zero are never going to be the same , how could they?

.

Agreed … So what point were you making with the reminder assertion that:

"Oh and by the way .250 and .250000000000000000 are exactly the same thing"

I'm not being funny … it's a genuine question.

MichaelG.

Edited By Michael Gilligan on 08/04/2014 17:42:52

[Another JohnSParticipant @anotherjohns]

Regarding the USA and Canada and metric –

1) the "kids" doing 3D printing over here are using metric, and seem to have no problems. That includes kids in the USA.

2) had a discussion with a retired construction guy in my Canadian province who stated "all construction plans are in inch", unfortunately, in the province of Ontario, the ton of construction site plans in the office where my wife works, are all in metric. When asked, the engineers said that it's been that way for at least a decade.

So, the old "inch to my grave" guys are out of touch with reality, and will be more so in the coming years.

I don't care one way or the other, but, gosh, the world is metric, for better or worse.

Another John.

[Author]

Posts