Mill spindle runout

I think with the arm set tangential to the surface it's trying to measure the runout will be a lot less. I'd do the test again

Edited By duncan webster on 10/09/2022 18:11:20

A DTI is not an absolute measuring device. It is a comparator. A dial gauge actually provides a proper reading.

This page seems to suggest something approaching the opposite of that:

"When it comes to measuring, the dial indicator excels in repetitive and comparative measurements. While the test indicator excels in a consistency measurement."

To be frank, I'm as puzzled by the distinction given on that web page as I am by yours.

As far as I understand it, both a dial indicator [or gauge] and a DTI can give actual or "proper" readings [indicated by the amount of travel of a needle (usually) round a dial] and both can function, as often as needed, as comparators.

I hope you, or someone, can clear up my confusion.

Bill,

It is my belief that the distinction is entirely artificial: Both instruments are more logically classed as Indicators; unless and until they are mounted in such a way that there is a Zero reference and the ‘system’ is calibrated.

Even then, the dial mechanism and the [likely] cosine error both serve to make them poor Gauges.

The use for which both were originally devised was that of a Comparator … a reference point is established, and the deviation from that is indicated rather than truly measured … The uncertainty can be high at any point other than a return to that reference.

Semantics make this a tricky concept to grasp, but that’s my best effort at this time in the morning !

MichaelG.

.

NPL Good Practice Guide No. 11 is a useful essay on Uncertainty of Measurement

Downloadable here:

https://www.npl.co.uk/special-pages/guides/mgpg11.aspx?ext=.

Edited By Michael Gilligan on 11/09/2022 07:11:52

DTI's and plunger-style dial indicators can both act as comparators and both take direct measurements.

One of the most common mistakes made whilst taking direct measurents with a DTI is the one shown in Steve's video (sorry Steve) where the lever is not tangential (or parallel, if it's a flat surface) to the work. Steve's is a particularly good example because the lever is so far from tangential the cosine error will be large. In the video it is showing 3 thou TIR but if the DTI were moved down so the stylus was parallel to the floor that reading would probably be less than 0.002".

Other things which can cause false readings are:

Using the instrument held rotated in the holder so that the lever is not swinging perpendicular to the axis of the shaft being rotated. Always make sure the instument is held perpendicular to the part.

Using the instrument with a different length stylus than it was designed to use. Fit a longer stylus, it will under-read. A shorter one will over-read.

Plunger type dial gauges will also over-read if the stem is not perpendicular to the work, though the error is usually very small. If you're taking runout measurements as in the video it must also be reading on the centreline of the part, to again avoid cosine errors.

Just throwing this pear-shaped pebble into the pond: **LINK**

https://mqs.co.uk/verdict-109-02610-v61-pear-shaped-stylus-length-0-770-for-verdict-indicators-metrinch-1-t8-and-t4b.html

MichaelG.

.

https://worldwide.espacenet.com/patent/search?q=pn%3DGB930287A

Edited By Michael Gilligan on 11/09/2022 08:54:56

Yes the pear shaped stylus is designed to reduce cosine error but it won't prevent the over-reading error that Steve's setup would incur.

Steve, looking at this video:

*LINK*

Since you are measuring directly over the vee block this should be showing no runout. Possible causes are:

• Bearing diameter out of round
• Burr or ding on the egde of the bearing register causing it to ride up on the vee block. You can eliminate this by putting slip gauges in the vee blocks to raise the spindle off the wider vee so it doesn't get affcted by any edge ding.
• Bent spindle
• Top bearing register not concentric with bottom bearing register, or it has a ding/burr.

Another good check you could do is put the vee block under the chuck register then use your DTI to check the taper and the bearing register. If the taper runs true but the bearing register does not, one solution could be to knurl the register to raise the diameter a bit then turn or grind it back down to the bearing diameter, whilst supporting the spindle in a lathe between a 4-jaw and a fixed steady on the chuck register. It won't be perfect but it'll get your TIR down to a couple of tenths rather than a couple of thou.

That's my understanding too - relying on gears limits this type of instrument's suitability for taking absolute measurements. To be repeatedly accurate the gears would have to be perfectly round, with perfectly shaped and spaced teeth, running in perfect bearings, with zero backlash. Such gearboxes are difficult and expensive to make, and highly prone to error due to wear. In practice though, geared instruments make excellent indicators because they're able to show small movements with considerable sensitivity and fit-for-purpose accuracy. But they become less reliable with increasing distance. Lever indicators are 'better' because they measure a smaller range than plunge types, and thus have less opportunity to accumulate error. The shortcomings show up when several comparators are used to take the same measurement; my dial is almost certainly different to yours!

Digital Calipers and DROs take a different approach; their accuracy depends on a 'scale' of microscopic proportions and a differential pair of encoders that detect capacitive, magnetic or optical changes. The mechanics are very simple, not prone to wear, and - although conceptually difficult - so are the electronics.

Most micrometers are based on a precision thread. The threads aren't made on a lathe or with a die because these methods are far too coarse; instead they're ground using specialist techniques. Despite the cost and difficulty of making micrometer threads, they're much easier than accurate gear-boxes and there's not much to go wrong when they're fitted in a micrometer. However, they can't be blindly trusted because micrometer threads and anvils wear out and frames bend if the instrument is dropped. For a micrometer to be considered accurate, it either has to be new, or recently recalibrated. In my humble opinion chaps who buy elderly second-hand 'quality' micrometers aren't buying accuracy. For rough work they can be checked with slip gauges, but here again I suspect too much faith is placed in ebay bargains!

In my workshop none of this matters much. My machines and measuring equipment all work to about ±0.02mm which is good enough for most of my purposes. I mostly work by 'fitting', whereby parts are combined by gauging and adjusting them relative to each other, not by absolute measurements. In this type of construction a certain amount of slackness is often helpful.

Dave

Dial gauges are the easiest thing in the world to check for reading accuracy. Pass a slip under the stylus or anvil, record the reading, pass another slightly different thickness slip under it. If the dial records the difference accurately the instrument is reading correct.

The accuracy of lever type indicators is less important than their ability to detect runout. For accuracy in actual measurement, the plunger type is designed to be fixed to a dedicated stand sitting on a surface table, and easy to check using slip gauges. With the lever type, each model is designed to work wilh a compatible length arm at right angles to the workpiece. As the angle changes, small errors accumulate. That has no importance if, for instance you are adjusting round work in an independent 4 jaw chuck, the final zero needle movement is all you are aiming for.

An instance of a dial indicator not being accurate in measurement, but functioning perfectly well is the co-axial type indicator used in milling for centering a bore with the spindle. The numbers on the dial are meaningless as there are a number of different types and lengths of lever probes and angles set used with one instrument, which mean the movement bears very little relationship to the pointer movement. Since the object is normally to attain zero pointer movement, that doesn't matter.

Slips are unlikely to be out by any measurable amount. If they are still shiny, they will very likely be good. If you really wanted to be anal then do the check twice or thrice with different sizes of slip. The chances that several slips will all be worn by a similar amount are infinitesimal.

The slip gauges at the museum would all fail testing even to workshop grade. A stack of 4 would probably be 0.0001" undersize, which is plenty good enough for our requirements. When I got the box, I had to remove a thin film of rust from most of them, with a scraper and oil, not abrasives. A new Stanley knife blade works well and won't scratch the surface if used carefully.

Thank you particularly to Michael, Pete and Dave for enlightening me.

Thanks, Bill, for enlightening you.

Slip gauges are likely perfectly adequate, for most of us, if they can be wrung together. Likely good enough, even when they won’t, for most applications as long as they are not actually rust stained.

Thanks for the example of a category mistake.

Our slips will never wring together, but rust is a danger as it stands proud of the surface and must be scraped gently off. The tiny pits left would be bad if they covered most of the surface, but ours are about 10%, fortunately. While I have instruments that can measure down to 1 micron, in real life, I would be content with 10 microns at best for actual work and 50 generally.

Gosh, a lot of posts since I last looked. I put some time into it at the weekend, including refurbing a tenths dial gauge I have to put the cosine error issue to bed. But alas, every time I measured I got different results. I will post a bunch of videos in the morning to illustrate.

after a very frustrating few hours I got fed up and decided to make myself a smaller scraper to go with the big one I bought from eBay. Big one, 60 quid. Small one, nothing.

DW Spindle Video

See latest video…..

Calibration of clock gauge

Runout on spindle

Runout on nose

Runout on test bar and drill adapter

Binding of spindle in bearings.

So it looks like 2 problems:

1) Not very well made spindle, more runout than would usually be acceptable

2) Bearing issue, binding up when only finger tightness is applied to the preload nut.

”Change the bearings” everyone will say… but, is that the cause? One bearing is fairly loose and the other is very difficult to get in or out. Could the quill not be straight? The bearing seats misaligned? how could I tell? They run smoothly with no preload. New bearings = another £80!

Any advice appreciated.

Steve

Edited By Steve355 on 13/09/2022 10:00:35