Anecdotes 2

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

There is always discussion of limits , fits and accuracy on this site so a I thought that I'd tell you about one or two oddball examples that have appeared correctly on real parts drawings .

(1) On a modest size and quite simple steel component :

Radius of 6000 inches +/- 200 inches

What do you think the part was and why such an odd dimension and tolerance ???

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

[Ed DuffnerParticipant @edduffner79357]

Peco Setrack 1:1 scale?

or part of the Hadron Collider?

[JasonBModerator @jasonb]

I sould think the decimal points were missing maybe it should be 6.000" +- 0.002"

[DurhambuilderParticipant @durhambuilder]

Hmm, doesn't necessarily need to be a complete circle, it could be a short arc?

[Martin WParticipant @martinw]

How about something like plates for a large fuel storage tank or a piece of railway track. Where they will finalise the diameter/radius on site by pulling into final shape.

[Sub MandrelParticipant @submandrel]

Something like a point frog? 500 feet +/- quite a lot sounds like the spec for full size rail radius.

Neil

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Its a dimension and tolerance which appears on a drawing of an aeroengine component .

Less sophisticated examples of the same component can be found in more commonplace engines .

???

[jason udallParticipant @jasonudall57142]

Moderate sized….?say 12"long .so approx 2 milli radians 0.1 degree or 6 minutes……or somthing like 3 thou on 6 "(3 thou each end of 12&quot..difference of straight line and arc segment….

I have programmed such radii to compensate for "barrel"shapes in journals…

6000+/-200…seems a realistic Toll…in percentage terms…

I have had (automotive production) drawings such that a drawn feature may disapear if other dimensions are at the extreem of their tolerence…(think three lengths ….over all. and two lengths in from opposing ends)….

I sometimes wonder if anyone reads these before issue.

[_Paul_Participant @_paul_]

Modest size, simple & part of an aero engine….. could it be the curvature of a rocker arm? or perhaps the prop boss?

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Hi Jason ,

Close enough !

The pins and journals on some Bristol radials were crowned to reduce local stresses .

6000 inches +/- 200 inches was a typical value .

Don't know what the good component yield percentage was in manufacture but they would certainly have been inspected and approved/rejected individually .

Regards ,

Michael Williams

Edited By MICHAEL WILLIAMS on 19/01/2013 09:19:32

[StovepipeParticipant @stovepipe]

Looking at Argos website yesterday – a "Dremel" accessory set, weight 450 KG. Blimey, some accessory !

Dennis

[Ed DuffnerParticipant @edduffner79357]

Posted by MICHAEL WILLIAMS on 19/01/2013 09:11:47:

Hi Jason ,

Close enough !

The pins and journals on some Bristol radials were crowned to reduce local stresses .

6000 inches +/- 200 inches was a typical value .

Don't know what the good component yield percentage was in manufacture but they would certainly have been inspected and approved/rejected individually .

Regards ,

Michael Williams

Edited By MICHAEL WILLIAMS on 19/01/2013 09:19:32

But how would you go about checking this tolerance? Isn't it more sensible to reduce the figures to workable/measurable values?

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Hi Ed ,

The crowning has to be of very large radius compared to the component – really its just a very slight reduction of diameter of pin (or whatever) at outer ends .

The 6000 inch +/- 200 inch figures appeared on engine drawings which were done in the 1930 to 1950 period .

Crowning is still very important on many engineering components but the means of specifying on drawings and means of manufacture have moved on considerably .

Regards ,

Michael Williams .

[Phil PParticipant @philp]

I once had a mate who got sacked for being a thou out on one job.

He worked in a bank at the time.

Phil

[jason udallParticipant @jasonudall57142]

Posted by MICHAEL WILLIAMS on 20/01/2013 19:14:58:

Hi Ed ,

The crowning has to be of very large radius compared to the component – really its just a very slight reduction of diameter of pin (or whatever) at outer ends .

The 6000 inch +/- 200 inch figures appeared on engine drawings which were done in the 1930 to 1950 period .

Crowning is still very important on many engineering components but the means of specifying on drawings and means of manufacture have moved on considerably .

Regards ,

Michael Williams .

It would be , "even" today be a near impossible dimention to check with linear 1D instruments..the "radius" would be Calculated from a cloud of 2D or 3D points….

many features today are more difficult to measure/verify than they are to generate..(both in the DO and in the factory)……

[Sub MandrelParticipant @submandrel]

Jason – how would you actually machine such a radius – some complex geometrical too to generate the radius without having a 6000 inch bar? or would you simply lap the journal concave?

Neil

[jason udallParticipant @jasonudall57142]

How would I?

Cnc if not ground.( or if less that 1000)

Centerless grinder from ("parallel" ) with profiled wheels if ground finish..(or more than 1000 pcs)

Diamond dresser cam / former CNC generated and verified as below…

The above deals with making it..checking/inspection of 6000" radius would be the real chalenge..I would use our 2D "height" gauge and rely on mathematics to derive radius from three points…..serveral times

Just how I work in the day job…

[Michael GilliganParticipant @michaelgilligan61133]

Posted by MICHAEL WILLIAMS on 19/01/2013 09:11:47:

Hi Jason ,

Close enough !

The pins and journals on some Bristol radials were crowned to reduce local stresses .

6000 inches +/- 200 inches was a typical value .

Don't know what the good component yield percentage was in manufacture but they would certainly have been inspected and approved/rejected individually .

Regards ,

Michael Williams

Edited By MICHAEL WILLIAMS on 19/01/2013 09:19:32

.

It would be interesting to know a little more about these components

… Length, Diameter, Surface Finish …

Whilst I have no historical evidence to support this; I suspect that 6000" radius +/- 200" was intended to prescribe "tramlines" within which the actual shape [including its Straightness and Surface Finish] could be whatever suited the manufacturing process.

Jason's expert description shows how he would make it now; I remain intrigued by how they would have done it in the 1930s

MichaelG.

.

P.S.   Whilst browsing, I just found this eBook about the Packard Diesel Radial.

 

Edited By Michael Gilligan on 22/01/2013 08:32:40

[jason udallParticipant @jasonudall57142]

in 30's… either centerless or between center grinding .

As to measuring it….well in production I would expect 3 "snap" mikes or dial indicators ( I have 1/10,000 " clocks in my collection…) …mounted in fixture controlling/fixing the postion along the part….. but I would be guessing…

[Anonymous]

Hmmmm, if I've got my calculations right, and assuming the journal is 2" wide, I make that a difference of 0.000083" on the radius. In other words less than a 10th of a thou. The tolerance variation would be rather smaller.

Under wartime pressures to get engines off the production line I suspect it may have been quietly ignored, plus a certain amount of muttering by machinists about draughtsmen who came up with the tolerance in the first place.

Regards,

Andrew

[jason udallParticipant @jasonudall57142]

Posted by Andrew Johnston on 22/01/2013 11:50:06:

Hmmmm, if I've got my calculations right, and assuming the journal is 2" wide, I make that a difference of 0.000083" on the radius. In other words less than a 10th of a thou. The tolerance variation would be rather smaller.

Under wartime pressures to get engines off the production line I suspect it may have been quietly ignored, plus a certain amount of muttering by machinists about draughtsmen who came up with the tolerance in the first place.

Regards,

Andrew

Quite… inspection would be challenging

[Joseph RamonParticipant @josephramon28170]

Perhaps just hand held emery to relieve the centre would do the job.

Joey

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Rollers in roller bearings are crowned in volume production and these are of even flatter curvature . More sophisticated methods are available now but for many years the crowning was done by tumbling .

Very flat radii on components are easy to measure optically – again there are new sophisticated methods but for many years checking was done by the fringe counting method .

Need to make a distinction between the diameter of the component as such and the crowning effect when considering measurement methods .

The original Bristol components were hand lapped . May sound hard work but it was a doddle compared to making some of the other engine parts . It may well have been that the actual dimension was just a ' guide in principle ' and that a slight fairing off of the outer ends of pin was usually in the right ball park dimensionally after a skilled worker had done a few by trial and error .

Quick and often used ' workshop ' test for crowning is to spin component on glass (barrel wise ) . If it spins freely crowning is present and probably adequate .

Other things are sometimes crowned – one example is of gear teeth for demanding applications .

Michael Williams .

Edited By MICHAEL WILLIAMS on 22/01/2013 23:32:52

[Michael GilliganParticipant @michaelgilligan61133]

Posted by MICHAEL WILLIAMS on 22/01/2013 23:27:59:

Quick and often used ' workshop ' test for crowning is to spin component on glass (barrel wise ) . If it spins freely crowning is present and probably adequate .

Michael Williams .

Thanks … I will remember that one!

Bound to come in useful sometime.

MichaelG.

[Author]

Posts