Surface plate support

[Paul L]

On Paul L Said:

Ideally a surface plate shoud sit on 3 points. These are called Airy (spelling?) points and are 22.6% of its length and width in from the edges. the 3rd point is on centre line.

Airy points come in pairs, not threes and, as DC31K says, are used on length standards, ie, a beam, to give minimum sag consistent with the end (vertical) faces being parallel. As such Airy points have no application for use with surface plates.

For a beam the two supports for minimum sag, ie, the centre sags the same amount as the ends, are slightly closer together than the Airy points.

Andrew

[Pete RimmerParticipant @peterimmer30576]

For a 12″ x 18″ plate just plonk it down on something it won’t bend to any degree you could measure or affect your work. I refurbish granite plates sometimes and I ran a test to see just how much they bend if not sat on their feet. I set up my autocollimator on one end of a 24″ plate and a very good mirror on the other (Hilger & Watts 0.000007″ flatness I think it was). Then I put a flat packer under one end of the plate And used a bar to lift the other end so all of the feet were off the bench and all of the weight was suspended between the two furthest edges. The Autocollimator showed about half an arc-second bend.

To put this in perspective, when I put my 10 second box square level on the floor of my workshop and walked from one side of it to the other, it registered the bend in the concrete slab on that 10 second scale.

I’ll point out here too that there is NO WAY to judge the flatness of a plate by it’s physical appearance alone. I have bought pristine-looking plates that were like the himalayas. I bought one which if you took a line diagonally across it, one side of the line was very flat the other side went up to a peak in the corner. One of my most flat plates is full of edge chips and pock-marks from poor storage/handling yet it works wondefully well for spotting and scraping. Long out-of-date calibration stickers or even grade stamps also mean nothing until you’ve had it checked. I have diamond lapped and surveyed a plate to grade A, scraped in one single straight edge on it then put a lap on it and the dip left behind is plain to see.

So, especially if you’re not going to have that plate checked, if you’re just going to use it, stick 3 cork coasters about 1/4-1/3 inside from the ends on the bottom two at one end and three at the other and stand it wherever you like.

[Michael GilliganParticipant @michaelgilligan61133]

This is an interesting note [referencing a standard]
https://optimalcalibration.com/wp-content/uploads/2018/06/3-Point-Support-System-Layout.pdf

The main ‘take-away’ must surely be that whatever choice the manufacturer makes is what we should use for subsequent mounting … This is not explicitly stated, but should be self-evident.

MichaelG.

[Robert Atkinson 2]

On Robert Atkinson 2 Said:

Dave (sod),

I have to disagree on the requirement for calibration. … Calibration only confirms accuracy. …

Robert.

Yes, but unless the item is calibrated the owner has no idea whether it’s accurate or not.   Like as not when an ancient surface-plate is calibrated, it will be found to be out-of-spec.   If so, what value is it?

My issue with folk buying random surface plates and other kit intended for high-precision work is that approach makes results  their untrustworthy.   Even though there’s little harm in hobbyists filling their workshops with rusty gauge blocks, because no-one else checks the actual accuracy of our work, that’s bodging, not engineering.  Real engineers try hard not to delude themselves,  or should, otherwise we fink we’re doing a bl**dy good job when we ain’t.    And paying customers don’t like that!

Dave

 

 

[VicParticipant @vic]

I bought a small Granite plate and mounted it into a box using the data from Michael’s post. Mostly though I use the cast iron table of my bandsaw …

[MikeKParticipant @mikek40713]

Yeah, I’ll probably do a simple support.  I basically just want to get it to a usable location and not blocking anything.  To be honest, I think my caveman brain never considered the possibility of just getting a flat plate instead of this 80lb burden.

Just found the spec, which I started to skim:

https://irp.cdn-website.com/4fbb675e/files/uploaded/Federal-Spec-GGG-P-463c.1.pdf

Mike

[SillyOldDuffer]

On SillyOldDuffer Said:

On Robert Atkinson 2 Said:

Dave (sod),

I have to disagree on the requirement for calibration. … Calibration only confirms accuracy. …

Robert.

Yes, but unless the item is calibrated the owner has no idea whether it’s accurate or not.   Like as not when an ancient surface-plate is calibrated, it will be found to be out-of-spec.   If so, what value is it?

My issue with folk buying random surface plates and other kit intended for high-precision work is that approach makes results  their untrustworthy.   Even though there’s little harm in hobbyists filling their workshops with rusty gauge blocks, because no-one else checks the actual accuracy of our work, that’s bodging, not engineering.  Real engineers try hard not to delude themselves,  or should, otherwise we fink we’re doing a bl**dy good job when we ain’t.    And paying customers don’t like that!

Dave

 

 

Are you sure this isn’t to some degree at least, a question of intellectual process rather than practical application – I don’t think you credit most of us with the realism of our own expectations.

 

[Michael GilliganParticipant @michaelgilligan61133]

< deleted >

Mike had already found the Federal spec.

[duncan webster 1Participant @duncanwebster1]

Unless you’re into scraping flat surfaces why do you need a super accurate surface plate? For marking out the traditional piece of plate glass or old mirror is quite good enough. In my bodger’s department, even tho’ I have a small surface plate, rather than digging it out from its hiey hole I often use the kitchen worktop. Compared to these I reckon almost any old surface plate is OK.

[Michael GilliganParticipant @michaelgilligan61133]

I wouldn’t argue against that at all, Duncan

… In fact the brief Starrett document that I linked previously is very clear on the matter.

MichaelG.

.

[SillyOldDuffer]

On SillyOldDuffer Said:

On Nicholas Farr Said:

…

 

I don’t know the accuracy of any of them, as they are all second hand, but they are good enough for whatever I will be using them for.

…

But what does that mean Nick?  I guess you have no intention using them for accurate work, but where do you draw the line:  ±1mm, ±0.2mm,  ±0.02mm, or ±0.002mm?

 

±0.002mm is the real mcCoy,  and a surface table would be valuable.   But a point often ignored by hobbyists is that this level is only achieved when done properly.   Buying a second-hand tenths micrometer, gauge block set, and surface table off ebay absolutely isn’t good enough.  They all have to be in good working order AND calibrated.  Plus, the owner has to know how to use them: ±0.002mm is only achieved by understanding what causes tiny errors – not ‘common sense’.

On arrival, second-hand tools might be ‘good enough’ to do a shade better than ±0.02mm, but even that is uncertain.

The reason I asked surface table owners to say what they’re used for, is I suspect surface tables seem like a good idea, but turn out to be rarely used in home workshops.  For us, I reckon they’re mostly a waste of money and space.   Most of our close fits are achieved by ‘fitting’, not by accurate precise measuring.  The most likely exception I guess is a repair shop needing to get accurate dimensions off random incoming objects.

Dave

 

Hi Dave, I’m certainly not making anything to the standards of NASA, or making tools to the standard of companies like Starrett or Moore & Wright, but I do do things that require a good level of accuracy, but let’s face it, a pair of spectacles is a compromise to 20/20 vision.

The photo below is not the best way to show how good my small rectangular surface plate is, and yes, you can see a smidgen of light coming past the 60 x 30 x 15mm block and the larger one beside it, on the left hand side, but the photo shows it worse than it actually is, and my thinnest feeler gauge of 0.0015″/ 0.0381mm, will not even try to enter that gap, not even the very tip of it, and this is what I mean by being good enough.

Regards Nick.

[larry phelan 1Participant @larryphelan1]

Seems like I might have to put my “surface plate” back in the skip !

A pity, because it looked nice.

[larry phelan 1]

On larry phelan 1 Said:

Seems like I might have to put my “surface plate” back in the skip !

A pity, because it looked nice.

It has long been demonstrated, Larry, that a discarded Rolls Royce can make a mighty-fine Chicken Coop !

https://www.iamfy.co/product/the-rolls-royce-chicken-coop-art-print

… I believe the modern term is ‘down-cycling’

MichaelG.

 

[Diogenes]

On Diogenes Said:

On SillyOldDuffer Said:

On Robert Atkinson 2 Said:

Dave (sod),

I have to disagree on the requirement for calibration. … Calibration only confirms accuracy. …

Robert.

…

My issue with folk buying random surface plates and other kit intended for high-precision work is that approach makes results  their untrustworthy.   Even though there’s little harm in hobbyists filling their workshops with rusty gauge blocks, because no-one else checks the actual accuracy of our work, that’s bodging, not engineering.  …

Dave

 

 

Are you sure this isn’t to some degree at least, a question of intellectual process rather than practical application – I don’t think you credit most of us with the realism of our own expectations.

 

Diogenes is on to something!  When hobbyists rush to buy second-hand precision kit, I am indeed questioning the realism of their expectations.  I suspect this stuff is bought because folk haven’t thought it through.  Rather than buying precision tools to meet a quantifiable precision requirement, maybe they believe in stuff!  Iffy motives like the appeal of a posh brand-name, or the idea that classic kit is better made than modern, bargains, time-expired workshop lore, plus a strong belief that the ravages of time don’t matter.

With many tools, not knowing the actual condition doesn’t cause any upset.   For example, if a posh second-hand milling cutter arrives with 50% of it’s working life consumed by the original owner, a hobbyist is unlikely to detect any problem.   The cutter works, and will continue to do so for many hours,  but it’s on the way out – sooner rather than later.    Whether the purchase was a bargain or not depends on how much was paid for it, and how much life the owner gets out of it.    Hobbyists are hazy about tool-life because we don’t have any way of measuring running hours, do a variety of cutting with the same tool, and don’t log what we do.   The potential for delusion is there, but mostly harmless, provided the cutter cuts for ‘long enough’ in the operators mind.

To my mind second-hand precision tools are in a different category.   They were specifically made to enable measurably precise results.  Their purpose is to know dimensions for sure within tight limits, not to provide a feel good factor.   And, unlike my cutter example, it is possible to quantify how good or bad they really are – have them calibrated!    This is the most important check, and rarely done by hobbyists because it costs money.    And embarrassing when it turns out our lovely shiny gauge blocks had been de-rusted with a sand-blaster, and none of them are on target!    Such sets make reasonable parallels, but are no good for measuring.  With precison tools there is a high-risk of delusion.

Can’t credit anyone with ‘the realism of our own expectations‘ because I don’t know what they are!  And I strongly suspect neither does anyone else.   Expecting an ancient tenths micrometer to be a shade more accurate than a new thou micrometer is an untested assumption, and it could be wrong.   Both micrometers can be tested if the owner has access to set of trustworthy gauge blocks, otherwise pie in the sky.

I might be deluded myself which is why I’m asking the folk who buy precision kit, especially second-hand, to explain themselves.

So far my favourite report is from Larry, who retrieved a section of granite worktop from a skip, used in conjunction with a T-square and 6″ rule.   He’s not spent any money, isn’t fretting about what’s supporting the plate or the importance of Airy points, and his T-square and 6″ rule suggest he has sensible expectations of his granite!

Dave

 

[Nicholas FarrParticipant @nicholasfarr14254]

Hi Dave, you do know that all engineering has a compromise, but the level of the compromise will depend on how deep your pockets are. Not everyone has the finance to purchase brand new tools, but many will get by very well by buying second hand stuff. My two surface plates I’ve shown above, were a fair price in my opinion, the price didn’t matter to me if they were not worth what I paid for them, and they do what I need, my large one didn’t cost me anything, it doesn’t get used very often, but has been very handy, and has coped okay. I could buy a new high precision surface plate if I wish to, but I’ve no reason to spend money on something that so far haven’t needed.

Regards Nick.

[Andrew Johnston]

On Andrew Johnston Said:

On Paul L Said:

Ideally a surface plate shoud sit on 3 points. These are called Airy (spelling?) points and are 22.6% of its length and width in from the edges. the 3rd point is on centre line.

Airy points come in pairs, not threes and, as DC31K says, are used on length standards, ie, a beam, to give minimum sag consistent with the end (vertical) faces being parallel. As such Airy points have no application for use with surface plates.

For a beam the two supports for minimum sag, ie, the centre sags the same amount as the ends, are slightly closer together than the Airy points.

Andrew

If supported on the Airy points, a uniform beam will be level at it’s ends. For equal droop at ends and middle you need Bessel pointshttps://www.engineeringnotes.org/metrology/airy-points-and-bessel-points/

[Kiwi BlokeParticipant @kiwibloke62605]

This is getting a bit silly, and beyond the reach of shed-based amateurs, I think. However, some of us chase the maximum precision we can achieve, sometimes – when patience allows and masochism drives us…

However, to continue the academic discussion…

‘If supported on the Airy points, a uniform beam will be level at it’s ends. For equal droop at ends and middle you need Bessel points’

Not sure about that. Bessel Points support a beam to maximise its length, so used for length standards, where the end surfaces are not required to be parallel. For minimum sag (which is presumably wanted for a flatness standard) the support points should be a smidgen (we’re talking precision here!) closer, at 0.5536 of the length (had to look up the figure…). Dunno whether they have an eponym though.

My 2ft square cast-iron plate was reconditioned by The Surface Flatness Co, near Manchester, for me, over a quarter of a century ago. Its calibration certificate shows a max. error of one tenth of a thou (“Inspection Grade”). Impressive! However, I had no way of checking it. Since then, it has survived a sea journey to NZ, probably involving large temperature swings, whilst in a container. Now I have an autocollimator, I must re-check it – it will have changed. But why bother? AFAIK, there is no-one around here to recondition it for me, and, although I enjoy scraping, a massive job like that won’t ever get to the top of the infinitely long to-do list.

[duncan webster 1]

On duncan webster 1 Said:

If supported on the Airy points, a uniform beam will be level at it’s ends. For equal droop at ends and middle you need Bessel pointshttps://www.engineeringnotes.org/metrology/airy-points-and-bessel-points/

So … does anyone have a reliable link to the derivation of these magic numbers ?

Duncan’s reference looks promising but [typical of many web-sites] it lacks substance … Just try the embedded PDF link and you will see what I mean.

MichaelG.

[Kiwi BlokeParticipant @kiwibloke62605]

Wikipedia references Airy’s original paper, from its ‘Airy point’ page. Pity my maths is no longer up to it…

[Michael GilliganParticipant @michaelgilligan61133]

Thanks for that, Kiwi

I’ve just downloaded it

… Mercifully brief at only four pages

MichaelG.

[duncan webster 1Participant @duncanwebster1]

I would have thought you wanted the deviation from perfectly flat to be a minimum across the whole plate. Hence Bessek points. If you move the supports closer together the middle will sag less but the ends will sag more.

[larry phelan 1Participant @larryphelan1]

Thank you Michael for those kind words ! I have decided to retain my “Surface Plate” and give it a second chance. Right now, it rests on a beercrate, but this will be upgraded.

Interesting observation regarding the chicken coop !, doubt if R&R would be too pleased.

[Michael Gilligan]

On Michael Gilligan Said:

Thanks for that, Kiwi

I’ve just downloaded it

… Mercifully brief at only four pages

MichaelG.

Regrettably, I am not much the wiser

Airy’s note is nicely descriptive, of a very specific set of circumstances, but is exceedingly difficult to apply elsewhere.

MichaelG.

[Kiwi BlokeParticipant @kiwibloke62605]

I’m relieved to find that no-one’s mentioned the possibility of floating the plate in a bath of mercury. No need for anyone’s points then…

[Howard LewisParticipant @howardlewis46836]

Never knew that hardboard was hydroscopic!

The idea was to ensure that the plate, being rather thin compared to industrial granite ones would be unstressed.

[Author]

Posts