Unwanted Taper ?

[fcheslop]

I think a tighter tolerance should be achievable to what level on those machines I have no idea
The approach needs to be methodical
My preference would be to use two test bars one that fits in the machines spindle and the other between centres and then do some measuring making notes as you go . So that you are starting from a known point
Level the machine in with the first with the machine bolted down or you a peeing into the wind
Set the tailstock with the second and it should also show if the bed is bellied
To check the head use a parallel and clock
Bogstandard did a very good write up on this and other forums about remachining one of youre lathes and detailed
My two lathes one a ML7R the other a Schaublin 70 TO will do what I want but as the temp and humidity changes they go out of whack. Both machines have cost as much as a medium SUV but have in the past made my daily crust
Chasing tenths can and is frustrating and often not needed depending what you want to make
Sadly precision costs both for the machines and the machines environment
Only you can decide to what level you want the machine to perform
If in the past I was machining below 0.001 then it would be a grinder jobby now if Im at 0.0003 Im a happy bunny
Im a little dyslexic so deliberately dont post often as it takes me far to long and life is to short
Other more wise will have better advice
cheers

 

[BaronJ]

More chance of the tool just rubbing and not cutting .

I don't know, I can easily take a tenth thou cut in most materials !

 

[BaronJ]

I think you are misunderstanding me - when I say a thou I am referring to 1/1000 of an inch or 0.001 inches this is one division on the cross slide so it is do able, I see you are thinking 0.0001 which is a tenth of a thou and you are right I know I couldn't dial in 0.0001.

A lot of getting fine cuts is down to technique ! The sine method will easily allow a tenth cut. The real question is can the machine achieve that level !

 

[BaronJ]

Those who are asking me (OP) why I am looking for less than .0005" taper in 2" or even 1", I was under the impression, maybe mistaken, that any lathe, without tailstock assistance, should be able to achieve this kind of tolerance. Please educate me if my expectations are too high.

You are talking about a Chinese lathe, by your own admission the lathe has a slight upward bow in the bed. Whilst old iron can achieve what you are looking for quite easily, they were made and built to do actual work to precise standards. Not like the toys coming out of the far East.

 

[goldstar31]

JohnB And Frazer- and I assume myself are perfectly capable of both removing 'barely perceptible swarf' and ' making such fine steel wool that it can be crumbled in the fingers. I have no doubt that we could all do it on a 7 x14 lathe. I can do it on a 8x16 which I have in addition to what I could do on the old Super7B and I have no qualms or doubts about it once I refit the 'funny ' motor on my nearly new Super7B PXF.

The common denominator in such things is that we all follow a classical professional approach-- and not try to to re-invent the wheel.
As John mentions, he and i are quite familiar the sine and a lot more of the many propositions of Euclid- and can instinctively apply our learning to something as trivial as has raised so much 'hot air'. I have no doubt that our good Frazer will have no problems either.

It does require getting things which are appropriate to the task in hand and not trying to use the wrong micrometer that will only measure at best half a thous when n appropriate one will measure in 'tenths'.

In other words, the right tool for the right job- to have the wit to blame ourselves when things go wrong- and not the tool.

My thoughts but the world may be full of people who disagree.

Norman

 

[IanN]

Those who are asking me (OP) why I am looking for less than .0005" taper in 2" or even 1", I was under the impression, maybe mistaken, that any lathe, without tailstock assistance, should be able to achieve this kind of tolerance. Please educate me if my expectations are too high.

Hi Danuzzo,

Nothing wrong with expectations!

There are two types of machine tool (in this case lathe) owner:

1) The person who uses their lathe to actually make stuff.

If you are this sort of user, there is no point in wasting time, money and effort making your machine any more accurate than you actually need - if your most demanding task only require a tolerance of +/- 5 thou then there is no point in chasing tenths

This is the sort of person who goes down the pub, shows off their latest beautifully finished steam engine and expects everyone to be impressed and buy them a pint

2) The type of owner who never actually makes anything and has “ownership of the tool” as their hobby

There is absolutely nothing wrong with this - if you get pleasure and a sense of achievement through making incremental improvements in machine accuracy then go ahead and enjoy yourself

Such an owner may never have any manufactured artifact to show off, but they do gain “bragging rights” allowing them to walk in to the pub and demand “My lathe can repeatable turn to one tenth over a two inch length - now buy me a pint!”

Personally I make stuff, and everything I make is a “one off” - the actual dimensions are not that critical as long as parts have a “good fit” with each other. Nothing has to be interchangeable with parts from a second item.

My lathe is over a hundred years old and is quite sloppy in a number of places - it is quite possible to produce very accurate and consistent parts on it, you just have to know what you are doing

Whatever sort of user you are, and whatever you do - just have fun and enjoy yourself

All the best,
Ian

 

[terryd]

I think you are misunderstanding me - when I say a thou I am referring to 1/1000 of an inch or 0.001 inches this is one division on the cross slide so it is do able, I see you are thinking 0.0001 which is a tenth of a thou and you are right I know I couldn't dial in 0.0001.

Hi,

If you set the compound slide to a fraction under 6° (actually 5° 44') then use the compound slide to feed in, each division is divided by 10. So if I feed in 0.001" on the cross slide dial the tool advances by 0.0001". a bit ott perhaps given all of the variables in lathe condition, material being cut, tool sharpness etc but it is actually very useful if you have backlash in your cross slide. e.g. if you need an accurate 5 thou' cut you dial in 50 thou' on the cross slide when offset as above and the tool advances by exactly 5 thou'. That perhaps may be a useful tip to stow away for future reference? So in the discussion above if you dial in 10 thou' (0.010") on the cross slide the tool advances by 1 thou' (0.001") - that can make for better accuracy - and all the better if you can lock the saddle and cross slides on your lathe.

And I understand that in the USA what we call a thou' (0.001") you call a mil. Is that correct? In the UK a 'mil' is short for a millimetre.

Regards

TerryD

 

[retailer]

terryd - thanks for the info on the compound slide setover I was aware that there was an angle somewhere between 2 and 10 that gave 1/10 thou feed on the cutting tool, but never got around to trying it out , I do believe it is a very useful thing to know, I might put up a small note on my splashback reminding me, I've had my lathe from new since the early 1980's and made a few steam engines etc but it is only since I machined the Quorn grinder that I became serious about working to fixed dimensions with in tolerances,I used to just machine parts to fit each other but it didn't occur to me that once the first part is machined the second part must be machined to a fixed dimension with a set tolerance so why not do it all of the time.
I'm in Oz and work in metric for just about everything except the model engineering where it is imperial and thou's, I can't quite seem to get my head around the metric engineering stuff 1/10mm is too coarse and 1/100mm is less than 1/2 thou and quite close tolerance, 1 thou seems just right. I've got an air spindle build in the pipeline perhaps I'll bite the bullet and try working in metric the closer tolerances can only help with the air spindle.

 

[ALEX1952]

I second IanN last post, the machine only needs to be good enough for the job in hand, yes you could strive for perfection but to maintain that perfection would be a nightmare, as mentioned previously a change in temperature or atmosphere would be enough to cause an issue. when I was working anything of the accuracy you are seeking was in a shop with controlled temp and atmosphere, some of the machines were never turned of so that they were allways at working temp, shafts would settle because the oil film would drain away, CNC machines had a cycle that would keep all the axis moving. This was to achieve what you are seeking, stop obsesing and enjoy your hobby get to know your machines failings, and work round them, as IanN says "mostly you are looking for a fit nothing needs to be interchangeble", which is why if something you buy requiring such accuracy you invariable buy it as an assembly or, they are made undersize for you to match and fit.

 

[minh-thanh]

Hi All !
Danuzzo !

This is my thinking at this point. The common denominator on all 3 lathes is Me. I am now reluctant to think that all 3 headstocks were out of alignment, and all in the same direction. The 1st one had a significant taper of about .001 per inch. The second one was a little better, especially after I bolted it to a bench with a shim on the tailstock end away from me.However, the next day, I went back to cutting a taper with the same lathe that did so well the day before.

Now on my 3rd one. At first, it was cutting a taper; but, I think it has improved because I shimmed the bench's foot at the tailstock end way from me. The lathe itself sits on the bench on the rubber feet. The lighter the cut (e.g. .001 for a total removal of .002), the greater the taper. The current piece I am testing is aluminum 1" stock sticking out about 3" from the chuck. The taper has been reduced to about less than .0015" over the 3"; except the last piece where I said "one last try for the night", and I took a cut of about .010 (actually .020 taken off) and I got much less taper. In fact, the first 2" from the tailstock end measured about .0005 off only at the first 1/4" or so. After that, it was probably less than .0001 off. But; then I get to the last inch and it tapers about .00075.

This was all yesterday, I haven't gone out to see what happens today. Help would be greatly appreciated.

All data you give are my dream !
Exact perfection is always desired, but if you are too focused on it you have to spend a lot of time and money ...
Focus on the projects you are going to do and adapt to your machines !

 

[goldstar31]

The lighter the cut on my taper issue, the greater the taper issue. Don't know what it means. Just my personal observation. Even after the shim, it went back to a taper on the aluminum. By the way, I took a Bridge City straight edge to the bed length wise, and I found a small gap between the ends. This was with the shim. I didn't measure without the shim. Don't know the significance of this; but, I did not think it was a good thing.

I missed this remark originally. Enough to question whether the lathe is bolted down or not when the straight edge is used. Of course the straight edge could warp as well

 

[SpringHollow]

My South Bend 10L is within around 0.001 over 36" after appropriate leveling.

 

[Cogsy]

I'm with the OP on this one (and it sounds like we're in the minority) but I would be quite upset if my lathe was delivering a taper of 1/2 a thou per inch as he describes. Relating that to making I.C. valve, that means my valve stem would be 0.025mm thicker at one end than the other, when I like to maintain a tolerance of half that (and that's on a very flexible part). Relating it to a crankshaft with 3" of shaft each side of the webs, there would be 0.05mm difference between the end of the crank and the web with the chuck end thinner than the tailstock . So to get the shaft end small enough to get a bearing to press on it would be sloppy and loose in it's running position.

Now expecting to cut on-size to a tolerance of 0.0005 thou might be a bit extreme (not that I think he was suggesting that), but in my opinion, expecting to have less than that amount of taper per inch of travel is certainly justified. I'd be interested to hear why I'm wrong though - change my mind!

Edit: I re-read the original post and I see that the taper is the opposite to what I remembered it, with the thicker end of the shaft being the chuck end, so my crank example means I would get the shaft small enough to get the bearing on then it would freeze part-way down the shaft. Still not ideal in my opinion.

 

[TonyM]

I have levelled and bolted down more high precision machines than I care to remember. Every machine was levelled before bolting. Bolting down, if required, was always done gradually checking levels as you go. I realise that many here have not had the benefit of experience and no real instructions are given when you buy a small lathe.
These chinese lathes are as flimsy as you like and invariably placed on a wooden bench. They only have two mounting holes and no adjustment other than shimming. This does not allow for any 'out of flatness' of the mounting surface. Bolting down tight to an out of flat surface will distort the bed in whatever direction regardless of how level it is lengthwise.

The head should be leveled in both directions then the bed should be leveled in both directions rechecking the head as you go.

If you want to ensure best chance of accuracy make some mounting pads with jacking screws on each corner and one holding bolt between these. Bolt one plate to the table and the plate with the jacking screws to the lathe using the jacking screws to adjust instead of shimming

 

[BaronJ]

Hi Cogsy,

I'm with the OP on this one (and it sounds like we're in the minority) but I would be quite upset if my lathe was delivering a taper of 1/2 a thou per inch as he describes.

Me too ! I have just less than half a thou over 6 inches. Around 0.0003" I'm quite happy with that. Particularly with bearing fits. I'm currently making a grinding spindle, so bearing fits are particularly important.

Now expecting to cut on-size to a tolerance of 0.0005 thou might be a bit extreme (not that I think he was suggesting that), but in my opinion, expecting to have less than that amount of taper per inch of travel is certainly justified. I'd be interested to hear why I'm wrong though - change my mind!

Edit: I re-read the original post and I see that the taper is the opposite to what I remembered it, with the thicker end of the shaft being the chuck end, so my crank example means I would get the shaft small enough to get the bearing on then it would freeze part-way down the shaft. Still not ideal in my opinion.

No I agree with you ! There are so many variables involved here you can spend weeks or months chasing your tail !

 

[IanN]

..... a taper of 1/2 a thou per inch as he describes. Relating that to making I.C. valve, that means my valve stem would be 0.025mm thicker at one end than the other,
:
Relating it to a crankshaft with 3" of shaft each side of the webs, there would be 0.05mm difference between the end of the crank and the web with the chuck end thinner than the tailstock.

Hi,

Let’s get real here. addressing the above points in order:

1) Your example of machining a valve stem with 1/2 thou/inch taper and a total end-to-end discrepancy of 0.025mm (1 thou) suggests a machined stem length of 2 inches - a full sized I.C. engine valve. If you are intending to make critical engine components for your car, holding a taper over 2 inches is the least of your problems (in real a production environment such items would be rough turned on a lathe then ground to size - much quicker, easier and more precise)

2) Your second point above considers two bearing journals 3 inches apart. As I’m sure you already know, no-one would ever seriously attempt to make a shaft to the journal diameter along its entire length - you would machine one journal to size, move along and machine the second journal to size then machine the shaft in between to about 10-20 thou under the journal size (and leave it rough turned - don’t waste time, effort and money putting fine finishes where they are not required)

The issue with machine accuracy breaks down in to two sets of criteria: Accuracy over short distances, and accuracy over long distances. For most of us, accuracy over short distances (less than 1/2 inch) covers 99% (often 100%) of a hobby user’s requirements. Unless you make full sized pistons for 3 litre car engines, I cannot imagine anyone actually needing to hold 1/2 thou over 3 inches

All the best
Ian

 

[goldstar31]

I sort of go along with the suggestions of TonyM. The difficulty is that the OP is not following a recognised test procedure. The only hint of proper measurement is the use of a straight edge. Whether the lathe when checked was 'bolted down' or not has yet to be answered. My own SiegC4- the slightly larger version at 8 x 16 is as good or perhaps better than most would admit. It's Chinese, it's crap, and so on. Mine has a powered milling attachment- and true to form- the motor was crap but replaced under warrantee. At the price- second hand for the lathe itself, it is surprisingly good value.

However the complaint of the present poster is not unknown. It was answered here and as far as I can recall, was only resolved with obtaining a new lathe bed. Again, I vaguely recall someone with a 'blued' straight edge discovering a shear which had literally sagged- after it was machined.
Whether our OP's machine is something similar is unknown.
More years than I wish to recall, my father would 'normalise' warped steel. He had little to do with castings but that was raised by a then apprentice who found mountains of part machined castings left out in the open to rust and weather before being brought in for final machining. Both he and the factory are- long gone.

All of this is nothing new. I can - or could shrink or otherwise normal sheet steels but I would be loathe to attempt this on high strength low alloyed steels.

So it remains with the OP. to actually correctly determine the situation before anyone attempts to find a cure for it

 

[goldstar31]

John
Lathe beds etc aside, I have written using your 'other' e-mail address.

Still trying to snow so I'm going to do my tax return-- in the warm

From one Grumpy Old Git to another

Best Wishes

N

 

[terryd]

terryd - thanks for the info on the compound slide setover I was aware that there was an angle somewhere between 2 and 10 that gave 1/10 thou feed on the cutting tool, but never got around to trying it out , I do believe it is a very useful thing to know, I might put up a small note on my splashback reminding me, I've had my lathe from new since the early 1980's and made a few steam engines etc but it is only since I machined the Quorn grinder that I became serious about working to fixed dimensions with in tolerances,I used to just machine parts to fit each other but it didn't occur to me that once the first part is machined the second part must be machined to a fixed dimension with a set tolerance so why not do it all of the time.
I'm in Oz and work in metric for just about everything except the model engineering where it is imperial and thou's, I can't quite seem to get my head around the metric engineering stuff 1/10mm is too coarse and 1/100mm is less than 1/2 thou and quite close tolerance, 1 thou seems just right. I've got an air spindle build in the pipeline perhaps I'll bite the bullet and try working in metric the closer tolerances can only help with the air spindle.

Hi,
I also work in metric and have mostly metric tooling with a little imperial stuff that I sometimes pick up from car boot sales but on the whole I tend to convert drawings to closely matching metric sizes. I realise that 0.1 mm is 0.004 thou' and therefore 0.01mm is 0.0004 so I can use 10ths mm to get reasonably accurate and let's face it, if you're making steam models as we say here 'half a gnat's ****' is close enough. Of couse if miniature pulse jets are your thing that's a different matter. After all the very early Newcomen atmospheric engines had large cast iron cylinders hand scraped and filed and the piston was sealed with a leather flap around the circumference and a layer of water on top to complete the seal. As you probably know the piston was only powered on the downstroke as the partial vacuum created by the condensation of the low pressure steam below the piston pulled it down. In those circumstances tolerances were measured in large fractions of an inch. James Watt's achievement was to improve the efficiency of Newcomen type atmospheric engines.

TerryD

 

[IanN]

:
The difficulty is that the OP is not following a recognised test procedure.

This is, of course, the root of the problem. There is a proven path to follow when testing and trueing a lathe and any omissions or deviations from the sequence will result in uncertain test conclusions and, therefore, the likelihood of faulty corrective actions

There is a motto in engineering: “Check everything, assume nothing”

Until the bed is proved straight and true against a surface plate, there is little point in doing any further corrective work

All the best,
Ian