Triple spider idea to remove runout?

[GummyMonster]

Morning,
I'm reposting my question, as my other thread question was discussing a topic that I unknowingly shouldn't have .

Ok,
Would the triple spider idea below ( taken from another discussion) work to remove any latent runout in a W210 style lathe?
To make sure your piece is EXACTLY centred with the tailstock tooling.
I'm very new to this, but I'm thinking it could work to remove any runout in the machine .
Thanks,
Ken

 

[RobinHood]

I think that that may be hard to do.

I would try and center the part in the spider to run perfectly true in the spindle bearing (the rotating part) and then center the tailstock tooling to run concentric with the bore of the part.

 

[cuslog]

I don't use a triple spider. I do have a spider in the outboard end of my spindle bore - 4 X 1/4 - 20 SHCScrews w/ brass tips.
On the chuck end I use a 4 jaw with a piece of 1/4" soft copper bent around the work piece, resulting in very narrow contact at the chuck end as well. If you "sneak up" on the tension (switch ends a couple or 3 times), I think you can set a workpiece up to run true on both ends without flexing in the middle.

 

[Mcgyver]

Thats a cats head, spider is usually something and the far outboard side of the spindle used to hold long stuck so it doesn't whip (although I get that nomenclature is tossed around to the point where there may not be universal agreement on that.)

Why 3? Seems any more than two and they'll be fighting each other. A cats head is a horrific pain to set up, probably why they are hardly ever seen lol. 4 Jaw is about 100x easier. Like cuslog says, there are ways to set things up in the 4 jaw and get them dead on in two planes. I just wrap the piece in copper, thin stuff like flashing, and as things just to get snug, I starting do iterations in two planes. get it spot on at the chuck, move out, indicate and this time mostly gently tap it true. go back to chuck, rinse, repeat. When you get it true in the 2nd it won't be in the first, with iteration, if done carefully with taps and very minimal jaw adjustment, the error between the two planes keeps getting smaller.

A lot of people just worry getting it concentric at the chuck, but thats not good enough for something were true axial alignment really matters. It's also why a good quality 4 jaw is worth it....if dead on at the chuck it will be close in the second plane

 

[GummyMonster]

Thanks for the great information.
Only reason I chose a triple cats head (spider) is that's the one that was being discussed in a different thread.
My lathe actually came with both 3 and 4 jaw chucks, but I really haven't spent any real time checking for accuracy of them. I've been too busy trying to finish up all the other things that require my attention so I'm focused when I start dedicating my days to trying out my lathe and milling machine. Right now I'm trying to absorb as much knowledge as I can.
Thanks again for your proper direction.
Ken

 

[Susquatch]

I'm a bit late to this discussion but can add some insights as to why a cat's head VS a 4 jaw and why three sets of screws.

The purpose of the cats head is to allow AXIAL alignment of parts that are too short to reach through the headstock to a rear spider. Any one set of jaws can only align the center of the part at the Jaws themselves. By providing more than one set of jaws, it is possible to align the center at each set thereby achieving axial alignment with the spindle.

The reason for three sets of screws in the cats head is to accommodate shorter parts. Normally only two sets are used at any one time. Whichever two sets provides the longest holding length. Distance is your friend on this task.

I too like to use short copper ears and sometimes even copper wire in my four jaw to allow the part to swivel with any axial alignment via the rear spider. But where parts are too short to reach a rear spider and where axial alignment is important, nothing beats a cats head.

 

[PeterT]

I don't really get what you mean Susquatch. A single 4J jaw (red) will contact the part all along its tangent on each of 4 jaws to part (green). So doesn't it have more part contact area than the (orange) point contact of screws at their spacing, short or long?

4J can displace the part in either 90-deg plane until the part axis is concentric to spindle axis. Whether that is ID or OD is wherever you spot the DTI. The cat head I guess does the same thing through point contact on each screw, but how would you know the left side of part is axial to front unless you get in there with a long DTI? I can see an advantage if part is irregular or tapered, the screws would better accommodate over 4J. Is that what they are used for? (I don't have one but that doesn't mean anything).

 

[PeterT]

My lathe actually came with both 3 and 4 jaw chucks, but I really haven't spent any real time checking for accuracy of them.
Ken

The key thing is to recognize is re-chucking repeatability. You can chuck a potato in a worn out 3J and it will turn a perfect round because its spinning about an axis against the cutting tool. But that's where it ends. If your 3J jaws / clamping repeatability is out say 0.003" which isn't unheard of, and you grip an absolutely perfectly ground cylindrical part, it will incorporate that same runout. Therefore turning the end or boring a hole of that perfect part will not be concentric to the OD. Every time you re-chuck its a new adventure. They do make adjustable backplate types of chucks & collet chucks to help with this but its a specific system.

A 4J (better to say 'independent jaw') chuck forces you to grip the part and dial it in simultaneously every time, typically with a test indicator on a surface so that it reads zero when its all tightened up. Now any secondary operation will be concentric because you have set it that way. And of course you can grip (non round) rectangular stock too but that's another thing.

If you have bad jaws in either 3J or 4J chick, all bets off. By bad I mean if the jaw grip surface is not parallel to the chuck axis, then how it grips at the front will be different than the rear & worse yet if jaws themselves are different. That's why they are hardened & then accurately ground & why hey cost moola.

 

[Susquatch]

I don't really get what you mean Susquatch. A single 4J jaw (red) will contact the part all along its tangent on each of 4 jaws to part (green). So doesn't it have more part contact area than the (orange) point contact of screws at their spacing, short or long?

4J can displace the part in either 90-deg plane until the part axis is concentric to spindle axis. Whether that is ID or OD is wherever you spot the DTI. The cat head I guess does the same thing through point contact on each screw, but how would you know the left side of part is axial to front unless you get in there with a long DTI? I can see an advantage if part is irregular or tapered, the screws would better accommodate over 4J. Is that what they are used for? (I don't have one but that doesn't mean anything).

Your drawings are great. You went to a lot of trouble to do that and I am impressed.

I think you actually worked out one of the many issues that a cats head addresses with your second drawing.

But I think I can expand your thinking a fair bit by suggesting that you picture that cats head about a foot long instead of so tiny as you show it. Each set of jaws (jack screws) on mine are about 4" apart. Alignment is usually accomplished with an internal dti or a precision mandrel.

Another way to look at it is to think about a cats head as a smaller lathe head with a shorter and perhaps bigger internal diameter spindle and a much closer spider on its rear.

The overall purpose is to facilitate axial alignment of a part that could not be accomplished with a two inch long fixed surface jaw and a two foot long spindle/head.

I totally agree with your subsequent post. If you extend that thinking a bit and imagine that external diameter of the part you are chucking is not perfectly round or concentric to an internal diameter - say for example a rough cylinder with an internal eccentric bore that is not just offset but also on an angle and you need to add features to the internal bore. Suddenly, you have an intense longing for a cats head!

I think most machinists have no need for a cats head. But when and if they ever do, they are a blessing for sure!

 

[Susquatch]

FWIW, someday I'd like to get myself a really nice smaller four jaw. One that isn't so heavy that I have to think about it before mounting it on my lathe. I LOVE my Bison Collet Chuck mostly because it is so small and so light and so damn perfect. When I got it the only thing I did was to rotate it in the cam locks from position to position until I found one that ran repeatedly true enough (turned out to be less than a thou), marked that position, and have thoroughly enjoyed it ever since. I usually go out of my way to find a way to use it instead of my 3 or 4 jaw chucks. I LOVE IT!

The other thing I'd like to do someday is to make myself a jaw grinder. But right now my priority is to get both my mills working enough to decide which one to keep.

 

[Mcgyver]

The key thing is to recognize is re-chucking repeatability. You can chuck a potato in a worn out 3J and it will turn a perfect round because its spinning about an axis against the cutting tool. But that's where it ends. If your 3J jaws / clamping repeatability is out say 0.003" which isn't unheard of, and you grip an absolutely perfectly ground cylindrical part, it will incorporate that same runout. Therefore turning the end or boring a hole of that perfect part will not be concentric to the OD. Every time you re-chuck its a new adventure. They do make adjustable backplate types of chucks & collet chucks to help with this but its a specific system.

A 4J (better to say 'independent jaw') chuck forces you to grip the part and dial it in simultaneously every time, typically with a test indicator on a surface so that it reads zero when its all tightened up. Now any secondary operation will be concentric because you have set it that way. And of course you can grip (non round) rectangular stock too but that's another thing.

What I think Susquatch is getting at, and I what tried to describe earlier, is that you cannot depend on the alignment, even with a high quality 4j chuck. You can indicate perfectly in one plane but unless you indicate in two, there will be wobble - i.e. work and lathe axis intersect where you indicated but are still not parallel. It'll be close with a good chuck, and good enough for lot of /most work, but it needs tweaking for exacting work.

For example, take something know to be round and straight to a high degree of accuracy. Say a cylindrical square (wrap it in something obviously to not mar the surface). Use your best tenths indicator and dial it in to the best of your ability, less than a tenth and barely a flicker of needle movement. Now, roll carriage out 3 or 4 inches and indicate. I can almost guarantee it will not be as concentric as it was at chuck. Photos below show what is meant by indicating in two planes - one close to the chuck the other further out. This is a spindle housing so bores for the bearings (expensive P4s!) need to be true - and each end needs to set up and bored. The only way to ensure those bores are aligned is reference the OD and indicate it in in two planes. Its a Rohm chuck so is good quality, but its still needs tweaking to get things spot on.

There are ways to to deal with it, and its somewhat of a pita and takes some time - basically iterations of dialing it in in each plane as well as judicial tapping.

I don't agree with the idea of the spider being for alignment, they're to discourage whip afaik, but the cat head can achieve perfect alignment. The one I have and have tried to use (and its a pita) is on a watchmakers lathe. making one is simple and a great trick if you don't have 4 jaw, but with a 4 ja just being for different lengths of work and only two are used at a time - hadn't thought of that!w I'd never bother with the thing. Susquatch makes a great point on the three screws;

 

[Susquatch]

What I think Susquatch is getting at, and I what tried to describe earlier, is that you cannot depend on the alignment, even with a high quality 4j chuck. You can indicate perfectly in one plane but unless you indicate in two, there will be wobble - i.e. work and lathe axis intersect where you indicated but are still not parallel. It'll be close with a good chuck, and good enough for lot of /most work, but it needs tweaking for exacting work.

For example, take something know to be round and straight to a high degree of accuracy. Say a cylindrical square (wrap it in something obviously to not mar the surface). Use your best tenths indicator and dial it in to the best of your ability, less than a tenth and barely a flicker of needle movement. Now, roll carriage out 3 or 4 inches and indicate. I can almost guarantee it will not be as concentric as it was at chuck. Photos below show what is meant by indicating in two planes - one close to the chuck the other further out. This is a spindle housing so bores for the bearings (expensive P4s!) need to be true - and each end needs to set up and bored. The only way to ensure those bores are aligned is reference the OD and indicate it in in two planes. Its a Rohm chuck so is good quality, but its still needs tweaking to get things spot on.

There are ways to to deal with it, and its somewhat of a pita and takes some time - basically iterations of dialing it in in each plane as well as judicial tapping.

I don't agree with the idea of the spider being for alignment, they're to discourage whip afaik, but the cat head can achieve perfect alignment. The one I have and have tried to use (and its a pita) is on a watchmakers lathe. making one is simple and a great trick if you don't have 4 jaw, but with a 4 ja just being for different lengths of work and only two are used at a time - hadn't thought of that!w I'd never bother with the thing. Susquatch makes a great point on the three screws;

Perfect compliment to my comments and spot on too! Along with nice photos to demonstrate.

Only one quibble. I replaced my factor rear spider with a custom made beauty. It is a joy to use when my work is long enough to be able to. The longer length make it easier to get excellent alignment. Alas, that is not always (or even often) possible, hence the joy of a good cats head.

Thanks for your input.

 

[PeterT]

Gotcha. I had no idea of scale, total length or end application from the initial sketch so made some (incorrect) guesses. Sounds like purpose fixturing for specific jobs vs. generic work holding. Would love to see some real pics of this kind of setup, more out of personal curiosity. One can always learn from the other team!

 

[Susquatch]

Here is a photo from the web.

Mine is on loan right now. When I get it back, I'll take a few photos for you as installed on my lathe.

I have no idea why this one has threads on the inside. Perhaps it was designed to fit a threaded spindle nose.