Coax indicator design

[BillTodd]

This  coaxial indicator design is one I've been working on for a while. The images shown do not include fixings or screwed joints.

 Before I commit it to metal, anyone have any comments on the ball angle changer ?

[Rob.Wilson]

Hi And welcome Bill .


Great idea/design     ,,sure will take up allot less room with the DTI mounted that way ,,,,,,,,,,,,



Rob 

[BillTodd]

Thanks for the welcome Rob

less room with the DTI mounted that way

That's the whole reason for making it. My little Haighton hasn't much room below the vertical head. I wouldn't be able to get a Blake type in there.

However, I can't take credit for the idea, it was inspired by this clever design by PM member Wes43:

http://home.comcast.net/~nonsense91/SALE/Coax%20Ind.JPG

judging by the pin to the top/left of the DTI, it uses a lever to turn the movement 90 degrees.

Bill

[Bogstandard]

Bill,

I use one of the normal ones, and one of it's major faults is the length. In fact, I have made shorter probes to allow me to use it with my RT with the chuck fitted.

I suppose the angled design actually shortens it by over one inch, which has to be good.


Bogs

[Rob.Wilson]

Hi Bill

Great idea all the same ,,, thanks for showing    ,,,,,,,,,, must add it to the list  .

Impressed with your cad drawings     ,, will you be producing some drawings 


Rob

[BillTodd]

will you be producing some drawings

Yes, assuming it works. ;)

I was expecting (still am) all sorts of criticisms of, and objections to, the ball angle changer

Hi John,

I don't like the idea of getting my hands anywhere near a spinning probe, so I'm wondering about a simple torque limit clutch between the drive and the rocker arm. Does your Blake device have anything like this?

I have also been thinking about a slow speed drive mechanism (which would also double as a clutch), but I'm not sure how to keep every thing concentric. Worth the effort?

Bill

[Bogstandard]

Bill,

I wouldn't bother with the clutch idea, it has a sort of safety feature because the probe isn't mounted solidly, a bit like a probe on a DTI, if any force at all is put on it, it just moves away.

What isn't there can't go wrong.

I don't know how fast you expect to run it, as when I run mine, having electronic speed control, I run it at around 20-30 RPM, otherwise I find my eyes can't keep up with the bouncing needle. I suppose a sign of getting old.
So maybe a speed reducer would be advantageous to people who can't get their machines to go so low, but in all honesty, it would be a pig to do while retaining concentricity, plus also it would be just liable to increase the overall length, a thing your version scores on.

People have said that they aren't very accurate, but in all honesty, in the time I have been using mine, I could never get anywhere as near to true as the coax gets it, plus the ease and speed of use is fantastic. I used to hate setting up to find concentricity, and started to grow a 2 foot long rubber neck to prove it (just joking of course, it was only 18") but now I really enjoy the complete control I have over the process.

I think Stew has now invested in one, and I think he has the same sort of feedback as myself.


John

[BillTodd]

I don't know how fast you expect to run it, as when I run mine, having electronic speed control, I run it at around 20-30 RPM, otherwise I find my eyes can't keep up with the bouncing needle. I suppose a sign of getting old.
So maybe a speed reducer would be advantageous to people who can't get their machines to go so low, but in all honesty, it would be a pig to do while retaining concentricity,

All good points.  My drill/mill requires a belt swap (a right PITA), but I can slow the Haighton with its VFD - you're right no slowmo drive.

plus also it would be just liable to increase the overall length, a thing your version scores on.

Speaking of which... I've removed the unnecessary top cap (a legacy from an earlier design) which has shaved another 8mm off the height. I think I should save another couple of mm by re-thinking the rocker arm/probe holder.

[RichardShute]

This  coaxial indicator design is one I've been working on for a while. The images shown do not include fixings or screwed joints.

 Before I commit it to metal, anyone have any comments on the ball angle changer ?

I can imagine a couple of aspects which might be less than ideal - I know, 'misery guts'.

The DTI is operated by riding up/down a conical surface so any radial play between the outer body/DTI mount and the inner shaft will give rise to a false reading. If you were to arrange the DTI to be actuated via a bell crank resting on a flat disc that potential error is removed and you may even be able shorten the whole assembly slightly.
Also, as shown, the DTI (obviously) does not rotate, but it does move up and down to give the reading. You will presumably hold the outer body by hand to stop it rotating and thereby have a very fair chance of inadvertantly raising or lowering it. By re-arranging it a little so that the DTI remained fixed vertically this potential error would be removed.

As it stands, I expect it will probably perform adequately, with care in use, but I think you can still improve the design. You have come this far, why not polish it a little more?

Richard

[BillTodd]

I can imagine a couple of aspects which might be less than ideal - I know, 'misery guts'.

Not at all Richard

The problem with pixel machining is that everything fits perfectly there's no wobble, no TIR or friction. Without someone to look over ones shoulder it is all too easy to miss the pit-falls.

Here's a perfect example:

any radial play between the outer body/DTI mount and the inner shaft will give rise to a false reading.

I had considered radial run-out of the shaft,  but assumed I could turn all the bearing surfaces at the same time, so keeping any run-out to a minimum.

I had however, completely ignored body-wobble as the upward force from the rocker moves around the bottom face (cheers Richard )

I can see a couple of possible ways to minimise the effect without changing the design:

1)The bearing surfaces have to be a very good sliding fit, so I could make them tight then lap them to a good fit.

2)The stay arm (that prevents the body turning) could be hung off of the same side as the DTI thus moving the centre of gravity out towards the DTI. This should apply a (vertical) torque to the body, pre-loading the top and bottom bearings and keeping them snuggled up against the shaft.

...arrange the DTI to be actuated via a bell crank resting on a flat disc

I looked at bell-cranks or levers (as used in Wes43's example above), but couldn't come up with a neat, compact arrangement. A lever would offer the possibility of some 'gain',  to compensate for losses at the rocker/probe end. Wouldn't it need a gimbal of some sort to remove all radial errors ?
 

By re-arranging it a little so that the DTI remained fixed vertically this potential error would be removed.

If the DTI/ body is fixed vertically then the problem of wobble & smooth parallel motion perpendicular to the shaft shifts to the rocker follower (IYSWIM) . I couldn't figure how to do fix that.  Had you something in mind?

Bill

[kwackers]

Just to throw something completely different into the mix.

I was contemplating making something like this and so I'm fairly interested in the issues and design.
What I considered doing though to make the whole thing small was to get a magnet to move and use a coil as a pickup (or the other way round). This could be used to drive a small meter. It auto zeroes when not moving (since we're only interested in the 'wobble') and could in theory provide the 'multiplication' required to overcome the mechanical 'loss' particularly when long arms are used.

It's not got much further than an idea, but I thought I'd bung it up here in case it had legs or could give someone a nudge in another direction.

[ken572]

Bill,

That is a really cool idea/concept. Time to build a proto-type,
and make a video clip of it in test use. 

Ken.

[picclock]

I like the magnet idea.

A multipole magnet with north and south horizontal close to a stereo cassette recorder read head might be an easy way to implement it. Just tuning for minimum output would work. By connecting the heads out of phase the balance between north/south fields would be very precise.

Somthing to think about 

picclock

 

[BillTodd]

Just to throw something completely different into the mix.

I was contemplating making something like this and so I'm fairly interested in the issues and design.
What I considered doing though to make the whole thing small was to get a magnet to move and use a coil as a pickup (or the other way round). This could be used to drive a small meter. It auto zeroes when not moving (since we're only interested in the 'wobble') and could in theory provide the 'multiplication' required to overcome the mechanical 'loss' particularly when long arms are used.

It's not got much further than an idea, but I thought I'd bung it up here in case it had legs or could give someone a nudge in another direction.

I had wondered about using a magnet to couple the rotating vertical movement of the rocker to a stationary lever but rejected it as too compliant.

I like the magnet idea.

A multipole magnet with north and south horizontal close to a stereo cassette recorder read head might be an easy way to implement it. Just tuning for minimum output would work. By connecting the heads out of phase the balance between north/south fields would be very precise.

Somthing to think abou

Yes.

The snag with all magnetic sensors like coils is their sensitivity to stray fields (hum from motors or just the rotating mass of iron coupling to any static magnetic field. I suppose if the parts could be shielded it might work

Or (getting excited now :LOL:) if there were four coils, stationary and located on the X  and Y axis.  With two magnets on the rocker, as they passed the coils, each magnet would induce a different current in the two coils (varying with distance). it would be possible to synchronously rectify the outputs of the two pairs of coils to give an X & Y error read out. Thoughts?

[Bluechip]

Some more 'orrible thoughts ... if you want to convert a wobble to a indicator ( meter ? ), you're making a seismometer.

Some 30 yrs ago I made ' variable capacitor / pll / amp / meter contraption that actually worked.

The cap was two bits of PCB, one bit stationary, t'other attached to a weighted beam. When it moved, capacitance varied, osc. freq. likewise, pll outputs a voltage proportional to freq. and so to the amp.

Maybe easier to screen than inductors/ magnets. Maybe not.

Dave BC

PS .... don't ask how I configured the pll, 'cos I can't remember. Only time I've ever used one   

It was not a 4046 IIRC. NE xxxx ?? But it is probably possible with a 4046. Suppose they're much the same.




 

[kwackers]

I was thinking of a low impedance coil, i.e. just a few turns around the (ali?) body. It would have very little susceptibility to mains hums and stray magnetic fields (I think) would be too low by the time they got there, it could also (probably) drive say a 50uA meter directly.

[RichardShute]

Just to throw something completely different into the mix.

I was contemplating making something like this and so I'm fairly interested in the issues and design.
What I considered doing though to make the whole thing small was to get a magnet to move and use a coil as a pickup (or the other way round). This could be used to drive a small meter. It auto zeroes when not moving (since we're only interested in the 'wobble') and could in theory provide the 'multiplication' required to overcome the mechanical 'loss' particularly when long arms are used.

The problem with that is that as you say you get no output when the magnet is not moving and with the relative speed being ineviably very slow you will get negligible output. You'd be much better using a LVDT (linear variable differential transformer) which are readily available to buy (for *real* money) or could be made. It is basically two colis on a common former with a magnetically susceptible slug up the middle attached to the probe. You apply a signal to one coil and compare it with the other. They can be easily accurate to a micron!

Bill
I have some ideas for a reasonably non-rocking slide scheme, I'll try to do some sketches shortly.... Not got time just now - off out to pickle some brain cells.

Richard

[AdeV]

The downside to all this coil malarkey, as I see it, is you're going to need to introduce circuitry and batteries 'n' stuff. But, if you're going to do that, wouldn't it be easier to use a hall sensor? As the magnet moves up & down in the unit, its distance from the sensor will vary, and (in theory at least) you can then read off the distance between magnet & sensor as a voltage output. This also means you can take a reading while everything's stationary, should you so desire. I don't know how easy/difficult it would be to use 2 hall sensors to get X & Y readings...

[picclock]

@ kwackers
even 3000rpm is only 50 hz which is a quite a low frequency for what you are suggesting, hence my suggestion of multiple poles to increase the frequency. At 100 rpm you would need some serious core material - which would make it quite large.

@billtod
at 50 rpm (just under 1 Hz) the frequency of wobble would be so low that filtering out the other induced interferance would be simple. Additionally, if it was encased the case would act as a faraday shield. But .. .

@bluechip
PLL is probably the best approach. Change of capacitance would be very small though and overall capacitance would be low, so high frequency needed. 4046 sounds just the job. Power it and a moving coil meter with 2 or 3 button cells, add some form of off on switch and it should be good to go.

It's quite a good idea that ought to succeed.

Just needs a reasonable mechanical design.

One for the 'tuit' list.

picclock


 

[kwackers]

@Picclock, yeah I always forget rpm and hz aren't interchangeable! That means my simple version is a non-starter...

[BillTodd]

Well Today I turned up a few of the parts....

The shaft and DTI holder are 316 stainless, the body is in an unidentified bronze. I'm pleased with the fit of the shaft, but I'll have to drill an air hole in the middle 'cos the thing acts like a piston pump ATM.

Tomorrow, I'll fire-up the milling machine.

[BillTodd]

What's that saying about best laid plans...

I was only able to assemble the the parts I had already made. The mechanism seems to work OK, I get about 0.12" of indicator movement between the top and bottom positions of the shaft.

[Brass_Machine]

The turning work looks good. Anxious to see how this works out for you.

Eric

[RichardShute]

I was only able to assemble the the parts I had already made. The mechanism seems to work OK, I get about 0.12" of indicator movement between the top and bottom positions of the shaft.

Looks like you are making a nice job of it Bob. Sorry I didn't have time to offer my suggestions over the weekend, I got bogged down in domestic drudgery... and now you are well on the way with your own design, I don't want to distract you from your purpose.

I'll watch with interest.
Rgds
Richard

[BillTodd]

The turning work looks good. Anxious to see how this works out for you.

Yes, my Hardinge makes it look like I know what I'm doing  

Looks like you are making a nice job of it Bob.

I'm sure Bob will make a far better job of it than I ever will

...I don't want to distract you from your purpose.

Don't let that stop you ;) I already have ideas for a second version, I'd love to see what you had in mind.

Bill