How accurately can you machine?

[Tim StevensParticipant @timstevens64731]

I am trying to find a 'standard' to which the home machinist, with a small lathe and/or mill, etc, can expect to produce work.

From experience, it seems that anything less than plus or minus a thou* on diameter or length is pushing things (ie it requires special care or an extra finishing op to do better than that). Is this reasonable?
*0.0025mm

And what would be just as reasonable in terms of angles?

So, it would help if you could let me know how expert you think you are, and how accurate would your output be, remembering that you would be working with a home or hobby machine, and not with stuff worth more than a three-bed semi in Alderley Edge.

Thanks for your help

Regards, Tim

PS I'm sure I have seen a page of helpful data somewhere in the Machinery's Handbook, or similar, but I cannot find it. Can you?

[Tim StevensParticipant @timstevens64731]

[Neil WyattModerator @neilwyatt]

I'm probably about half as good as I claim to be

Neil

[JasonBModerator @jasonb]

So assuming we are only allowed basic home/hobby machines I assume similar will apply to measuring equipment and we won't have anything that measures better than a thou either

Most of the time I would say better is not needed, I tend to do 99% of my measuring with the digital callipers so not the most accurate measuring device despite reading to 0.0005.

If a part needs to fit another then I would tend to turn until I'm a couple of thou of the measured mating part and then offer the part up until it fits so thats feel not measured.

I rarely get out the *micrometer but will use that if I need to get to a good size that can't easily be test fitted, something like a crank pin for example and will usually polish that down to size wit emery on a stick. Pistons and bores may need lapping or honing and the piston would be fitted to teh bore not measured.

J

* Note I said THE mic, as I only have one that is 0-1" anything else gets the digital callipers and I do make quite a few larger size models. Tell a lie I do also have a digital 0-1" but don't like the feel of it.

[Chris Evans 6Participant @chrisevans6]

I always try and do the job as accurate as is needed. I have friends who chase silly accuracy when the accuracy of a plough would do. As a retired toolmaker I chase a thou only when really required, most jobs I do a couple of thou is good enough and gives the speed to get on with the job. Greater care is taken if boring/turning for bearing fits ETC. If I am making nuts and bolts for a tractor or motorcycle then the thou or two is usually good enough.

[norman valentineParticipant @normanvalentine78682]

I don't understand 'why' you need to know. As Jason pointed out you only need to make things fit. If you are building a one off model it doesn't matter how accurate other people make their work, as long as yours fit. Unless of course if your hobby is making things as accurately as possible. If so, that is an honourable pursuit.

Edited By norman valentine on 06/06/2016 15:39:16

[John RuddParticipant @johnrudd16576]

I'm with Jason….a fitter turner….lol.

You can have the Rolls Royce of machines capable of maching to accurate finish , but if the tools for measuring are poor, then thats as good as you will get….

I tend to aim for finished sizes as per drawings, but in the absence of such, its a case of if it fits its right….

I once turned a pin for some guy up in Scotland, gave me dimensions to work to….I got to within a thou when finished…Never did get to know whether it fit or not…

[Steve PaveyParticipant @stevepavey65865]

Just to be pedantic, a thou 0.001" is 0.0254mm (or 0.01mm is about 4 tenths). If I'm on the lathe I'm disappointed if I get worse than 0.01 mm out from my target. But it all depends – sometimes you need a clearance hole – I've just turned up a Vesconite cutless bearing for the boat, which needs around 0.15mm clearance on the shaft, plus 0.10 minus 0.03, so not exactly difficult to achieve that.

The other point is that it helps to know a few dodges for when things go wrong – machining a bore for a ball race demands good tolerances, probably to under a thou in old money, but Loctite or even a few strategically placed centre punch marks can sometimes rescue a job that has gone south on the final cut (so I've heard…..).

[Steamer1915Participant @steamer1915]

Posted by Tim Stevens on 06/06/2016 14:43:56:

 

From experience, it seems that anything less than plus or minus a thou* on diameter or length is pushing things (ie it requires special care or an extra finishing op to do better than that). Is this reasonable?

*0.0025mm

Just to avoid confusion. 0.0025mm is nearer to a tenth of a thou. A typo obviously, but thought it needed pointing out.

Best regards,

Steve.

edit : Steve beat me to it.

 

Steve.

 

 

 

 

Edited By Steamer1915 on 06/06/2016 16:47:16

[Neil WyattModerator @neilwyatt]

I think the difference we have as hobbyists is having the time to work to tolerances that are closer than needed, while not doing enough work to develop the skills of a jobbing engineer. Like Jason, I generally work to what fit is required and what 'looks' acceptable. Sadly all my care comes to nothing once I put a coat of paint on top and ruin it.

Neil

[BazyleParticipant @bazyle]

Do people working in metric work to tighter numbers (note I didn't say accuracy) because the 'round figure' is a smaller dimension? We know that many early lathes didn't have dials as standard and the early Myford topslide screw was 12tpi with an 80 division dial (optional extra).
Not so long ago Modellers didn't have mics and might be used to working to fractions rather than decimal so when the drawing said 3/8 in did they convert it to decimal and work to a thou – I doubt it. A thou is after all a human invention that just appeals to our fondness for round figures.

[MuzzerParticipant @muzzer]

If you have the skill and patience, it looks as if you may be able to approach 1um accuracy. I think you can safely rule me out on that score though.

BTW, this links to one of our suppliers. I visited their factory in Dongduan back in February and saw some of this machinery in the flesh. The machines claim to be able to attain accuracy in the 1 – 5um range. Being mostly Swiss and Japanese, I have no reason to doubt them.

Edited By Muzzer on 06/06/2016 17:09:09

[Rik ShawParticipant @rikshaw]

When I did this for a living the production turners did not need to chase thou's as an allowance of .005" – .010" was invariably left on for grinding. The same applied to most milling. Making one off bits – as we do as hobbyists – to a more accurate size is more demanding and certainly more time consuming. This is where emery cloth, grinding paste and Loctite come to the rescue. In the final pursuit of precision and without having the benefit of grinding machines etc. I think that constantly aiming for machined super precision when using lathe or mill is for the most part, time wasting. Viva the bodge – long live the fudge!

Rik

[EugeneParticipant @eugene]

As a raw beginner I'm happy enough when turning; if I want to be super precise inside a thou is OK using a good micrometer or dial gauge as appropriate. The key is good sharp tooling.

Where I come a ball of chalk is in tasks other than turning; cutting to length is definitely iffy, and milling in the lathe is a bit of a nightmare at the moment. It's mostly lack of technique plus some poor instrumentation, so I'm doing my best to rectify both.

I find the cheap digital calipers pretty poor; not a patch on my steam driven Moore and Wright micrometers.

Eug

[Martin 100Participant @martin100]

Posted by Eugene on 06/06/2016 17:12:52ing my best to rectify both.

I find the cheap digital calipers pretty poor; not a patch on my steam driven Moore and Wright micrometers.

Yes but with calipers you can make the figures match to within 0.01mm / half a thou regardless of the whatever the real fit is Vernier is allegedly French for very near.

I went digital on almost all my measuring instruments some time ago. The mechanical ones (mainly M&W + Starrett) are still there in their well worn cases from a generation or two ago but are rarely used.

[JasonBModerator @jasonb]

Posted by Bazyle on 06/06/2016 17:00:07:

Not so long ago Modellers didn't have mics and might be used to working to fractions rather than decimal so when the drawing said 3/8 in did they convert it to decimal and work to a thou – I doubt it. A thou is after all a human invention that just appeals to our fondness for round figures.

They set their firm leg caliper with the aid of a steel rule to 3/8" and machined the part until the feel of the jaws over the part was right

[duncan webster 1Participant @duncanwebster1]

Is it only me that often turns things exactly 0.5mm wrong when using a metric micrometer. Pestilential things, use a digi caliper to get near then switch to the mike if that sort of accuracy is needed.

[RichardNParticipant @richardn]

 

Vernier is allegedly French for very near.

I like the quote!!

I think it may have been Ivan Law or Sparey I was reading who was saying in the home workshop using mics and verniers as comparators is appropriate since absolute tolerances from parts made 'outside' are irrelevant- I tend to aim for the first part to be working within around a few thou, and make the second to fit- I don't have the luxury of time in chasing the last microns round the shop…

Edited By RichardN on 06/06/2016 18:50:23

[JonParticipant @jon]

As others depends what it is, the only way I can get 0.0005 accuracy on mill is to send it down several times, where 0.001 will start to become sloppy thats a thou. Its due to the flex of cutters, machine, vice, job etc.
Just so happens got to finish off a dovetail friction fit no slop which means 0.005mm off each face becomes sloppy, done by trial and error as Chinese Super Lux machine wont hold tolerances within 0.35mm whether one off job or production.

Lathes worn out only had it 5 years struggle to hold 0.06mm, so experience takes over from there on to get tolerance.

Careful how use calipers have to make sure its square by feeling for both flats on job and pulling in for external. Theres a 0.04mm discrepancy on my Mitutoyo coolant proof internal to external measuring maybe because dropped dozens of times. Also changes with temp as much as 0.06mm often around 0.03 within 20 mins, have to be careful. Other Mitutoyos not as bad had 6 and used daily without respect, its a dispensable tool.
Mitutoyo digi mic can have discrepancy of 0.001mm needing constant zero before use every time.

[RichardNParticipant @richardn]

Has anyone on here any thoughts on the difference between these two methodologies for obtaining repeatability?

1- first cuts take relatively deeply- maybe 30thou/pass, then as you approach the final measurement make progressively lighter passes, 15thou, 5thou, then shave the last thou to suit.

2- take first cut at around 10thou, measure before and after so you know what that tool actually removes at a pass (which may vary due to sharpness, material, speed, time of day etc) then remove whatever amount is required allowing for a final pass or two at the 'known' depth of cut.

I assume 1 thou cuts with a dull cutter are just rubbing and polishing, so not guaranteed to be reliable and consistent?

[mechman48Participant @mechman48]

I'm in the process of building the S10V & am working to kit dwg sizes; I am using digi vernier callipers having converted certain dimensions to 'thous' although I have the Stuarts book 'Building a vertical steam engine' by Andrew Smith, revised by Pengwern, where all dimensions are in metric & imperial, some of the kit dwgs have certain dimensions missing so this is where the book comes in to play… I digress, I am mainly using the digi vernier callipers, but recently have used my 'Mitutoyo' digi mic to check sizes in the crankshaft & webs, to check the hub of the flywheel I have used a pair of spring outside callipers then checked with the vernier between the nibs for 3/4" … this brings into play the question of 'feel'…

I also have Starret / M & W analogue mics which I use on occasions so have plenty of ways to check sizes if needed, at present on this model I am working to + / – .001"- .002" … .03 – .05 mm which I find is more than acceptable for my needs. So what is accuracy… however accurate you need it to be, to meet dimensional requirement which usually includes some form of + / – figure, & how accurate are your measuring tools, usually set up to some form of 'Standard' … so, if all your equipment has no play whatsoever, all your gibs are set perfectly, you don't get any frequency vibration from your electrics, your bench / tables are 100% rock solid… & you want to work to a tenth of a thou', & your tools can measure that… then go ahead, no one is going to stop you, I have on the very rare odd occasion but must have hit a lucky day…

George

[Ex contributorParticipant @mgnbuk]

Is it only me that often turns things exactly 0.5mm wrong when using a metric micrometer. Pestilential things, use a digi caliper to get near then switch to the mike if that sort of accuracy is needed.

Nope, happens to me too & I have taken to checking the mike with a digital caliper to make sure I'm not 0.5mm out. Frequently get asked by operators at work to check a Metric mike reading for this – they also have digital calipers & don't think to use these as a quick check on the mike for confirmation.

Don't seem to have the same problem with Imperial mikes & have a poundshop large button calculator double-sided taped to a cabinet above the lathe for instant conversions when the 0-1" is to hand & Metric dimension required. The Polish 0-1" mike was a subscription "gift" with ME or MEW quite a few years ago – pity such useful "gifts" are not offered these days. I have Metric mikes to 100mm & Imperial to 6", so can measure most things I can fit in the lathe. These are not "officially" calibrated & most are used (some of the Metric are "new old stock" DDR made, with standards) but a check with a slips set from work suggest they are all pretty close – though should an accurate measurement be called for I would be wanting to check more carefully. Like others above have said, most times use a digital caliper (Aldi / Lidl) while roughing out & mike to finish.

I try to get within +/- 0.01mm on diameters at home (Super 7 with mainly HSS tooling) whether the part really needs to be that accurate or not – more to try to get into an "acccurate" habit so that there is (maybe) more chance of getting an accurate part first time when it does matter. Same goes for surface finish – practice getting it good so that good is the norm. Frequently don't succeed on either count but, hopefully, regular practice will make for more reliable results !

Milling / shaping probably get within +/- 0.03mm – I seem to be able to get reliably closer than this a work on an XYZ turret mill with DRO & new cutters, so probably a combination of less rigid machinery, no DROs & less use of new cutters at home gives the worse result (FB2 clone mill & Boxford shaper).

Have not yet tried the Ingar RT612 surface grinder in anger, but used the same make/model at the last employment & could get as close to a desired size as I could measure – just took time & elbow grease !

[Jeff DaymanParticipant @jeffdayman43397]

Posted by RichardN on 06/06/2016 18:59:15:

 

Has anyone on here any thoughts on the difference between these two methodologies for obtaining repeatability?

1- first cuts take relatively deeply- maybe 30thou/pass, then as you approach the final measurement make progressively lighter passes, 15thou, 5thou, then shave the last thou to suit.

2- take first cut at around 10thou, measure before and after so you know what that tool actually removes at a pass (which may vary due to sharpness, material, speed, time of day etc) then remove whatever amount is required allowing for a final pass or two at the 'known' depth of cut.

I assume 1 thou cuts with a dull cutter are just rubbing and polishing, so not guaranteed to be reliable and consistent?

Richard your method is exactly the opposite to that taught to me by journeymen turners all those 1000 years ago when I was training. The one guy I learned the most from about machining in general was an ex British Army armourer. There was nothing he couldn't make, and all work done to a very high standard. The method I was taught was to plan your cuts 30 thou deep for all including finish cut. If the depth of cut was not a perfect multiple of 30 thou you could take a couple passes a little less than 30 or a little more than 30 but all cuts should be in that sort of depth. The main thing was to take that final full depth cut under load and with a known good tool / feeds / speeds setting, established in earlier passes, for good finish. He was always saying "don't take skin cuts!" loudly, meaning don't try to get to final size with .002 – .005"  cuts – you can get surprises that way, and ruin the work. 

Now this training work was on a good big Colchester Student that was rigid and had some HP on the spindle. Since then when using any big lathe I still follow the 30 thou/cut advice and normally get good results right on size. However on my home shop South Bend I reduce the 30 thou cut to 15 thou cuts but again the last cut is full depth and to size.

On Chinese hobby lathes, Atlas, Myford etc small light lathes this heavy cut to final size strategy does not always work due to lack of rigidity of the lathe design. If the lathe is not rigid enough you may get "surprises" but then you may get them with light cuts also.

Shaving or rubbing cuts with flat ended tools at shallow angles to the work can do OK but I have had these ops go pear shaped if workpiece materials don't co-operate.

Just my $0.02 worth. Your mileage may vary. JD

Edited By Jeff Dayman on 06/06/2016 19:48:35

[Anonymous]

I can't answer the OPs question as I'll get excoriated for not having typical home workshop equipment or measurement gear. However, I would say that it is the wrong question. The tolerance on a part should be set by its function and usage not by what the operator can, or can't, achieve. That means I work from tenths to inches depending on the requirements, not based on what I can achieve if pushed.

I'm not even going to attempt to tell you how I made the hole in an el cheapo stainless steel bowl this evening so that it fits upside down on the bird feeder in an attempt to defeat the squirrels. But it didn't involve anything precision.

Andrew

[Clive HartlandParticipant @clivehartland94829]

I agree there Andrew, parts made to a spec. for the purpose required. Force fit, running fit, sliding fit. Machining parts individually is OK as a one off. In the terms of business, parts need to fit all items so tolerances must be kept and adhered to. Imagine the car industry if tolerances were not kept.

Surface finish, type of materiel and sizes are up to the individual. It is quite easy to maintain sizes to less than 1/2 thou. which is adequate for most purposes. The design of parts is another factor in where only part of a shaft is machined to a high tolerance and the rest left a rough finish ie. bearing seats. Most people would machine a shaft to size all the way along its length when that is not required, only where the bearing is seated matters.

Clive, Andrew did you put a bearing in the bowl so that it spins. A fishing line swivel will give hours of fun watching squirrels fly off the edge of the bowl.

[Author]

Posts