How do I measure approx 60" ACCURATELY ?

[awemawson]

My Jones and Shipman 1300 EIUR cylindrical grinder spec says that it's  swivelling table is 60" long.

Now I need more information than that as I am measuring linear displacement of the table end to set a very accurate angle. The pivot point of the upper table is totally obscured but I think that a fair assumption is that it is dead in the centre of the table. The ends of the table are radiused, but again obscured by the table clamps but I think that I can project a tangent from these radiused ends forwards past all the obstructions using steel rulers and set squares to measure between them to get the length at the point that I am making my linear displacement.

So now I (theoretically) have two steel rulers projecting into free space and I need to know the distance between them - what options do I have:

A/ Tape measure - I'm sure that this isn't accurate enough - my displacement is to  tenths of a thou !

B/ Using my Leica Disto laser measurer, but it's only accurate to 1 mm

C/ Length Bars - I don't have any and certainly don't have enough Jo Blocks to stack up to 60 inches !

Now the accuracy of the 60" leg of my triangle can be less than the displacement (about 3.5") by swivelling the table by the ratio of their lengths and keep the same overall accuracy (I think) but any geniuses out there do please comment !

. . .all suggestions welcomed 


(best guess so far using a carpenters steel tape gives about 58.75" )

[BillTodd]

Are trying to set the angle of the table relative to the bed , rather than the angle of the work piece to the bed?

Sounds like a job for a good sized sine-bar jo blocks and an indicator.

Not sure why you're talking about steel rules?

Ok re-reading a couple of times...for  some reason,  you are trying to measure the angle of the table.

I would rig up a straight edge along the T slot  (to use the slot as reference)  setup a sinebar with something like adjustable parallels as the spacer ( you need something adjustable and measurable with a mic)  set an indicator on the head an sweep the bar , adjusting the parallel until sweet, then measure the parallels and whatever blocks to get the the value to plug into the appropriate bit of trig for your length of sinebar.

[Pete.]

Why not reverse-engineer the solution? Set a sine table to a known shallow angle using Jo blocks and put the base of the sine table against the side of the grinder table. Rotate the table until a tenths-reading (or better) dial reads 0-0-0 on the sine table as you traverse the grinder table and measure the amount of displacement at the point of measurement. From that you should be able to extrapolate the leg length from the pivot to the measurement point with at least as much accuracy as that system of measurement would provide from a known length.

The only difficult bit is that you won't be measuring from the actual pivot but off-set from it by half the table width. Just thinking about the trig for that is giving me a headache.

[awemawson]

I’ll put pictures up tomorrow but basically I am using the machine as a sine bar with an approximately 30 inch long leg (half table length) with the jo blocks replaced by a nifty long travel displacement gauge that indicates to tenths of a tenth.

My issue is measuring accurately the “long leg” ie the equivalent of the spacing of the sine bar rollers. Clamping a far shorter sine bar to the ways will give me far less accuracy.

(The angle that I need to set is 7 degrees 7 minutes and 30 seconds)

For years I’ve used the approximate  value of 30 inches but for this it is not accurate enough.

[Pete.]

I think that the problem is going to lie with the fact that your corner of the triangle that you're intending to use for trig calculations is moving with the table, given that it's not on the pivot point but on the front edge of the table adjacent to the pivot.

[russ57]

What is the longest you can accurately measure? Assume 6". Make yourself 2 accurate 'guage blocks' and leap-frog them down the length of the table, measuring the last gap with whatever you have.
Or make 10 and line them up end to end.

Not as good as a 60" vernier but they are not common i understand..

-russ

[WeldingRod]

I would expect a rational number.
Thus, I suggest setting an easy gaugable angle based on the assumed length and do a sine bar check on what you actually got.

Sent from my SM-G892A using Tapatalk

[efrench]

Interestingly enough,  that angle (7.125 degrees) makes measuring the leg lengths easy.

[awemawson]

Now that must be why they chose that odd angle, the ease of setting it up !

This is the angle of the  spindle nose taper of the A1 / A2 series of lathe noses. In my case I’m grinding test gauges for A2-5 and A2-6 chuck recesses.

That’s obviously more than just a mathematical co-incidence. I’d cottoned on to it being 7.125 degrees but not WHY it was chosen in the first place.

[awemawson]

To set the swivelling upper table to 'zero swivel' the only accurate way seems to be grind something and do iterative adjustments until it is parallel, so at this point in time the headstock centre and tailstock centre are dead in line and on zero. Now sliding the tail stock to another position inevitably slightly alters this setting. Even tightening the table clamps makes slight movement of the swivel angle.

So my action plan is to grind a bar parallel (the one you see currently in the machine), and this bar is the same length as the arbor the gauge I want to finally grind to that odd angle is mounted on. So the bar can be removed and the arbor fitted (sprung loaded tailstock)  WITHOUT moving the tailstock.

Then using the calculated swivel displacement the upper bed can be swung over and the taper ground (hopefully!) accurately.

The tailstock obviously obstructs the bed ways when it comes to measurements of length.

In the past I've made up a long fat parallel test bar hoping to set it between centres and use a dti to set zero - it just doesn't work. In fact I can remember John Stevenson telling me to forget the idea (he used to run one of these grinders at one stage of his career) saying that you just have to keep making tiny adjustments and re-measure. In fact googling the question of setting a cylindrical grinder parallel I came across his answer to my question back in 2004 !!!!)


OK as promised, some pictures:

[Joules]

Or as previously demonstrated, use DRO as giant caliper.  Using a razor blade against the toopost, mark stock and place in 3 jaw (assuming lathe) align razor to mark then advance saddle 15-30" depending on machine, new mark, repeat, this will be about as accurate as you can get.  At worst you are going to be within the scratch mark at that "temperature" you did support the material used to eliminate sag, at both ends.   You could also do this to check how accurate your tape measure is.   No really good way to do this in the home shop unless you have 60" worth of gauge blocks, and a true flat surface to place them on.

https://www.madmodder.net/index.php/topic,12110.msg143732.html#msg143732

[Joules]

Another thought occurred, the laser measures, once you eliminate the initial error, how consistent are they.  Is the 1mm error always 1mm, ± how much ?   I will try doing some measurement experiments and see if the laser exhibits any unusual results over 60 or so inches using the lathe carriage to move the laser.   Lots of assumptions on the internals of the laser pointing true, but look for phase shift errors in the electronics of the laser, i.e each mm differs by up to ± 0.2 or the error is consistent.   You may also be more accurate reading your 60" from a few meters away.

[Joules]

Far from the most rigorous lab setup for metrology 

Not accurate enough for Andrew, but some interesting results just the same.  I am using a very cheap ALDI laser measure, if I remember they quote something like ±2mm.   During the test I noted very good reading to reading results and was suspicious, damn right, if the reading doesn't vary by the error, they take the last reading.  So take a different distance and remeasure.  If the change is only a few mm's the error can be quite large, but point at another distance target and come back, I could consistently get to within 0.5mm.   Now you may note the Heath Robinson test setup and I haven't got space to try over a larger distance unless I clear some room round the mill.   You can certainly test how consistent a laser measure is over a set distance and use gauge, 1-2-3 blocks to offset the laser and check consistency.   I suspect the 1mm precision lasers should be able to do around 0.25mm accuracy.  A linear rail and bearing block with stops could improve this test a great deal.

Sorry for taking the thread off on a tangent (I believe thats what you are after) Andrew.


I should add, this accuracy was achieved only over this distance, it may vary better or worse at different distances, will try some targets laid out on the floor at some point, over say 3.5m 60" being 1524mm.  Probably far better using two metre steel rules end to end.

[awemawson]

Joules no problem with tangents !

I've been taking in all everyone has been saying and am very grateful for the contributions, but not actually 'doing' a lot recently - this whole business, which is taking me into diversions off detours via interruptions, and is itself a side shoot off the main picture is frankly a bit depressing. It supposed to be a quick 'let's make a gauge' one day job but there's a 'gotcha' at every corner it seems !

Never mind I won a HUGE (estimated 100 kg) crate of soft jaws for the Beaver TC20 lathe on ebay yesterday and my son (bless him) has just picked it up in Cirencester and is bringing it over at the weekend. Chap was asking a fortune, but in the end accepted a pittance (barely over the price of one set of blank jaws!) And I get two pallet sleeves that the wife uses as raised beds for veg in the polytunnel 

I'll try and get back to the grinder tomorrow !

[philf]

Andrew - What's wrong with turning a gauge on your CNC? If it's not going to be hardened, grinding seems a bit over the top - particularly when you consider how awkward it is to set up the grinder.  I was going to attempt to be amusing by suggesting you use an Oxymoron gauge to measure "approximately 60" accurately" but thought better of it!

[awemawson]

Phil, it’s a ‘chicken and the egg’ sort of thing.

Yes I could turn it on the Cnc lathe, but my only ‘standard’ is the recess in the back of the chuck that would be holding it. OK if it’s spot on first time but zero opportunity for adjusting later as the part has to be removed from the chuck to remove the chuck, and being a three jaw cannot be tweaked back into truth.

By grinding, not only do I get the far superior finish, but I can try the taper still on its arbor  in the removed chuck with blue, and remount the arbor spot on if I need to adjust, this way I can ‘creep up’ on engagement depth.

I’m pretty certain it doesn’t need to be as exact as I’m aiming for, but that’s just me being obsessive!

Also of course I’m trying to solve the ‘general case’ for future use of the grinder.

[awemawson]

So i was determined to push this a bit further today.

I realised that J&S have engraved a nice angle scale on the table for setting up, and presumably their engraving has some mathematical foundation. If I could measure the 'chord length' between angular graduations for a given angle I could calculate the radius !

OK due to vernier caliper limitations the largest angle I could reasonably measure was the  '10 degree included'  one, and plugging the figures into an online chord calculator gave me a radius length of 26.44 inches. To this needs to be added a further 1.31 inches to get to my place where I measure the displacement giving a radius of 27.75 inches

Is this a correct figure  I have no idea but it must be closer than the assumed previous thirty inches !

Meanwhile I got on grinding the parallel part of the taper gauge, covered over in the other thread.

[awemawson]

The plot thickens!

Perhaps foolishly I decided to do a rough and ready table swing radius test using my Disto D3 laser measure. I clamped some flat bars to either end of the table, and measured the distance between them with the laser. Then measuring 'end corrections' of 49.80 mm and 51.54 mm from the flat bars to where I'm actually measuring displacement and applying them I got . . .

29.46" radius

So I now have values of 30" (assumed) 27.75" (calculated) and 29.46" (sort of measured)

That's quite a spread!

So what is the implication of these figures - well the displacement that I need to set to achieve 7 degrees 7 minutes and 30 seconds is :

27.75 gives 3.4486 displacement
29.46 gives 3.6611 displacement
30.00 gives 3.7282 displacement

I think' the 27.75 value can be discounted - it's method of calculation is rather dubious

But that still leaves a displacement spread of 67 thou in a measurement that I try to set to tenths of a thou !

 . . .argh lets run round like lost sheep 

(oh two more of the blighters born this morning)

[vtsteam]

Andrew don't you have a manual lathe that you could cut the taper with?

[vtsteam]

Also another thing that I don't understand is, if the tapered part is only a few inches long why is it necessary to essentially use a 60" long standard?

Can't you make up a simple gauge say 12" long at the angle you want accurately enough to use that to set the table against?

Well as usual, not being there I'm probably not picturing the machine(s) and true situation, so forgive the naivete.

[vtsteam]

And one more thought, instead of grinding a straight test bar, can you grind a tapered one that through iterations can be blued and checked against the chuck until it fits, then substitute the "real" part and grind it at the same angular setting?

[awemawson]

Steve if you read the thread about making a taper gauge all will be revealed !

Anyway it’s now done.