DTI

[Graham Williams 5Participant @grahamwilliams5]

[Graham Williams 5Participant @grahamwilliams5]

Just fitted a magnetic scale DRO to my Tom Senior M1, sourced from EMS, who were great, more helpful than any other of the suppliers. Checked the Y axis read out against the machine dial and the difference was within 4 tenths at worst position, could have been the alignment between scale markings I suppose. The puzzle was that I had a DTI on the table at the same time and over 100 thou the DTI was reading 107 thou . Now the DTI wasn't an expensive one, sourced from a well known supplier a while back but that reading threw me, am I expecting to much of a DTI over that length of measurement?

[LambtonParticipant @lambton]

Graham,

Are you sure the axis of the DTI plunger is exactly aligned with the table axis. If it is not accurately aligned in this way the reading will always be too high.

[Mike PooleParticipant @mikepoole82104]

A dial gauge can measure accurately but it must be dead square to the change being measured. The lever type DTI can be used for comparison but will not measure accurately.

Mike

[FatgadgiParticipant @fatgadgi]

Graham

As Mike says, strictly a DTI has a lever arm and these are accurate for repeating a position, but not for measuring distance.

The Dial Gauge (or Dial Indicator) with a plunger should measure accurately over it's length. To get 7 thou out over 100 the angle needs to be quite big – something like 20 degrees, which I'm sure you'd have twigged.

If it's a Dial Gauge, try double checking with a digital caliper and if it's really 7% out ask the supplier for another one!!!

Cheers – Will

[Neil WyattModerator @neilwyatt]

It's an urban (workshop) myth! A DTI does NOT need to be dead square.

A bit of trig shows that to be 7% out, as in this example, the DTI would need to be angled at 21 degrees.

COS angle = 100/107 = 0.935

angle = 21 degrees

It needs to be 8 degrees out in order to have an error of 1 in 100, so setting a DTI by eye (which should be within a degree or so) will give a error of well under one part in 10,000 which is more than adequate even for the most demanding applications. The same applies to using a rule stand to set a height gauge, a degree either side of upright is more than adequate.

Neil

[Mike PooleParticipant @mikepoole82104]

It's easier to aim for dead square than say less than 8deg or whatever tolerance you can stand, when the variables are eliminated the rack operated dial gauge is very accurate as in a dial calliper.

Mike

[Neil WyattModerator @neilwyatt]

The secret to efficient working is 'good is good enough' don't waste time getting your DTI set up 'perfectly' when 'roughly' will be accurate to a tenth of a thou…

Neil

[Graham Williams 5Participant @grahamwilliams5]

Thanks for all the advice/comments guys. Dial gauge(plunger type) was eyeballed to be square on to table, got an old British plunger dial gauge somewhere that doesn't have the travel that the one I used has and will check using that, though for whatever reason in the past, was replaced by the, presumably, foreign manufactured one. Will try it with a digital caliper as well.

Graham W

[nigel jones 5Participant @nigeljones5]

Graham gets a detailed explanation and I get a link to a sewing website……….

[Graham Williams 5Participant @grahamwilliams5]

Battery flat in Caliper again, but John Bull DTI pressed back into service. Cleaned both shafts with a solvent and set them up by eye again. Readings somewhat different this time with the 'foreign?' DTI out by maximum of 14 tenths at both 200 and 250 thou table advancement, at 60 thou it was dead on. At 10 thou increments variation otherwise compared to DRO was between 4 tenths and 7 tenths. The John Bull was also dead on at 60 thou but at 10 and 20 thou was 8 tenths out but got better the greater table movement until only 2 – 3 tenths out at 200 and 250 thou table movement. 7 thou worried me but I'm happy the DRO is working OK now and much more accurate than me. Bit suspicious of the one DTI but if the last readings are the norm I think they are good enough for anything I can do.

Have a word with management Fizzy LoL.

Graham W

[John HinkleyParticipant @johnhinkley26699]

Here comes a slightly different spanner to go in your works ……

Maybe you've got an iffy thread on your Y-axis feed screw and all your gauges are 100% accurate!

If my neighbour slams his door on the way out, I can easily lose or gain a couple of (metric) thou here or there. I'm happy to work to that sort of tolerance.

John

[LambtonParticipant @lambton]

Graham,

How are you mounting the dial gauge?

It is very easy to flex or displace slender mounting rods , clamping joints, magnetic base etc. without realising it.

Try mounting your "inaccurate" gauge as rigidly as possible over a surface plate, or even the mill table, then test it against metal blocks previously measured for thickness using a good micrometer. If you have some slip gauges then so much the better. Carefully done this test will quickly show if the dial gauge is good or bad.

Eric

[SillyOldDufferModerator @sillyoldduffer]

For some reason lost in the mists of time I've always believed that the Dial Test Indicator was intended for use as a comparator rather than for taking accurate measurements.

I found a Mitutoyo Guide on the web that includes this table, presumably for a quality instrument:

I interpret the table to mean that the wide range accuracy of this particular range of DTIs can be out by as much as 0.01mm. Again, happy to learn I'm wrong, but I think this confirms that DTIs aren't the best way of measuring over a distance. In addition, I assume that my cheapo DTI is not as accurate as a Mitutoyo, and is even less satisfactory over a distance.  I would expect a much used quality brand to develop equivalent errors over time due to wear and tear on the mechanism.

Consequently I mostly use my DTI for jobs like centring work in a 4 jaw and wouldn't be puzzled by the apparent contradictions in Graham's original post: am I wrong?

Dave

Edited By SillyOldDuffer on 15/12/2016 10:53:21

[HopperParticipant @hopper]

Posted by SillyOldDuffer on 15/12/2016 10:51:22:

For some reason lost in the mists of time I've always believed that the Dial Test Indicator was intended for use as a comparator rather than for taking accurate measurements.

That's what I was always taught. In a type toolroom situation, if we wanted to measure something with with a dial indicator, you set up a stack of Jo' blocks (slip gauges) to the desired thickness, then set up the dial indicator on a solid stand on the surface table and set the dial indicator to read zero over the Jo' blocks. You can then measure any number of parts or items to compare them, within a small margin, to the recorded thickness of the Jo' blocks as indicated by the zero.

Same on the old jig borers before DRO came out. They had a V groove running beside the table in the x and y axes into which you laid an internal "stick" micrometer. One end registered with a stop on the table, the other with a dial indicator mounted to the bed so you always set the needle to zero to ensure correct positioning, and so you did not bow the stick mike. You never, ever, used the dial indicator to measure the movement of the table, even if you were only moving the table a small amount. That was always, always, done with the stick mikes.

When you look inside a dial indicator, there's an awful lot of gears and pivots and pinions stuffed in there to accumulate error over the full range of its movement.

That said, I have been measuring motorbike cam lift etc for years with dial indicators. But accuracy there is probably not as critical as some super tuners believe. The things are dancing all over the place at 10,000rpm!

[Michael GilliganParticipant @michaelgilligan61133]

Exactly, Hopper

The clue is in the words 'Test Indicator'

MichaelG.

[Graham Williams 5Participant @grahamwilliams5]

John. Scales on the machine seem to agree to what the DRO is saying within my ability to line up the divisions so don't think the screws are too far out.

Lambton. Had both DTI's on normal magnetic base set up attached to the mill column, mill is on a concrete base. Got a slip gauge set somewhere so will do as you suggest on a surface plate and with rigid setup and see what it gives me, though mostly used the DTI to set things true in the lathe, just curiousity to check table movement this time this way..

Thanks for the suggestion.

Graham W

[Martin ConnellyParticipant @martinconnelly55370]

The instructions for setting the pulse per inch for imperial CNC with Mach3 suggests using a 4" gauge block and a suitably mounted surface to press it against. You set your DTI to zero with the block in place, remove the block and move the table 4". If the DTI does not return to zero when against the fixed surface your pulse per inch count is wrong and needs adjusting. This is using the DTI as a comparator as earlier posts have suggested it is best used. For metric you would move 101.6mm to get the same effect or use a metric equivalent to the 4" gauge block.

Martin

 

Edited By Martin Connelly on 15/12/2016 14:23:44

[Neil WyattModerator @neilwyatt]

Posted by fizzy on 15/12/2016 08:11:55:

Graham gets a detailed explanation and I get a link to a sewing website……….

😛

That's because I thought you needed neatly sewn regulator covers (and I don't know the answer to your question).

N.

[Author]

Posts