Sine Blocks and Cosine Error on Test Indicators

[CalgaryPT]

I always wondered about this but never fully understood the implications for a test indicator—I guess because I do mostly fabrication and not machining.

I like some of the information in This Old Tony videos, but I find his attempts at being humorous all the time distracting.

Anyways, here he explains why test indicators all have a specific angle they are intended to be used at, and how changing the stylus can affect your results. Also he explains how to correct for cosine error if you know your indicator's operating angle. The whole video is here, but the part about test indicators starts at 26:14:

 

[Dabbler]

I watched Bert use test indicators and asked him about SIN error. His reply was to dismiss SIN error, but he had an interesting reason: Since a test indicator is used to make RELATIVE measurements, it affects the total deflection, whereas he is always working to null the deflection anyway.

I use test indicators so that I don't get SIN error, but I thought his point of view was interesting!

 

[CalgaryPT]

I wondered about that too. That’s what got me interested in the vid. I rarely use one myself but wondered why it would be an issue as I can’t recall a time when I used one for anything other than relative measurements. I just assumed there were instances where true machinists (I.e., not me) used them for measurements from a known distance. Interesting eh?

 

[Dabbler]

When measuring on a surface plate you can use a test indicator to measure precise distances with far more accuracy than any other method. You build a gauge block stack and measure the item and the stack and compare. Measuring of parts of a tenth is possible this way.

 

[PeterT]

You can get angular (most say cosine) error with dial test indicators too. Comparing from a datum or gage block stack is easy & reliable. The point is dont believe the relative gap value & minimize error by orientating the arm correctly
https://store.gaging.com/blog/whats-wrong-with-my-test-indicator-a-cosine-error-discussion.html

But there are instances when you need to take absolute measurements with a dial indicator. Maybe you have one set on a lathe or mill measuring travel of a table or a quill. Now there is no convenient reference datum analogous to the granite plate surface. You are hoping that it is measuring displacement accurately once the dial has been zeroed or nulled. Sine error is just a way of saying the dial plunger axis must be correctly aligned to the relative surface displacement axis, otherwise it is measuring the hypotenuse of an imaginary triangle. His (exaggerated angle) graphic pretty much tells the story.

So lets say the block of metal in his picture represents the end of your lathe carriage or a stop on a quill. If you want to measure displacement like 0-1 inch, you wont find a dial TEST indicator in that range unless you want to bring gage blocks into the picture. So now you have to be aware that the reading on the indicator dial can only be accurate if the plunger axis is coincident with the movement or displacement axis.

Another example, imagine that the dial is attached & locked to an arm being held in the drill chuck & we are measuring table plane tilt (or mill head angle & nod). Now it matters less that the dial indicator is canted slightly off axis because we are looking for the needle to remain stationary in one position assuming it is a spherical ball tip. Actually sine error is helping a bit here - its exaggerating the runout, but again, we just cant trust the runout number. But here we are using a dial indicator more like a TEST indicator looking for relative movement.

 

[PeterT]

Sidenote - another thing you hear about is indicators giving false readings when moved from upright to inverted. I cant find my pictures but they are in post #52
https://www.hobby-machinist.com/thr...-tailstock-need-help.77404/page-2#post-660884

The mag stand / arm mount assembly can be significantly influenced by gravity, even the good ones. I have read tests where dial indicators (alone) can also drift when inverted. I think TEST indicators less so but still measurable. This may seem like useless trivia but a common example might be mounting a DTI to an extended arm held in our lathe chuck & rotate it around the tailstock quill barrel or tool looking for concentricity. But we cant trust the 12 & 6 o'clock relative readings as they are most influenced by gravity.

 

[PeterT]

From Richard King

 

[Dabbler]

I have the utmost respect for Richard, however, he is showing a thing and calling it something else... Any assembly will be subject to gravity. However, an indicator (especially the Mitutoyo .00005 one he shows there) is compensated already. He is demonstrating the sag of the arm assembly.

I have taken my Mitutoyo .0001 test indicator, and chucked it into my lathe, and rotated it over the entire circle, and it did not move - not even a fraction of the indicator arm. Try it youself and check your indicator. I haven't checked my newly-acquired-used Interrapid .00001 yet.... Hmm. I should. right away.

He is demonstrating a real thing here: he is confusing it by blaming the indicator.

 

[PeterT]

At 0:40 he says "all this stuff sags" & waves his hand over the mag arm assembly but you're right he never distinguishes the drift of arm vs. indicator alone or tests the indicator alone. I did the same thing you did & could not detect discernible needle movement upright to inverted on my (Mitutoyo) DTI. I have not tried with dial indicator but now I'm curious. It has a heavier stem assembly but maybe there is some kind of compensation inside?

My DTI kit came with a solid arm & some dovetail rod stems. I use them whenever I can because they are more rigid. But their range & orientation is limited. Arms are kind of an evil synergy - in many instances we require it to reach into places. The bottom line is that indicator arms (even the good ones) are subject to drift under gravity in the thou range we are looking to measure so don't believe the up vs. down numbers. The longer the arm is outreached, the more exaggerated the reading. The heavier the gauge at the end of the arm (ie. dial indicator vs DTI) the more error.

 

[CalgaryPT]

Humm.....interesting video, but it strikes me that you'd have (at least) three variables here: the indicator stylus, the stand holding the indicator and the base plate. How do you know the sag isn't a function of the stand or the base plate v. the indicator? If you remember the feather and the ball inside a vacuum experiment from high school physics, all are affected equally by gravity. From a scientific method perspective you'd want to eliminate the base plate and the stand as variables. If the indicator were attached to the base plate w/o the stand you'd have a more accurate demo methinks, although you'd still have the compound effect of the baseplate. Nonetheless, I get the point he is making and agree there is an effect, irrespective of which variables are causing or compounding it. Very cool. Thanks for posting. I love this stuff.

 

[Dabbler]

He is pointing at a thing and saying 'it sags' No real explanation of what or why this is happening. This , I understand from those who have experienced it, is much like his courses. "Do it this way" is the theme without a good reasons why. Great guy, lots of knowledge. Just not a natural teacher and it shows.

The 2 1/2 tenths the indicator shows is more likely due to the arm than the granite plate. I'm very sure the indicator is just fine in all orientations (for good brands like Mitutoyo and Interrapid, Starret, etc)

***************

This is important to me because I teach how to align the tailstock by chucking up an indicator and chasing the barrel of the tailstock in two places. If it were an indicator sag problem I'd show them wrong every time!

 

[PeterT]

That's kind of why I like the centered dumbell alignment bars like the one made by Edge. Wherever your DTI indicates on the HS ring should match the TS ring by traversing the carriage with DTI on board. So for example you zero DTI on HS ring at 12:00 then traverse the carriage & measure the TS ring at same 12:00 position. No gravity effects on the arm, nothing rotates. Lets say this indicates the TS is 0.001" high. To confirm you re-adjust indicator so DTI indicator ball on the underside at 6:00, repeat. The TS should confirm read same amount. Measuring TS lateral displacement is the same procedure with DTI ball on 9:00 position of rings.

Note this procedure may be an iteration if TS is out in 2 dimensions. For example if the TS is say 0.003" inward, then the DTI test across the HS ring top at 12:00 position wont match the TS 12:00 because the TS ring circle is displaced slightly, the ball will rest on one side or another.

 

[PeterT]

Or..... meet the BIG KAISER! lol

 

[CalgaryPT]

Or..... meet the BIG KAISER! lol

Holy cow. Now that's a tool.

 

[Dabbler]

Or..... meet the BIG KAISER! lol

wow. just wow.