Home workshop accuracy & tolerance?

[mechman48Participant @mechman48]

Posted by Martin Kyte on 23/09/2020 12:14:11:

. . . and then there is that age old process called easing the fit, used where you have generated your exactly mating parts and have to open everything up to make it run. I have done my fair share of that in the past.

regards Martin

Exactly! I served my apprenticeship as Fitter/Turner & after 4 years was given a choice of which trade to follow, as I didn't fancy being stuck behind a lathe for the rest of my natural ' I went down the fitting route. Being conversant with turning I could solder bearing halves together, bore out a bearing allowing for solder gap, then file & scrape bearing to 'FIT' the shaft. The number of times I've ' fitted' parts togetther to run as required, as any 'fitter' knows is inumerable.

George.

[Clive FosterParticipant @clivefoster55965]

Whilst I have every sympathy with folk who point out how things were done back in the day with minimal measurement capabilities by exploiting cut'n try, hand finishing and careful fitting its a slow old process. Partly inherently and, more importantly, due to the sheer amount of time it takes to learn to be good at it. Not to mention the waste of inevitable do-overs.

Inescapable in 1920 'cos the equipment to do better was unaffordable. Not so in 2020. These days the same relative expenditure to the left over fraction of a 1920's working mans wage that (eventually) mounted up to a Drummond round bed or similar pedal lathe and basic rule'n scriber level measurement gear will cover something adequately accurate, with dials and a decent mike, vernier et al.

Fair enough for folks who have worked old style for many years and attained appropriate skills to continue in that manner but anyone considering starting out today with a "cheap" Portass, Drummond Round Bed, Pools, Zyto, Grayson et al found in a shed has to be out of their mind. All of which I've used or at least kicked the tyres on.

Bluntly those things were crap when new, and even crappier now in old age, but when crap was all you could get you made the best of it that you could, cutting your ambitions to suit your cloth. Probably a Stuart or similar set of castings, maybe the ME beam engine if a little more ambitious. Even a Tich would have been a stretch for one without work honed skills. These days something like SouthBend, Atlas, Myford, Boxford or one of the better imports with proper dials has to be the starting point.

Learn to work off the dials so everything fits first time and reserve the handwork for places where its inescapable. Lapping pistons into bores et al.

Jason typifies the modern way. At his levels of productivity there is no time to spend on handwork when reading the dials or DRO will do.

Amateurs seek perfection. Professionals make it to fit first time. It may be a hobby but being a bit professional about things seriously cuts the frustration factor. Frustration may be said to be good for the soul but mine is doing just fine on the amount inevitable to life without seeking more by doing things the hard way on the hobby side.

Clive

[Brian HParticipant @brianh50089]

Where I first worked things were supposed to fit within a gnats (private parts).

Perhaps that's where the expression "cock on", came from, meaning that the job was just right.

Brian

Edited By Brian H on 23/09/2020 16:59:32

[Neil WyattModerator @neilwyatt]

Posted by Martin Kyte on 23/09/2020 11:26:40:

Interesting to note that when I made my telescope mirror I was able to achieve a variation of less than 50nm over a 10 inch glass paraboloid disc surface using purely hand work and simple optical measurements . It still astonishes me.

regards Martin

A classic example of where relative accuracy is what matters. Things like Ronchi tests make it relatively straightforward (if a long drawn out process, requiring skill) to achieve a parabolic mirror to within an extreme level of accuracy.

But a hand-ground a mirror is not made to exactly a particular focal length – something within a few inches of nominal is fine.

To compare with what we do, we might finish a shaft to within 0.0002" diameter but the accuracy of the actual circular profile may be accurate to within 10 or 100 times that.

Neil

[Clive FosterParticipant @clivefoster55965]

In the optical world hand methods too a very long time to die away. When I took redundancy in 2004 the experiemtal optical workshop at RARDE / DERA / DRA / DSTL to list the name acronyms during my employment, was still using the sort of techniques that Martin and Neil allude to.

But focal lenght of a paraboloid mirror within a few inches of nominal would certainly not have been fine. To be fair most mirror work would have gone to Optical Surfaces at Kenley who (mostly) used similarly unsophisticated methods to produce focal lenghts as near dead on as could be desired. I've measured 10 thou in a metre errors.

Clive

[Martin KyteParticipant @martinkyte99762]

Posted by Neil Wyatt on 23/09/2020 17:16:57:

Posted by Martin Kyte on 23/09/2020 11:26:40:

Interesting to note that when I made my telescope mirror I was able to achieve a variation of less than 50nm over a 10 inch glass paraboloid disc surface using purely hand work and simple optical measurements . It still astonishes me.

regards Martin

A classic example of where relative accuracy is what matters. Things like Ronchi tests make it relatively straightforward (if a long drawn out process, requiring skill) to achieve a parabolic mirror to within an extreme level of accuracy.

But a hand-ground a mirror is not made to exactly a particular focal length – something within a few inches of nominal is fine.

To compare with what we do, we might finish a shaft to within 0.0002" diameter but the accuracy of the actual circular profile may be accurate to within 10 or 100 times that.

Neil

Actually ou can get to better than 1/4 to 1/2 an inch without too much trouble especially as you can measure the radius of curvature when still on the grinding process rather than the polishing although most people make the mirror and then design the rest of the scope round it.

I does however illustrate my point about repeatability. To produce two identical mirrors with the same focal length takes much more effort.

Industry cheats there too. The chinese optics manufacturers effectively beat the rest of the world by high volume output and then measuring and grading what came off the line according to quality with the best lenses and mirrors commanding the highest prices. I am told similar things used to happen in race engine building when using production engines. Buy a slew of engines, strip and measure everything and build a few out of the most accurate parts. Tecnically it's a production engine but the chances of all those 'good fit' parts finding their way into the same unit commercially is remote.

As the man said "There is no substitute for cheating" or when someone is admiring that perfect model "yes but no one see's all the bits you threw under the bench"

regards Martin

[Andy StopfordParticipant @andystopford50521]

Industry cheats there too. The chinese optics manufacturers effectively beat the rest of the world by high volume output and then measuring and grading what came off the line according to quality with the best lenses and mirrors commanding the highest prices.

The same sort of thing is done in the production of computer processors: many blanks are sawn from a large disc of silicon, the ones from near the edge* can be run reliably at a higher clock speed, and are sold as the top-level, high performance, premium price model.

You can 'over-clock' the cheaper version, it is after all the same thing as the posh ones, but it may cause errors (which will usually result in the system freeze- or exiting).

*Or the centre, I forget which.

[Bill PudneyParticipant @billpudney37759]

This sort of thing has gone on since they invented factories….for example

1 Machine tool manufacturer, makes a lathe. At final test if the machine comes out with a particularly good set of numbers, it gets called a "Toolroom" lathe, and is given a special higher price tag.

2 Motorcycle manufacturer builds a bike, if the motor produces more power than a nominal figure it's called a "Sports" motorcycle, given a special Sports sticker and an elevated price

Ans so on….

cheers

Bill

[CirclipParticipant @circlip]

" Jason typifies the modern way. At his levels of productivity there is no time to spend on handwork when reading the dials or DRO will do "

Whoa there tiger, thought Muddle Ingineering was a HOBBY. Was taught how to use a Tally rond and a Tally surf but how many M/Es have them in the workshop? Or actually NEED them. "Orses for corses" or cut your cloth etc.. Disposable income has certainly altered the skyline.

Regards Ian.

[samuel heywoodParticipant @samuelheywood23031]

Thank you all, a fair bit of info to digest!

@Neil Wyatt~ a timely reminder of "push fit, sliding fit" etc, in my naievity i've more than once made a shaft 'precise' enough it didn't fit the bearing as intended!

@Circlip. I'm not worried about super accuracy per se, call it an intellectual enquiry of those better skilled & experienced than myself.

Hobby engineering has taught me to be more precise however.

Before taking it up, i pretty much lurched through life with approximations & rules of thumb.

@Jason B~ I've not noticed it taking any longer working to 1/10 thou than a thou yet as i'm not at that level

But you are right, working to ones best possible accuracy is definitely more time consuming.

@Martin Kyte~ yes it's amazing how optical workers can work in microns & by hand.

That's the main thrust of my hobby engineering at the momment, telescope parts.

.Commercial 1.25" eyepiece fit is very sloppy these days, only one i've come across that wasn't was old' Vixen Japan' stuff~ slide the ep into the holder & you can feel the air cushioning gravity slightly.

That's a correct 'fit 'for the intended purpose IMO.

Going to have to re read entire thread, a lot to absorb, thanks folks.

[Author]

Posts