ER collets in South Bend lathe

[Mosey]

Bog,
You are correct on all counts, of course.
I don't have the metric gears as yet, and someday, maybe. I know the method of cutting them, just haven't done it.
If the chuck maker thinks that .003" runout is acceptable, I should cancel the order and return what he sent, as this is not good enough.
I have no grinding capacity either, but as a small-time hobbiest, I am happy to buy these things instead, provided they are of sufficient quality. I wonder if a Dremel in the toolpost would do the job? Not very rigid or precise. And I risk ruining what I have. Likely.
Thanks for your thoughts.
Best solution me thinks, is to buy better quality stuff, though I haven't seen it offered, and the quality USA guys don't appear to have it.
How about the 2 plate method? Buy a small face plate to fit the SB. Buy an ER 32 collet plate from anyone. Bolt the 2 together with alignment by me using the indicator.
Mosey

 

[lazylathe]

Thanks for the update Mosey!

Like i said earlier these are better known on the woodworking world.
0.003 run-out is nothing to worry about when making wooden pens.
As for metal working it is way too much.

Any ideas on what you are going to do?

Andrew

 

[Mosey]

Yes, the Beall is going back. If they can furnish one with <.001" TIR, then I will accept that. Otherwise, I will keep looking, either for a 2 plate assembly, or an affordable USA one. Keep you posted.
Bog, Are you aware that SB shows a metric 4 disk threading dial in their old catalog? Has anyone ever seen one? Couldn't I make one?
Mosey

 

[arnoldb]

Bog, Are you aware that SB shows a metric 4 disk threading dial in their old catalog? Has anyone ever seen one? Couldn't I make one?

Hi Mosey. The metric threading dial is for use on a metric pitch leadscrew, so unless you replace your leadscrew with a metric one, it would be a worthless exercise to make the threading dial. It's not the dial that determines whether you can cut metric threads; it's the ratio of the change gears and pitch of leadscrew that determines that. I screw-cut metric threads often on my non-metric lathe; this is done by simply selecting a ratio for the change gears that closely approximates the metric pitch needed.

Kind regards, Arnold

 

[Mosey]

Your'e right! I jumped the gun.
Now I just need a metric lat...........

 

[Blogwitch]

Mosey,

I do have to apologise for taking tooling as a normal possession, I have been collecting now for over 40 years, so can cover almost any eventuality, it just totally slipped my mind at this time as I wanted to help you so much.

I have seen people use Dremel tools for grinding tapers such as this. It would entail making a decent rigid two position holder for it, but once done, it would be a tool for life, whenever needed.

As for cutting metric threads on your imperial lathe. I only ever had one extra gear for doing metric, and I think that was a 63 tooth one, I know I didn't buy any extras from the normal set that came with my machine, so the previous owner must have bought it. If you have a full set of change gears, then at least some metric pitches should be easily within your reach. For a thread length to fit the ER collet nut, I reckon your pitch could be around 1 or 2% out, and the nut would still fit. But usually when cut on an imperial machine, they are only fractions of a % out.

I don't have any charts for your lathe, but I am sure there must be a member on here that can let you have a copy of their metric threading charts for your lathe. In fact, I think Marv has a program that will do the calcs for you.

BTW, I have a metric machine, so I have to do the same sort of thing for imperial, on the other hand, I do have a true conversion gear that gives me exact imperial, but I still have to do it with the half nuts closed. In fact I do my metric the same way, I find it a lot easier than a drop in dial when using a swing up toolholder.


John

 

[arnoldb]

Your'e right! I jumped the gun.
Now I just need a metric lat...........

Mosey, no need to get a metric lathe. From what I've observed you will most likely usually work in imperial, so it might be better for you to stay with the imperial lathe.
What lathe do you have, and what change wheels do you have available for it, or do you have a quick-change gearbox on the lathe ?
Given that information and the lathe's leadscrew pitch, I'm sure the members can come up with an idea for you to cut a close-enough 1.5mm pitch metric thread for the closer nut. The threads for the closer nut does not need not be super accurate; it's the taper in the chuck that provides the accuracy. If you have a single-point threading bit, you can use that, as being from the US, it will already be 60o.

I didn't respond earlier with a link to my own collet chuck build, as the ones presented give a lot more detail. But if you're interested, here is mine. I built that when I was a lot more inexperienced than I am now, so it was right at the limits of my capability - if not over those. There was a bit of heartache in that build, but the finished chuck has much more than made up for it; I use it nearly every shop session and it was worth every bit of effort. As it was built on my lathe for my lathe, it's much more accurate than a purchased one will be.

;D Well, John beat me to the reply - but I wasn't going to throw away a lot of typing. What John Said Thm:

Kind regards, Arnold

 

[mklotz]

As for cutting metric threads on your imperial lathe. I only ever had one extra gear for doing metric, and I think that was a 63 tooth one, I know I didn't buy any extras from the normal set that came with my machine, so the previous owner must have bought it. If you have a full set of change gears, then at least some metric pitches should be easily within your reach. For a thread length to fit the ER collet nut, I reckon your pitch could be around 1 or 2% out, and the nut would still fit. But usually when cut on an imperial machine, they are only fractions of a % out.

You probably have a 63 tooth gear because

80/63 = 1.2698... [only -0.0125% error relative to 1.27 = 2.54/2]

It's one of the best approximations to this value using only two gears.

I don't have any charts for your lathe, but I am sure there must be a member on here that can let you have a copy of their metric threading charts for your lathe. In fact, I think Marv has a program that will do the calcs for you.

The GEARFIND archive on my page contains several programs that allow one to find gear trains that yield high accuracy approximations to any input ratio. The 80/63 result above comes from one of those programs.

 

[Mosey]

You guys are just great!
I will take awhile to digest all of this important intellligencee.
I do not have any metric change gears, but will get some in the nearside, as my lathe has a QC gearbox. I suppose I could make the gears also.
I do have a rather complete SB library of metric threading, parts. accessories, etc., so I am steady in that area.
I have cut one metric thread, approximated, but it was a fine buttress thread, and it was junk. I will try again. I have an ongoing need for metric buttress threads to make collet devices for my F1.
What I am doing right now is buying a threaded backplate, to which I will mount a proper collet chuck to it.
I did think of something else...why not make an Imperial chuck nut and collet holder? As long as the internal taper is done right for the collet, why wouldn't an imperial nut (which I can cut easily) be just fine? I can harden it also.

Marv, thanks for the program.

Not to pollute this thread, but I am smiling from ear to ear today. I re-leveled my lathe and adjusted the tailstock setover. Made a test bar 6" long with no more than a couple of thou. taper. Then made a new crank bearing that has 0.00 measured taper from end to end. I did it in an unorthodox way. I took a pass the length of the bar. Measured the taper. set the tailstock over 1/2, and re-ran the length of it until I got it on the money. That little Last Word indicator didn't even shimmy as it went the length of the bar. I'm happy.




Mosey

 

[user 284]

Mosey,
I doubt there's all that many here that could or are at least setup with the equipment to make a proper collet nut. Since the collets get snapped into and out of the nut. Your looking at some pretty tight tollerances on a hardened blank nut including eccentric grinding, taper grinding to fit the very end of the collet, And even the internal threads appear to be thread ground on my Bison nut. It's not an impossible job, Just very difficult. A pretty high end and rigid toolpost grinder along with a lot of practiced skill would be needed also.

Pete

 

[Mosey]

Whoops! Of course. If it were easy, the cheap ones would be accurate. But they are not.

 

[user 284]

Yep, And there's even more complicated ones than the standard type too. A internal ring of Ball bearings inside the collet nut. I haven't used one ................yet. But they really help with getting the nut tight enough. ER-40s for example take at least 80 plus ft. lbs on the nut for proper tightness. Around $80-$100 just for the nut.

Pete

 

[Paulsv]

You might give these folks a call, and ask what kind of runout they will guarantee. A bit spendy, but I have heard they do pretty good work.

http://shop.tallgrasstools.com/product.sc?productId=15

 

[Mosey]

Spendy, right. I'm waiting for their email response.

 

[steamer]

How about 47/37?

1.27027....not quite as good, but it works just fine on my lathe.

I have extra 16 dp gears cut this way...but I think the SB is 18 dp

 

[Mosey]

From the factory parts list, 12/1995


The Quick Change Metric Transposing gear set furnished for a SB 10k with gear box included the following gears:
Metric Transposing Gear 27 and 100

Change gears
48
44
36
32
28
26 teeth.

Plus charts, index plate, etc.

So how does this relate to the ratios mentioned?

Well, I think that they meant 100/127 gear pair, not 100/27, even though that is exactly what their brochure says.

 

[steamer]

The only problem is the physical size of the gears. A 100/127 combination is an exact match (1.270000) . For a thread that usually never exceeds an inch in length....OK 25.4mm length....the error is so small as to usually be ignored

A 47/37 will fit in the gear guard, and has a ratio of 1.27027 or put another way, it has an error of .00027mm per revolution of the leadscrew

.00027 mm is .0000106" per revolution of the leadscrew or about 11 millionths of an inch/rev..........I don't know about you but I can measure that lead error...... period!

I can bet you that your threading tool has way more form error than that.

Dave

 

[Mosey]

ALERT! ALERT!
Chronos has them but cannot give runout figure. XXXX not interested.

RDG Tools has them. Waiting for runout figures.

Tools4cheap has the adaptor plate and holder style, bolt the 2 together. If single holder doesn't pan out, I will go this way.

 

[Mosey]

I received the backing plate and ER 32 collet holder today from Tools4cheap. Very quick response. The collet holder is pretty nice, and the backing plate is a nice, fine-grained cast iron. I machined the hub on the backing plate to fit the holder with a tiny amount of play for adjustment. Mounted the holder, and tapped it into alignment. It now has a couple of ten-thousands of runout inside the collet cone. That is OK for a unit that cost $125.00. Now, stop the BS, and make things.

 

[steamer]

Glad it worked out of you Mosey!

I turned my Bison 5C into an adjust true in a similar way that you did yours and have been glad that I did it.

Dave