mini lathe dial accuracy

[s dParticipant @sd79178]

Hi, first post, hello all.

I've bought a mini lathe and have a question about accuracy of the dials on the cross slide and compound. These seem to be consistently out by 0.025-0.05mm, 1-2 markings when making a full rotation. I'm checking them against an indicator and it seems pretty consistent on both dials, requiring a bit more than indicated to get to a full 1mm travel. Is this normal?

[s dParticipant @sd79178]

[Clive Brown 1Participant @clivebrown1]

Have you checked the accuracy of your dial indicator over the travel that you're using?

[John HaineParticipant @johnhaine32865]

No, it is not normal, but since it applies to both axes it may be your indicator. These are not precision devices. better would be to use a feeler gauge ideally calibrated against a micrometer.

[s dParticipant @sd79178]

Posted by Clive Brown 1 on 08/06/2023 19:15:25:

Have you checked the accuracy of your dial indicator over the travel that you're using?

No, but I did cross check with a digital vernier guage and got the same result. That could still be inaccurate, but it is the same degree of error.

[s dParticipant @sd79178]

Posted by John Haine on 08/06/2023 19:15:59:

No, it is not normal, but since it applies to both axes it may be your indicator. These are not precision devices. better would be to use a feeler gauge ideally calibrated against a micrometer.

OK that's good to know in terms of expectation. I did also check at multiple regions of the indicators travel and it was consistent at least. I don't have any other guages unfortunately.

[John HaineParticipant @johnhaine32865]

Was the travel of the indicator exactly parallel to the axes? .025 error in 1mm travel would be caused by 1.4 degrees out of parallel I think. If the error is real and the lathe new you should seek a refund.

[Keith LongParticipant @keithlong89920]

How are you doing the measurement, have you eliminated the effects of any backlash?

[Tim HammondParticipant @timhammond72264]

I have a Clarke CL300 mini lathe, and although the dials are graduated in Imperial measurements, the various screws have metric pitches. Hence 0.040" on the dial (one complete revolution) is actually only 0.039" ( 1 mm = 0.03937"  ) Could this be the source of your problem?

HTH.

Edited By Tim Hammond on 08/06/2023 21:02:11

[Anonymous]

Posted by Tim Hammond on 08/06/2023 21:01:24:

….Could this be the source of your problem?

That would be my guess as well. Plus, the slide screws are not going to be precision ground, so a thou or two error is probably to be expected from a low cost product.

Andrew

[DC31kParticipant @dc31k]

If you are using an indicator, instead of measuring only one turn of the handwheel, measure as many turns as the indicator travel will allow.

That will do two things: average out any local errors and confirm if the errors accumulate.

You could also set the compound parallel to the cross slide and wind one in and the other out (after having taken up backlash) and check that net movement is zero. That might eliminate the effects of angular error referred to above.

Edited By DC31k on 08/06/2023 21:33:49

[s dParticipant @sd79178]

Posted by John Haine on 08/06/2023 20:42:35:

Was the travel of the indicator exactly parallel to the axes? .025 error in 1mm travel would be caused by 1.4 degrees out of parallel I think. If the error is real and the lathe new you should seek a refund.

Morning all…

I repeated the measurement with a much more precise setup. I think an out of parallel error is one of the main things I was seeing. The indicator has a very short collar, not much to judge by. That and I've learned that indicator dials are sensitive but not necessarily that accurate.

I clamped a digital vernier guage under the toolpost (a cheap one!) which allowed me to get it as absolutely parallel to the compound as I could visually (sighting a hair-width gap between the straight edge and the compound over about 15cm, assuming the latter is parallel to its dovetail) and rigid and measured off the tailstock.

I got essentially no error. 10.01mm over 10mm of travel.

This taught me a lot about these tools and methods. Thanks for all the responses helping to figure this out.

Edited By s d on 09/06/2023 06:39:10

[s dParticipant @sd79178]

Posted by DC31k on 08/06/2023 21:32:53:

If you are using an indicator, instead of measuring only one turn of the handwheel, measure as many turns as the indicator travel will allow.

That will do two things: average out any local errors and confirm if the errors accumulate.

You could also set the compound parallel to the cross slide and wind one in and the other out (after having taken up backlash) and check that net movement is zero. That might eliminate the effects of angular error referred to above.

Edited By DC31k on 08/06/2023 21:33:49

 

Great tip re setting the compound parallel to the cross slide, I'm curious to try it just because…

Edited By s d on 09/06/2023 06:33:25

[s dParticipant @sd79178]

Posted by Tim Hammond on 08/06/2023 21:01:24:

I have a Clarke CL300 mini lathe, and although the dials are graduated in Imperial measurements, the various screws have metric pitches. Hence 0.040" on the dial (one complete revolution) is actually only 0.039" ( 1 mm = 0.03937" ) Could this be the source of your problem?

HTH.

Edited By Tim Hammond on 08/06/2023 21:02:11

Yes good point. I did consider this, the markings are for both 0.001" and 0.025mm. But 0.001" is 0.0254mm, so it should reach a whole unit slightly before a full turn, and accumulate, if the threads were imperial. My more accurate test suggests they are metric so all is good.

[John HaineParticipant @johnhaine32865]

Result! Good work.

[Howard LewisParticipant @howardlewis46836]

Good!

Wondered if this might be the case.

Gives a bit more confidence in the machine.

Howard

[s dParticipant @sd79178]

Posted by John Haine on 09/06/2023 06:54:48:

Result! Good work.

Yes. Thanks!

[HopperParticipant @hopper]

Posted by s d on 09/06/2023 06:30:56:

… I got essentially no error. 10.01mm over 10mm of travel.

…

Don't expect to consistently get that level of precision when actually cutting metal, due to tool wear, machinery flex, material flex, tool clamping shift, backlash, lathe bed aligmment etc etc.

Accuracy of the dials on the cross and top slides is a bit of a moot point in daily use. Even if you did have the original (disproved) 0.025 per 1mm error, you would not normally notice it in daily use in a home workshop. A 1mm deep cut is a roughing cut so 0.025mm error would not matter. Finishing cuts are taken at a depth circa 0.1mm to 0.02mm, so the error over that short distance would be immesureably small.

You will find once you get to know your lathe, you will not look at the dials much for roughing out. You know 1 turn of the handle is 1mm so that is near enough, or half a turn for half a mm etc. Then the dials come into play for the fine finishing cuts. Even then you usually need to measure between every cut and "sneak up" on the desired dimension.

[s dParticipant @sd79178]

Posted by Hopper on 09/06/2023 11:01:58:

Posted by s d on 09/06/2023 06:30:56:

… I got essentially no error. 10.01mm over 10mm of travel.

…

Don't expect to consistently get that level of precision when actually cutting metal, due to tool wear, machinery flex, material flex, tool clamping shift, backlash, lathe bed aligmment etc etc.

Accuracy of the dials on the cross and top slides is a bit of a moot point in daily use. Even if you did have the original (disproved) 0.025 per 1mm error, you would not normally notice it in daily use in a home workshop. A 1mm deep cut is a roughing cut so 0.025mm error would not matter. Finishing cuts are taken at a depth circa 0.1mm to 0.02mm, so the error over that short distance would be immesureably small.

You will find once you get to know your lathe, you will not look at the dials much for roughing out. You know 1 turn of the handle is 1mm so that is near enough, or half a turn for half a mm etc. Then the dials come into play for the fine finishing cuts. Even then you usually need to measure between every cut and "sneak up" on the desired dimension.

These are good points, thanks. I figured I would need to be measuring and sneaking up for fine work so it makes sense that in practice it wouldn't matter much. I think its still useful for understanding where exactly I am in the measurement landscape as it were, what tools I can trust, what not etc and in which circumstances. And just to know how (im)precise my lathe is. As a beginner there's a lot to take in and reducing variables helps.

[HopperParticipant @hopper]

Oh yes . It is definitely good to know exactly where you stand at the outset, as lots of small errors can add up to one bigger one, usually right when you least expect it!

[not done it yetParticipant @notdoneityet]

My initial thoughts were the same as Tim and Andrew.

Next was when the OP said ‘sesitive but not necessarily that accurate made me wonder if a dti was being mistaken for a micrometer.

Other thoughts were whether the carriage and/or cross slides were locked down if/ when checking the compound travel. A decent thread gauge may well be good enough to prove/demonstrate if the threads are imperial or metric?

Don’t ‘sneak up” is my advice. The cut will be consistent only as long as the increments are the same. Get used to taking the last couple of cuts (at least) with the same depth of cut and feed rate. Better to take a deeper cut than trying to cut the last 0.02mm on the lathe – most especially if using un-polished carbide cutters/inserts.

[old martParticipant @oldmart]

I have just popped down to the garage to measure the leadscrews on my old Warco 7 x 12 mini lathe. The original owner traveled from Frome to Warco's at 12 miles SSE of Guildford to pick it up. He had ordered an imperial model, but they had mistakenly provided a metric one. They gave him an imperial leadscrew and nut and one or two other things to pacify him. Measuring ten turns of the cross slide leadscrew, I got 9.97mm twice with backlash accounted for. The compound gave 9.70mm for ten turns. For normal turning these figures are plenty good enough because a micrometer would be the proper way to measure diameters.

[Howard LewisParticipant @howardlewis46836]

Certainly, when boring whether with carbide or HSS, take at least one, or more, spring cuts without altering any settings, measuring adter each pass, even after sneaking up on the size..

As Jason says, everything flexes (How much depends on the factors he mentioned) so is JUST away from finished size, take a spring cut .

If in doubt, before taking the finishing cut, sharpen the HSS tool, or replace the carbide insert, and reset to centre height.

Howard

[Chris MateParticipant @chrismate31303]

I think in the end it boils down to how accurate you can measure the movement on your lathe or mill as a direction to accurac(Equipment like DRO/Other means), and how accurate you can measure the result you obtained just before accurate point or at final point achieved(Product with other means).

I think it becomes tricky when you have to produce anything to a certain spec of accuracy without the ability to fit parts together, so temperature as well. I struggle going from a micrometer on outside to using the same micrometer with telescopic boring gauges on the inside diameter relying just on measurement to achieve a wanted pressfit..

Edited By Chris Mate on 10/06/2023 07:22:21

[CirclipParticipant @circlip]

If you want to machine to RR accuracy you're going to have to fit DROs .

Regards Ian.

[Author]

Posts