Lead screw accuracy

[Julia]

I have just fitted a SINO glass scale to the cross slide of my new lathe.

When I checked the new scale against the lead screw in the cross slide I was surprised to discover a discrepancy of 10 thou built up over 6 inches of travel. I wound the cross slide out and zeroed by the digital scale and the lather dial. I then carefully wound the lathe slide in 6" inches (exactly 60 turns) and the digital scales read 5.99035.

At first I thought there was a problem with the scales so I checked them for parallelness with a clock gauge. While I was fairly confident the SINO scales were correct (these are expensive professional glass scales not the cheap Chinese type) I cross checked my results using a digital calliper. These reading confirmed that the lathe lead screw is in error.

Question: is an error of 1.6 thou per in an acceptable/typical error to be expected on the lead screw of a lathe cross slide.

[Deleted]

Julia. Here is a lathe test sheet from my (much) younger days. It's not that legible but just about everything is required to be within one thou per foot, leadscrews included. The permissible spindle run out is half a thou max.
There is no spec for the cross-slide screw but I would have treated it as a leadscrew because I would be attempting to work to the same accuracy in every direction.
In summary, if you are confident of your measurements then I'd send the lathe back!
Sorry to be so negative, but you did ask.
JB
EDIT: It would be interesting to see what results one would get nowadays...........

[rodwilson]

Apologies if this is a dumb question.........
If you expect near total accuracy from the lead screw what is the benefit of the digital scale? Is it the lack backlash or what?

[havoc]

That's 0.2%. Both your SINO and your caliper aren't perfect either, look in the manual to see what their fault can be.

[Deleted]

Rod, backlash is a red herring, you can have as much as you like provided you make allowances for it. I think the DRO reduces the need to do sums, and it eliminates pencil marks on the dials!
JB
Havoc, my cheapie caliper reads spot-on against 6" slips: I expect Julia's will do the same.
JB

[jasonb]

I'll give mine a check later as we have the same machine.

Given that I would take a measurement of the part being machined when its say 0.020" over finished size with either mic or veneers and then turn the wheel to give a 0.010" depth of cut at 0.2% error that would leave my part oversize by 0.00002" which I can live with.

have not done the maths but what length would the screw expand by if your workshop is say 60deg F in the winter and 80deg F in teh summer?

Jason

[havoc]

Expansion coefficient of steel is 13 x 10^-6/K This is relative to the lenght. Not much I would say.

Everything has errors. The question is if you can live with it. I would have no problem with that error as I turn out worse parts that that.

JB: I took the "manual" of my calipers and it says (+/- 0.02mm < 100mm), (+/- 0.03mm 100-200mm). This is a recent one, the older ones were (+/-0.1 +/- 1 digit). I would not trust them more than that.

[Julia]

Thanks for the replies. I used two different callipers and my gut feeling is that that are accurate to +-1 reading digit. I have had measurement crises before and have ending up with everything agreeing. I expected to get the lathe to fall in line! When I fitted a DRO to my mill I don't recall any significant errors.

I am fitting the DRO to make life easier and so that I don't get confused where the tool is. I have a quick change tool holder and I hope to able to change tools without getting lost. I still expect to use a mic but only just before the finishing cut.

I wonder whether it has anything to do with it being a imperial lathe made in a metric factory. I am suspicious as the change wheels for metric screw cutting use a 63 tooth cog and not a 127 tooth cog which would provide an exact ratio. (2 * 127 -> 254 4 * 63 -> 252)

[jasonb]

Just been out and measured mine

with the digital callipers (Mitutoyo so should be better than the machine DRO ones) I got 0.005" over 3" or movement

With a dial indicator 0.0015 over 1"

So both are similar to your machine

Jason

[Deleted]

Just to add another 'pennorth. Whilst in practice an operator would rarely require to register a large critical distance, I still think that the figure of .0016/" obtained by Julia is excessive. I just measured my MaximatV10P, admittedly only over 35mm using the caliper, but the total error was less than a hundredth of a millimetre, barely measurable assuming no visual parallax or cosine error. But that result may be something to do with the fact that it's Austrian......

I take your point Havoc about overall caliper accuracy. It reads spot on with careful handling, but I didn't say that I would trust it over that distance! For critical dimensions I ALWAYS zero the caliper on slips and work +/- around the zero. In a lot of respects it's better than a DRO.......
JB

[steam4ian]

G'day all

Julia your lathe looks quite new and is probably Chinese.
Some Chinese machines do not have true inch scales but have metric feed screws typically with a pitch of 1mm.
The collar is calibrated to 40 implying 40 thou". However 25 turns does not move the slide 1 inch but 25mm which is 0.4mm short and 0.4mm = about 16 thou".

There may be "nothing" wrong with your lathe at all.

Regards,
Ian

[Julia]

Ian
I suspected (see above) that it may have something to do with a metric imperial approximation but the error is considerably less than 16 thou per inch. It is a be weird that the error is aprox 1/10 of this figure at 1.6 thou per inch.

If the imperialness of the lathe is not exact I would expect the literature to make this clear. Knowing what I know now I would have gone for the metric version as these come with dials marked in both metric and thou. While these dial don't have a nice number of thou per turn when coupled with digital scales I would have had the best of both worlds. The catch may be that the metric lead screw might not be exact either.

I have not yet fitted the scale to the longitudinal axis but when I have I intend to check the screw cutting lead screw. If this is not exact I shall be very unhappy. Errors elsewhere I can live with but not when it comes to screw cutting.

[ozcutter]

Julia,
A currently available commercial trapezoidal leadscrew (Brisbane) is specified at pitch <=0.3mm / 300 mm ( 1 thou / inch ) with lead <0.1mm / 300 mm (0.3 thou / inch ) accumulating error.
These leadscrews are sold as replacement items for industrial machines.
(I am about to try one on a Tom Senior M1 horizontal mill)
The above totals to about 1.3 thou / inch, not too far from your measured results.
John.

[ettingtonliam]

Hi
Jason mentioned temperature. I seem to remember that machine tool accuracy is only guaranteed at a specified temperature - seem to remember around 20C. If Julia's lathe temperature (not air temperature) is much different to this, it may be significant. A cold machine will give a shorter travel. Don't know what the difference in expansion coefficients between glass and steel is though, or what effect that would have.

Regards
Richard

[44767]

Hi All,

Have a look at www.roton.com . This is a leadscrew manufacturer and quote a tolerance of +/-0.009" per foot for Acme, trapezoidal etc. threads and +/-0.010"per foot for ball screws. Modern CNC machines compensate for these inaccuracies with factory set parameters. In our applications we should be roughing out and measuring then taking final cuts to suit with much smaller errors involved. Roton come with my recommendation. They produced a one off 4' length of 7/8-4TPI-LH for me within a couple of days and I had it in NZ within a week of ordering!

Cheers,Mike

[Tel]

Julia, you should treat any dial reading as an approximation - I would be more than happy to live with an error of that magnitude.

[Staffordshirechina]

Just a thought but as soon as you fit a DRO to any machine, you never need to look at the dials again. Then it doesn't matter how rough your feeds or backlash are, the DRO gives you the correct reading. I haven't looked at the graduations on my big lathe and mills for years. I only use the dial on my Myford cross-slide because I am too mean to buy a Newall scale for it. The saddle has an ordinary glass scale but they don't fit the crossslide.
I suspect that if you complained to the supplier of your lathe, they would just say either lump it or have a refund. They would not (or could not) offer a better accuracy.

Les