3d indicator design

[BillTodd]

Add my standing Ovation to a super presentation of a wonderful design.

Thanks   

The thrust washers and bearing balls have arrived, and I finished re-painting the kitchen so, hopefully, I'll get a start on the prototype this weekend.

I've decided to screw the probe into the with a M4 thread since this suits commonly available  Renishaw probes (and I happen to have one here) . I still have yet to come up with a simple locking mechanism so a tommy-bar will have to do.

Bill

[BillTodd]

Interruptions, interruptions... 

only managed to get a few bits made.  Not too happy about the finish on the mounting shaft, sizes are OK but  too many rpms on the finishing pass left banded marks.

The 45° cone was easier to cut that I expected , I was able to use a tiny boring bar to enlarge and finish after a an initial cut with a 45° spotting drill.

More tomorrow...

[BillTodd]

Made the upper taper and lapped  it into the main shaft. The two cones work as expected

The two ball cages were a bit of a disaster - I forgot to change the design (PCD of the ball holes) when changed it to use larger thrust washers    so the balls fell out when assembled 

Also, real world tolerances of the thrust washers  (20.7 diameter instead of 21 with a 0.2 bevel on the edges etc.) mean that the over-lap between the washers is smaller than expected, leaving the slide a little short of travel (I am aiming at 4mm in all directions). So, although I had hoped to use standard parts, I may have to make the thrust washers myself

Bill

[BillTodd]

Made the larger thrust washers from stainless steel (after three attempts!) - just need to lap the surface to a good finish.

Made the body from a piece of 40mm aluminium bar.

I still have to make the correct sized  bearing cages and the DTI mount.

[BillTodd]

Made the DTI mount and pressed it into the body, then spent all afternoon and most of this evening wondering why the mount hole is in the wrong place...

..slightly ironically,  given that I'm making and edge detector,  it's because I forgot to subtract 2mm from the DRO when I set the edge of the vice with the (4mm) edge detector . 

Never mind, I'll just have to bore the body 2mm deeper and reposition the cir-clip groove.

[DaveH]

Damn clever Bill 
 
DaveH

[BillTodd]

 

To save making two of the thrust washers, I've redesigned the slide so that the slide-tube and nut do the job (they're very lightly loaded so don't really need to be hardened). I've also found some SS mudguard washers in suitable sizes (16x50)  which will avoid having to make any more bloody washers!.

 

[BillTodd]

I found the time this afternoon to fix my cock-up and make the two new ball cages, so now it is almost finished.

However, I've discovered a snag with the cir-clip idea; The large thrust washer has to be held against the step in the main body if the zero reference it to be stable. The cir-clip does not hold the washer tight enough (there's no room for a spring) .

To solve that I've redesign the body to have a screwed in locking ring at the bottom. As a simpler alternative, it would be also possible to screw the washer in place with a half-a-dozen small screws.

The next problem is because everything is tight with minimal clearances, I keep getting random jams - one minute it'll slide nicely, next it'll snag and jam  - a major PITA 

[Brass_Machine]

Fast work Bill.

How do you plan on dealing with the random jams? Will you use any type of lubricant or keep it dry? I was thinking maybe one of the NanoLubes that are always up on ebay? Dry? maybe time for some uberpolishing or will that lose too much tolerance?

Eric

[BillTodd]

Hi Eric,

Well, your polishing idea has helped a lot   (it also showed me I had a slight eccentricity in the upper taper part ) Although, I think the main problem is a poor choice of materials (aluminium is too reactive to make a good sliding material).


It is now basically working (apart from the loose main thrust washer). I'm getting about 3.5mm movement in all directions.

However, I have spotted a fundamental flaw in the design; Most of the play in the sliding parts can be removed or ignored (i.e. the ball-slide can be tight and it really does not matter if the upper taper is a bit sloppy because it is biased one direction by the forces on it). 

But, the bottom taper has to slide perfectly in the vertical direction without side play or the centre 'zero' position will be inaccurate 

Back to the drawing board...

[BillTodd]

[andyf]

Bill,

I can see the value of this device for edge-finding in X and Y, in that it enables a tool held in the spindle of a miller to be accurately located in either axis in relation to the the edge of the work, or an edge to be accurately positioned to lie along one or other axis.

But do you anticipate any use for it in the Z axis, apart from levelling the top surface of a workpiece, or tramming a milling vice? Either of those functions could be accomplished equally well with a dial indicator held vertically in the spindle.

I'm not trying to be a wet blanket; it's just that I have the feeling that I may be missing the point where Z is concerned.

Andy

[BillTodd]

The idea is that such a device  provides a zero reference for all axes

 I.E. One would fix the device into a holder at a known distance from the Z reference point (all the tools & holders should be set to the Z reference) then zeroing all axes can be achieved with the one device.

[andyf]

Thanks, Bill.

I'll have to stay with cigarette paper. I have no reliable way of setting tools to project from collets by a distance calculated from a datum which is derived from a gauge like the one you're making 

Andy

[DaveH]

Bill,

Nice video.

Seems to work very well 
 
DaveH

[BillTodd]

But do you anticipate any use for it in the Z axis, apart from levelling the top surface of a workpiece, or tramming a milling vice?

That's a good point  Andy . One quick fix for the wobbly vertical slide would be to remove it - i.e. fix the probe in to the XY-plane slide and just use it as a 2D probe 

[BillTodd]

Seems to work very well

Thanks Dave - it is, but just not quite well enough. For it to measure accurately, it has to move smoothly - it does not at the moment  

[PekkaNF]

I'm not compentent on this issue, but my first advice is not to give up. Often when all looks lost, answer is just around corner. Has happend to me many times. On some aplications I see that there is an accurate hard surface as a reference and then a member is pushed against it. Really hard to get many close fitting parts work together.

Pekka

[BillTodd]

my first advice is not to give up

Thanks for the encouragement 

Adding a threaded lock nut has fixed the zeroing problem and makes the whole thing feel better (although I won't be using a 40tpi thread on the next one - way too fine and far too easy to cross-thread !)

Thinking about it some more; The play in the vertical slide only messes up the initial zero measurement  i.e. the rest position will not be exactly 2mm from the centre of a 4mm probe. However, as long as the play is consistent in all directions (which it should be), the edge position should be reliable.

Next , I'll make a short 4mm diameter probe  and test it.

[BillTodd]

Ok, a little more progress...

I've replaced the ballslide's  brass tube and washer design with a steel 'top-hat' & nut that runs directly on the balls.

The fit between the new brass bottom taper part and the ball-slide tube is now ten times better, and enlarged the top of the taper is less inclined to tip. All this has increased my confidence that this thing might work

I just have to tweak the depth of the shoulder in the body to suit the new parts and it should be ready to test properly 

[BillTodd]

 

Well it kinda works...

To test it I set it up in the 4 jaw on my lathe so I could access the accuracy using the lathe's Sony DRO (good to a few microns)

I zeroed the dial then used the DRO's touch sensor to set the DRO to zero for the 'depth' (Z axis) test. Then wound the carriage in by 2.000mm (see test1.jpg)  . The dial reading was about 50um short. This was not unexpected as I don't know how accurate the DTI is or how accurate my 45° tapers are, and absolute accuracy it is not too important for an edge detector.

Then I zeroed the DTI (at 2mm in) to test the XY plane using the cross-slide (see test1a.jog). I zeroed the cross-slide DRO just as the tool post hit the probe (couldn't get the touch sensor in a position to test this so there could be a small error) . Then moved the cross-slide in 2.000mm. This time the DTI read 80um short (meaning there is about 80um movement in the X+ direction, i.e towards the indicator, before the DTI moves) (pic test 2a.jpg).

I re-zeroed the DTI and tested it in the three other orientations (X-, Y+, Y-) see test2.jpg.  It seemed reasonably consistent  perhaps 20um out in the Y- direction (could be chucking run-out  or a gravity effect?).

All of which makes me wonder if this is ever going to be accurate enough to be useful  ?????

There is only the tiniest play in the ball slide (I chose to bond the parts together this time instead of screwing them) and the fit of the other parts would be difficult for me to improve on, so I'm now not sure if I can get this to work 

[Fredbare]

I have have been following this from the beginning, and admire your skill,  and tenacity, hope you succeed, and finish this  to your high standards.

John

[Divided he ad]

Bill,

Sounds like you're in for a bit of "fun" testing!?

I'm sure you'll come up with something.... even if it's a good hard launch at the wall!   


Nah. You'll get it 





Ralph.

[BillTodd]

I'm tempted to go back to my original prototype idea, which was a tipping beam. It is inherently non-linear and needs careful scaling to be anywhere near correct, but still it might be better than what I've got ATM 

[PekkaNF]

I'm tempted to go back to my original prototype idea, which was a tipping beam. It is inherently non-linear and needs careful scaling to be anywhere near correct, but still it might be better than what I've got ATM 

Three prongs/legs? Or a cunning plan to get all level and square?

PekkaNF