DRO on lathe cross slide accuracy/resolution

[JmccrackParticipant @jmccrack]

I just fitted a 6' igage on my cross slide. I have a grizzly 14" x 40". I am please with the way it works, But as with any DRO you still have to know how to use a mic Here in Canada it cost me $40.00 well worth it if only to use as a reference with the dial on the machine.

[JasonBModerator @jasonb]

Although the Sinos will read out to five places behind the point when set to imperial, they actually jump in 0.00019 or 0.00020 steps so that is where the 0.0002" resolution comes from.

In Metric they read in 0.005mm steaps

Its worth looking for a readout that combines the carrage and topslide readouts into one line of display as you may find it hard to set that last 2/10ths cut with the carrage handwheel, the topslide being easier to put on a small cut with

The

[John Stevenson 1Participant @johnstevenson1]

Posted by Bogstandard2 on 30/03/2015 23:04:34:

Stuart,

I have two Sino display heads (four measuring heads) on my largish lathe (Chester Crusader) and the display heads can be set to read to 5 digits behind decimal point.

I tried to use it at that setting and it was almost impossible, a fly farting as it went past would cause the display to jump all over the place.

I now only ever use four places on both my lathe and mill, and life is very sweet.

John

 

.

 

Many years ago when DRO's were mega expensive and very basic one of my customers had machines fitted with Heidenhain DRO's that read to 1/2 a thou. Last digit went 0, 0.5 0, 0.5 etc

 

When they upgraded to BS 1066 or whatever it was before BS9000 they had to fit all new DRO's to read in tenths

0.0001, 0.0002 etc and I bought a couple of the older two axis ones off them to fit on my round column mill drill, remember paying £400 for a 2 axis second hand setup in those days.

 

Next time I went into their works ALL the machines had masking tape covering the last digit because the operators were fed up of the dancing numbers. God knows what that operation cost for no outcome.

Edited By John Stevenson on 31/03/2015 09:28:39

[Ken GillottParticipant @kengillott]

I fitted 4 DROs to my Chester machine centre/lathe about 4 years ago. The milling head and tailstock DROs are read directly no problem, however I had to fit two remote head LCD readouts for the lathe saddle.

The main problem I had was reliability of the cables at the DROs, which are non standard rubber 4 contact affairs. In desperation I dismantled one unit and with some difficulty soldered in a miniature USB socket which has been ok so far.

I still found the DROs & head readout units themselves to be unreliable – which is hopeless if undertaking co-ordinate drilling. Replacing the DRO LR44 batteries with 100uf capacitors fixed part of the jitter problem.

Next one of the head LCD readouts switches fell apart – impossible to find a replacement, so I bit the bullet and decided to design and build my own and used nice bright 0.7 " LEDs (Microprocessor based systems is another hobby).

I had always thought the half thou last digit a bit daft so I wrote the software to display in .0001" increments……oh dear, the digits do leap about. I put up with them last year but have changed to half thou to try this year. Hopefully I will now return to producing scrap metal instead of programming bugs!

[Ady1Participant @ady1]

I only use direct cheapo DROs but I do have a tip for anyone who has live wire weak point issues because my e-cig charger gets heavily used and the wires will soon fail if they are not properly supported

I whipped all the weak points with thread, like a fishing rod whipping, then squirted cheap superglue onto those whippings to strengthen them

Instead of a charger lasting for a few weeks my latest one has been going for months

[Stuart BridgerParticipant @stuartbridger82290]

Just as a FYI, it is about a 30% premium on the scale price to go from 5um to 1um

[CrocadillopigParticipant @crocadillopig]

[Just as a FYI, it is about a 30% premium on the scale price to go from 5um to 1um]

Stuart, that ain't necessarily so. Check out **LINK** I bought a system from this firm at 1um

for my lathe, and one for my miller from M-dro at 5um at virtually the same price!

[Brian Rice 1Participant @brianrice1]

Would not do with out mine on mini lathe my accuracy as Improved 100 percent no more guessing.

[Neil LickfoldParticipant @neillickfold44316]

It is not that much more money to have the more accurate scale on the X + Z axis, you can choose if it reads in 2 or 3 decimal places as well, apart from radius or diameter.

Digital readouts have lots of conveniences. Some have the ability to set tool numbers and zero them , so interchanging tools is very easy without having to re zero or reset the tools.

Neil

[Gordon WParticipant @gordonw]

It is worth remembering that setting , say, 1 thou' on the slide does not mean that the machine will cut 1 thou'.

[Neil WyattModerator @neilwyatt]

The capacitative scales (regardless of resolution) work to a code that uses 20,480 positions per inch.

So you can exactly denote any power-of two fraction from 1/1 down to 1/2048.

1 thou = 20.48 units.

0.1 thou = 2.048units

0.01mm = 8.063 units.

0.005mm = 4.031 units

Note these aren't errors and that the rounding required (if done properly) only introduces errors far smaller than the resolution of the units.

The code allows for more sensitive scales than currently made, and you can make a stand alone readout that works to 0.0001" but the jitter will make it unusable like in John's example. My homebrew displays round to the nearest 0.0005".

If the errors in the output follow a genuinely gaussian random distribution (which can't be guaranteed) you could use oversampling to increase the resolution to spuriously high levels.

I don't know if the lower resolution readouts still output every step of the 20480 per inch.

Neil

I feel a metrinch moment coming on…

[Les Jones 1Participant @lesjones1]

Hi Neil,
While I suspect all the scales read the data from the scale at 20480 CPI resolution only the original Chinese scales and caliper output it in that format to the data connector. The BIN 6 scales output at 2540 CPI when the display is set to metric and 2000 CPI when the display is set to imperial. The 7 BCD scales output at 2540 when set to metric. I have not made a note of the CPI value when set to imperial. The igaging scales outut data at 2560 CPI. I think there will be some rounding errors generated within the scales for BIN 6 and 7 BCD type scales

Les.

[JasonBModerator @jasonb]

Posted by Les Jones 1 on 01/04/2015 14:13:27:

Hi Neil,
………………….. The igaging scales outut data at 2560 CPI.

Les.

Neil, does this mean the Igaging scales are Metrinch compliant?

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Neil Wyatt on 01/04/2015 12:39:17:

I feel a metrinch moment coming on…

.

Neil,

Given the SI definitions of the metre and the second; it's quite likely that the EU will [next April?] ban the use of both decimal and binary, as both systems are incompatible with Cæsium transitions.

The Prime Factors of 299,792,458 are:

2 • 7 • 73 • 293339

 

The Prime Factors of 9,192,631,770 are:

2 • 3 • 3 • 5 • 7 • 7 • 47 • 44351

 

MichaelG.

[Nick_GParticipant @nick_g]

.

As you are contemplating tight tolerance scales I am 'presuming' that you are not on a tight budget.

If that is in fact the case I would consider spending those extra coins purchasing magnetic scales.

1) They are a doddle to fit.

2) They are already coolant and swarf proof.

3) They are of a smaller profile. This matters quite a bit on a small lathe so that the amount of capacity reduction is minimised. If a normal (even slim) is fitted by the time a protective cover has been fitted a couple of inches movement / capacity may be lost. This may not be an issue on a large lathe but can easily become a problem on smaller machines that are associated with most model engineers.

Nick

[Stuart BridgerParticipant @stuartbridger82290]

I did look at magnetic, but based on machine DRO's specs, they quote an accuracy of +/ 0.015 mm "typical after installation" even on the 1um resolution, which doesn't compare very favourably with the +/- 0.005mm quoted for optical. The slim optical scale with cover would mean losing about an inch with the tailstock being further way from the chuck.

[Anonymous]

Posted by Neil Wyatt on 01/04/2015 12:39:17:

If the errors in the output follow a genuinely gaussian random distribution (which can't be guaranteed) you could use oversampling to increase the resolution to spuriously high levels.

It's not so much the over-sampling as the averaging that may help, assuming that the input signal is time invariant, or at least quasi-stationary. The process does not improve the S/N ratio of the input signal, it reduces the effect of noise that is inherent in the conversion process. If the noise is small enough such that the output word doesn't change for a steady input then the process is of no help.

For time varying signals differential non-linearity in the ADC transfer function is often more important in determining the accuracy of the output. Deliberately adding random noise to the signal can help; known as dithering.

This link **LINK** may help explain some ADC noise effects and techniques.

Regards,

Andrew

[Ken GillottParticipant @kengillott]

Neil

I can confirm that the cheap Chinese digital scales I use do indeed output down to give .0001 increments. I mask off the least significant two bits of the 24 bit data field before doing the divisions.

During testing using a 6" digital vernier gauge I have a stable display at this resolution, however that is in the warm dry study. In the garage using the scales on the lathe is when the problems with stability etc arise. This may well be down to damp conditions as suggested by others and indeed I have less trouble in the summer months, but I think static/induction and lead lengths also play a part.

Sometimes the scales just "lock up" and the way around this is to fit a 3 way switch to enable isolation and discharge of the capacitor in the scales. I feed the scales with 1.65v from my homebrew readouts.

Ken

[Steve WithnellParticipant @stevewithnell34426]

So when buying scales then does it actually matter if they are based on 24bit (typical chinese) or 21 bit (tWarco) electronics when the context is cold damp workshop and no machines mill / lathe etc more than £1k list? ie not an Aerospace toolroom environment.

Steve

[Ken GillottParticipant @kengillott]

Er, quite. I had a bit of lateral thinking and proposed moving the machine centre to the study to cure the damp problem. However there was great opposition to this solution from the boss!

Ken

[Stuart BridgerParticipant @stuartbridger82290]

Well, after all this discussion, I have managed to get the first serious amount of "shop time" for ages. This has made me question whether I really need a DRO on my lathe???? As always, practice makes perfect and getting good hands on time has built my confidence and I find that I am getting very good results in a sensible time the traditional way with dials.

Stuart

[Michael GilliganParticipant @michaelgilligan61133]

Posted by Stuart Bridger on 03/04/2015 13:14:01:

… I find that I am getting very good results in a sensible time the traditional way with dials.

.

Good Result, Stuart.

Nowt much wrong with Analogue.

MichaelG.

[Neil WyattModerator @neilwyatt]

> For time varying signals differential non-linearity in the ADC transfer function is often more important in determining the accuracy of the output. Deliberately adding random noise to the signal can help; known as dithering.

I was over-simplyfting, having been brought to book on this by professional electronicists once before.

Neil

[Anonymous]

Posted by Stuart Bridger on 03/04/2015 13:14:01:

This has made me question whether I really need a DRO on my lathe???? As always, practice makes perfect and getting good hands on time has built my confidence and I find that I am getting very good results in a sensible time the traditional way with dials.

Apaart from the woffle on ADCs and noise I have kept quiet so far on DROs. However, I concur with Stuart; I simply don't need a DRO on my lathe. I have a 2-axis DRO on the vertical mill, and it is probably the most useful accessory I have bought. But I just don't see the need for one on the lathe. The DRO on the mill does not interfere with any of the table travels, or get in the way. However, on the lathe it would do so. The cross-slide unit would definitely interfere with tailstock positioning. In addition I have a hydraulic copy unit fitted, and the bars for holding the patterns for same make it difficult to fit a DRO. If I was going to fit another DRO to anything it would probably be the horizontal mill to turn it into a poor mans horizontal borer.

Andrew

[Bob Brown 1Participant @bobbrown1]

I have thought about a DRO on the lathe but I can not justify the cost, I have a micrometer stop on the saddle so I can traverse the saddle accurately and measuring and using the dials I can easily get to 0.005mm or 0.0002".

DRO is very useful on the milling machine.

You can buy a lot of other kit for the cost of a DRO ststem, or beer if it comes to that.

Bob

[Author]

Posts