What to get: Imperial or metric

[Neil WyattModerator @neilwyatt]

Fractions generally imply a precision of ~1/64 while decimals imply the precision they are written to. With care it's easy to work to 1/64" with a rule.

It's also much easier to see how fractions add up to an overall length e.g. for a 3/4" stroke piston, if the piston is 1/4" thick and the end clearance is 1/32" and the end covers are 3/32" thick the cylinder bock is 1 1/16 high and 1 1/4" over the covers. If all those dimensions were give in decimals I would have to reach for the calculator or at least a pencil and paper.

Neil

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Neil Wyatt on 28/08/2016 08:47:30:

Fractions generally imply a precision of ~1/64

.

Oh …

Now what do I do about those 1/128" dimensions ?

MichaelG.

[Bill PudneyParticipant @billpudney37759]

Generally, in the days when draughtsmen used pencils, drawings had to be concise, at least in organisations that took things seriously. So that somewhere they (the company) had a set of guidelines/rules, draughtsmen for the use of, and in there it would be stated that if fractions were used they had a tolerance of +/- some amount, decimal dimensions had a tolerance of +/- something else. "Something" may vary with the number of decimal places. So I'm broadly lining up with our Esteemed Ed. However, very little was implied, simply because that can easily lead to whoopsies. The key phrase is "….took things seriously…"

What really does, (or did when I was paid to care about such things) float my boat is poor and/or sloppy drawings. As I somewhat flippantly mentioned previously, inappropriate tolerances are one of the things that really matter and if the end product is destined for production can really increase the cost by huge amounts. Trust me, I used to run an manufacturing estimating section.

cheers

[Bodger BrianParticipant @bodgerbrian]

Posted by Neil Wyatt on 28/08/2016 08:47:30:

It's also much easier to see how fractions add up to an overall length e.g. for a 3/4" stroke piston, if the piston is 1/4" thick and the end clearance is 1/32" and the end covers are 3/32" thick the cylinder bock is 1 1/16 high and 1 1/4" over the covers. If all those dimensions were give in decimals I would have to reach for the calculator or at least a pencil and paper.

Neil

That's only true if you're used to using fractions. Presented with a bunch of them, I'd have to reach for a pencil & paper.

Brian

[Danny M2ZParticipant @dannym2z]

I cannot understand why people have problems converting fractions to decimal units.

Fire up the workshop calculator.

For example 1/64" equals 0.0156" (divide 1 by 64)

Multiply the result by 25.4 to convert to mm

So 0.0156" x 25.4 equals 0.3969mm.

So now, decision time – go for 0.4mm or hunt down a 1/64" drill bit

* Danny M *

 

 

 

Edited By Danny M2Z on 28/08/2016 11:17:07

Edited By Danny M2Z on 28/08/2016 11:17:47

Edited By Danny M2Z on 28/08/2016 11:34:02

[not done it yetParticipant @notdoneityet]

Nooo. 1/64 impies an accuracy of plus or minus 1/128.

 

In a similar vein, one cannot convert one unit into another and quote the answer to more significant figures (ie 0.0136 is not the same as 0.3969mm). A simple mathematics principle.

 

The simple maths principle calculates the mean of 1 and 2 as 2. NOT 1.5! Many on here perhaps do not understand why this is so – and some draughtpersons either, by the experience reported by Bill.

Edited By not done it yet on 28/08/2016 13:39:54

[Nicholas FarrParticipant @nicholasfarr14254]

Posted by not done it yet on 28/08/2016 13:38:09:

Nooo. 1/64 impies an accuracy of plus or minus 1/128.

In a similar vein, one cannot convert one unit into another and quote the answer to more significant figures (ie 0.0136 is not the same as 0.3969mm). A simple mathematics principle.

The simple maths principle calculates the mean of 1 and 2 as 2. NOT 1.5! Many on here perhaps do not understand why this is so – and some draughtpersons either, by the experience reported by Bill.

Edited By not done it yet on 28/08/2016 13:39:54

Hi not done it yet, nobody said it did, but it does equal 0.34544mm

sorry I guess you just did a typo.

Regards Nick.

[Nicholas FarrParticipant @nicholasfarr14254]

Hi, some people may not like my drawing here then:-

I measured the broken plate to the finest measurement I could and then made the new plate as close as I could allowing rounding up/down where it was reasonable to do so. This is a one of and I would expect anyone else who made one to these measurements to adjust them to there own abilities. I have put a photo in this album with this and another drawing converted to metric by request of one member, but have kept them by and large to an exact amount, leaving one to decide their own tolerance.

I think the choice of buying imperial or metric machines, has to be decided by whoever is going to use them and the units they prefer to work in.

Regards Nick.

Edited By Nicholas Farr on 28/08/2016 14:23:58

[Nicholas FarrParticipant @nicholasfarr14254]

Hi Jason, They are shown on the other drawing mainly for clarity on my part and are in fractions, thus avoiding mixed terms and do have other information associated with them, Metric converted one in mm's. (see Hobbymat Lathe Refurb Album)

Regards Nick.

Edited By Nicholas Farr on 28/08/2016 15:06:40

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Nicholas Farr on 28/08/2016 14:23:28:

Hi, some people may not like my drawing here then:-

.

Forgive me, Nick … I mean no offence, but:

"Dimensions ? … He can't even spell 'em"

.. MichaelG.

[not done it yetParticipant @notdoneityet]

You forgot to make any mention of K, the temperature at which you measured it!

Please tell us how you measured to one millionth of an inch (or to one part in about four million, or so).

What was the real tolerance for the build? To the nearest 10 thou close enough?

[Nicholas FarrParticipant @nicholasfarr14254]

Hi MichaelG, over the years I've found my Spells never work, however you seem to have interpreted the correct word

[Neil WyattModerator @neilwyatt]

Posted by not done it yet on 28/08/2016 13:38:09:

Nooo. 1/64 impies an accuracy of plus or minus 1/128.

I didn't say it was! Perhaps I should be clearer; in general fractions mean work to 'rule' dimensions.and you can generally assume someone using a rule will be accurate to within 1/64"

Naturally if you are faced with fractions smaller than 1/32", you need to take extra care but you will generally find dimensions less than this are typically the nominal sizes of materials or fixings or for adjustment (e.g. set valve so open 1/64" at TDC or leave a clearance of about this amount) generally not dimensions to be cut to – although there are exceptions few people would be expected to work to 1/128" with a rule.

Neil

[Michael GilliganParticipant @michaelgilligan61133]

Nick

MichaelG.

[Russell EberhardtParticipant @russelleberhardt48058]

Posted by not done it yet on 28/08/2016 13:38:09:

The simple maths principle calculates the mean of 1 and 2 as 2. NOT 1.5!

I suggest you revise your maths!

Russell.

[duncan webster 1Participant @duncanwebster1]

Posted by Russell Eberhardt on 28/08/2016 15:58:51:

Posted by not done it yet on 28/08/2016 13:38:09:

The simple maths principle calculates the mean of 1 and 2 as 2. NOT 1.5!

I suggest you revise your maths!

Russell.

If this was integer arithmetic, the mean of 1 and 2 would be 1. I've been caught out by this type of thing when using integers to reduce memory requirements in code for processors. Even wierder is for example 2 * 99/100, which comes out as 0. Why? It obeys BODMAS, so it does the divide first, which comes out as zero because it rounds down to the nearest whole number, then multiply anything by zero and you get zero. I'd much rather have 2 as the answer (real answer in floating point is 1.98), but you have to be careful how you write it

Edited By duncan webster on 28/08/2016 16:32:31

[Ray GandertonParticipant @rayganderton56482]

Base 10 metric is very simple but…and a BIG but!

As I grew up in an Imperial world my natural instinct is to convert metric to imperial and then

make a judgement. I have no feel for metric, which only time provides, and I haven't that much left!

[SillyOldDufferModerator @sillyoldduffer]

Posted by duncan webster on 28/08/2016 16:30:33:

Posted by Russell Eberhardt on 28/08/2016 15:58:51:

Posted by not done it yet on 28/08/2016 13:38:09:

The simple maths principle calculates the mean of 1 and 2 as 2. NOT 1.5!

I suggest you revise your maths!

Russell.

If this was integer arithmetic, the mean of 1 and 2 would be 1. I've been caught out by this type of thing when using integers to reduce memory requirements in code for processors. Even wierder is for example 2 * 99/100, which comes out as 0. Why? It obeys BODMAS, so it does the divide first, which comes out as zero because it rounds down to the nearest whole number, then multiply anything by zero and you get zero. I'd much rather have 2 as the answer (real answer in floating point is 1.98), but you have to be careful how you write it

Edited By duncan webster on 28/08/2016 16:32:31

I hope 'not done it yet' explains because he's set me wondering too!

I guess he may have been thinking about rounding when he wrote " The simple maths principle". In that case 1.5 would round up to 2.

After repeatedly being caught out by BODMAS as a programmer I started to use brackets galore. Colleagues good at maths were always annoyed by unnecessary brackets but it reduced the embarrassing numeric accidents I kept having!

Dave

[SillyOldDufferModerator @sillyoldduffer]

Another thought, sometimes it's useful to calculate the geometric mean. On the dangerous assumption I got the sums right, the geometric mean of 1+2 is 1.414 (-ish)

So depending how you do it, the mean of 1+2 could be 1, 1.414, 1.5 or 2

I think it was muzzer in another thread who warned against over thinking things…

Dave

Edited By SillyOldDuffer on 28/08/2016 17:22:50

Edit It was a dangerous assumption that I'd get the sums right: the square root of 3 is 1.732, not 1.414.  Serves me right for trying to be a smarty pants.

Edited By SillyOldDuffer on 28/08/2016 17:26:40

[Neil WyattModerator @neilwyatt]

Posted by duncan webster on 28/08/2016 16:30:33:

Posted by Russell Eberhardt on 28/08/2016 15:58:51:

Posted by not done it yet on 28/08/2016 13:38:09:

The simple maths principle calculates the mean of 1 and 2 as 2. NOT 1.5!

I suggest you revise your maths!

Russell.

If this was integer arithmetic, the mean of 1 and 2 would be 1. I've been caught out by this type of thing when using integers to reduce memory requirements in code for processors. Even wierder is for example 2 * 99/100, which comes out as 0. Why? It obeys BODMAS, so it does the divide first, which comes out as zero because it rounds down to the nearest whole number, then multiply anything by zero and you get zero. I'd much rather have 2 as the answer (real answer in floating point is 1.98), but you have to be careful how you write it

Edited By duncan webster on 28/08/2016 16:32:31

So… use Reverse Polish Notation

Different processors & languages use different standards for rounding – it creates havoc for Java as it has to run the same code in many environments, there's a long page on the paedia of wiki.

Neil

[Russell EberhardtParticipant @russelleberhardt48058]

Posted by Neil Wyatt on 28/08/2016 17:52:44:

So… use Reverse Polish Notation

+1 I'm still using a Hewlett Packard RPN calculator from the 1980s and have Reverse Polish calculator apps installed on my smartphone and tablet.

Russell

[Anonymous]

At one company where I worked we convinced one of the secretaries that Reverse Polish was so advanced that it wasn't even in the Kama Sutra.

Back to engineering; given the number of errors that have been posted in this thread when converting from imperial to metric, or vice versa, I'm amazed that anybody manages to make anything at all in any sort of unit.

Andrew

[Nick WheelerParticipant @nickwheeler]

Posted by Andrew Johnston on 29/08/2016 11:51:51:

Back to engineering; given the number of errors that have been posted in this thread when converting from imperial to metric, or vice versa, I'm amazed that anybody manages to make anything at all in any sort of unit.

Andrew

I follow my grandfather's advice on that: converting gives you another chance to f^#k up, learn to work in both. So although my machines are metric(as you would expect when they were bought new in the 21st century), on the rare occasions imperial jobs come up, I push the button on the digital caliper, and get an inch micrometer out of the drawer.

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Nicholas Wheeler 1 on 29/08/2016 13:53:28:

… learn to work in both.

.

I do …

I'm quite happy with dimensions like 20mm +/- 0.005"

MichaelG.

[area3fitterParticipant @area3fitter]

Thanks guys,

the thread took an interesting direction! I did my apprenticeship in imperial, am now 60 and wanting to set up my little workshop as I step back a little from work. I can handle the maths no problem for conversions etc. I'm thinking of metric and looking at the Warco lathes and mill.

Looking at the previous posts, will these suffer from the leadscrew discussions mentioned earlier?

thanks for the replies, they have been very helpful. I have only a benchtop lathe at the moment.

The lathe I am thinking of is the GH1330 or similar and the mill is the HV.

thanks all.

Gareth

[Author]

Posts