PDA

View Full Version : Machining a Casting



projectnut
01-26-2012, 04:45 PM
I got the castings from Andy Lofquist for his MLA-21 collet chuck yesterday. Unfortunately it opened a whole new world of machining. I've never cut any rough castings, and I'm having a terrible time getting the thing to run true. The instructions say "special care should be taken to assure all body features are running as true as possible".

I put the body in a 4 jaw chuck expecting to be able to indicate it to within a few thousandths. That's when the trouble began. I can get 1 surface to within .020 but that's as far as it goes. I move to another surface and it's a good .100 out. Moving to yet another surface it's out .060 in another direction.

When holding the thing in my hand I can easily see the outer flange doesn't run true to the body, and the nose is oval shaped. I'm sure this is typical of rough castings. I just don't know which surface I should use as the critical reference. Any help would be appreciated.

Thanks, projectnut

Mcgyver
01-26-2012, 05:13 PM
what they mean is that you have to be like Michelangelo looking at piece marble and see the machined part inside of the casting, ie you want to make sure all of the machined elements are more or less centred. As no casting is perfect its a exercise of compromise. For example it might not be possible to have all the bores perfectly centred in their bosses, but whats the best compromise...what you don't want is one perfectly centred and the next breaking through the side...that sort of thing

you've got to survey the casting, measure, figure out what centred means then decide on a datum surface and figure out the sequence and clamping. This is head scratching and layout work.

Castings are much difficult to machine than bar stock where everything is neat and easily held

Black_Moons
01-26-2012, 05:24 PM
Well, Its not so much that you need it to run true, its that you need each feature to run true enough that theres enough metal to remove

Ie, if you need a 2" finished diamiter, and you measure the casting as 2.5", you can have up to 0.5" runout before its a problem (Other then making interrupted cuts in cast iron dulling your tools making everything take longer)

As for what feature you use to 'mainly' true it up, Id say you'd first measure the casting and figure out where you have the least leeway, Then true it up to that, and check some of the other surfaces to make sure they still have some leeway, And if not, you can steal some leeway from other surfaces to make up for it.

lynnl
01-26-2012, 05:35 PM
Well, I've never machined a casting either, so I'm not qualified to advise on that.
But I did build that collet chuck from bar stock a few years ago. I'm can't say for certain from where Mr Lofquist drew his inspiration for his castings design, but I'd guess it was from an article in HSM by Pat Loop, which is the design I used.

I don't know how much 'meat' is available for waste on your castings, but recalling, as best I can, my experiences, I'll offer these thoughts.

1) First, get the backplate machined to marry up with your spindle. But don't yet cut the groove (race) for those balls to run in. Best to leave that until such time that you are prepared to cut that groove in all four mating surfaces. That is to insure all four races are on the same circle and concentric to the spindle axis.

2) Save the boring on the collet bore and taper until last. You want all the other pieces fitting together and fitting to your spindle by the time you do that.

3) Most, or maybe all, of that exterior machining would also best be done after you have it running on the spindle. (By "exterior", I'm excluding the mating surfaces.)

Actually, I guess if I had that casting kit, I'd go grab that Pat Loop article and study it carefully and thoroughly before proceeding.

(edited to add the following correction)
Re my item 3) above: Instead of "exterior machining" I should have said "exterior FINAL machining." As rkepler points out (below) you will likely (almost certainly) have to do some cleanup machining passes on some of the exterior surfaces in order to properly grip and align for other, subsequent operations. Initially you'll want to keep those to just minimum, skimming cuts; just enough to achieve a true surface.

tdkkart
01-26-2012, 06:17 PM
Don't be afraid to call Andy and talk to him about where the best place is to start. Andy is cool, loves to talk about this stuff, and is a wealth of information. Call him.

RussZHC
01-26-2012, 06:21 PM
No help but interest as I am 99% certain I will end up later this spring with the S-4382 cross slide. I suspect there is minimal amount I can personally do due to lack of skills and or equipment (though the latter may change)...but to me it is still a very good option, even if someone else ends up doing the work with one of the reasons being Paula has a very well documented build over on PM.

I do not know how well or how much documentation these pieces have but as I understand it, there are three main casting parts, so I would start with the backplate and work forward, fitting each piece to the last...after, at the very beginning as others have said, checking the documentation and seeing how all the bits are supposed to interact together...just saying, if it were me, I'd be most concerned with surfaces and fastening points as opposed to cosmetics...I've seem lots of things like pulleys and gears where a hole for example is nowhere near the center of the boss around it, of course it does not look all the best but its got to work first...

Richard Wilson
01-26-2012, 07:03 PM
Castings need marking out first before you get them on the machine, so you can decide how best it works. We used to use whitewash, but these days I use Tippex (typists correction fluid), mark out the centres of holes etc, then mark on the diameter of the bores, and dot the circles with light punch marks. Then put it on the machine.

Richard

lynnl
01-26-2012, 11:14 PM
Projectnut, here's some discussion on that collet chuck, and other's issues in building it, that you might find interesting.
http://www.practicalmachinist.com/vb/south-bend-lathes/mla-loop-style-collet-chuck-195959/

Do the two castings come with cored rough holes, or are they just solid chunks?

Dr Stan
01-27-2012, 01:52 AM
The term I learned when I first began to machine castings was "balancing" the casting. In other words setting it up so that you had roughly equal amounts of material to remove to make the finished surfaces.

I never used a dial indicator on a casting, but instead used a pointer like the scribe from a surface gage, or even a piece of welding rod.

As others have said machining a casting is very different from machining from bar stock, especially sand castings. I have known machinists who worked their entire lives in the trade and they never machined a casting. On the other hand I have machined castings ranging in size from those for a 4" angle vice up to & including the main casting for chain making machines, essentially a horizontal punch press.

So, take your time setting up your parts. The comment about a sculptor looking at a piece of stone is very appropriate. Plan your cuts from beginning to end and you should be just fine.

projectnut
01-27-2012, 08:43 AM
Projectnut, here's some discussion on that collet chuck, and other's issues in building it, that you might find interesting.
http://www.practicalmachinist.com/vb/south-bend-lathes/mla-loop-style-collet-chuck-195959/

Do the two castings come with cored rough holes, or are they just solid chunks?

Thanks everyone for the assistance. This is a whole new world to me. A friend of mine worked his first 5 years in a shop that machined castings for pumps used to remove water from mines. He always complained about it being nasty work. I thought he meant because the castings were huge and machining cast iron is dirty work. I'll bet there was more agrivation associated with trying to determine how to set the things up rather than how to machine them.

The casting for the back plate is a solid chunk with a raised section that will register with the body, and a dimple where the spindle mount willl be bored. The collet nut is also just a blank piece of metal.

The main body has a hole cored through it and all the exterior features. Most of the exterior features will only need some clean up but the interior will essentially be bored from scratch. To give you some idea of how much metal needs to be removed the raw castings weigh 22 lbs. The finished chuck weighs 9 lbs.

Mcgyver
01-27-2012, 09:27 AM
A friend of mine worked his first 5 years in a shop that machined castings for pumps used to remove water from mines. He always complained about it being nasty work. I thought he meant because the castings were huge and machining cast iron is dirty work. I'll bet there was more agrivation associated with trying to determine how to set the things up rather than how to machine them.
.

There's some amusement in seeing the revelation occur in another

Often when the uninitiated are looking at something you made, say an engine, that involves castings, it will dawn on them that you didn't make the castings and they'll ask. "Oh the casting were bought, got them from such and such" Of course they don't say so but I suspect from the look on their face they're thinking "ah ha, that part can't be to difficult because so much is already done for you - the casting is already nearly the right shape" :D

I said earlier castings are more difficult. I take that back, none of this difficult. Difficult implies some can, some can't and there'd be a high failure rate. Anyone can do it - what it does require tough is some new extra steps and some thought - ie planning.

Do do the layout on it. Its essentially a planning exercise on how within this hunk of iron the final part will emerge, something the machinist has to plan for and control

rkepler
01-27-2012, 11:12 AM
I've found it best to go over the casting and make sure that there's enough meat for all the critical features to maintain their relationship when done. Sometimes the cope & drag will shift in relationship to one another and while there might be enough meat for the cylinder to be bored doing so will take you far enough out that the outside of the glange on the face won't fit or that there isn't enough room for the valve face. Sometimes things can be 'corrected' (more like 'fudged') but if there's an opportunity to swap out the casting for a good one you're usually better off doing so.

After that's done I'll choose a reference surface - that's some feature that I can make an initial cut on that everyone else will relate to. On a steam cylinder it's usually the valve face - the cylinder is perpendicular to that and it's a convenient face. Setting up for the first cut can be interesting with dodgy setups that only allow a minimum cut, but once you're done the finished surface can be the clamping surface and progress proceeds. Sometimes the surface you want to use as a reference cannot be reached until a couple others are roughed in, that's OK, just make real minimum passes just barely taking off scale so they're useable, cut the reference, then clean things up with the rough machining.

I always rough machine the casting first, never taking it to finish dimensions until after I've roughed in all the critical dimensions and have all the relationships established. You never know when you're going to find a rough spot or a blowhole in a casting that requires things to shift over for a cleanup. Once you know that you have good, clean material under ever machined surface you can finish things up.

Workholding for all this can be creative. I've mounted castings inside right angle plates put on their ends in the mill to make a reference bore. I've shimmed with aluminum strap, lead sheet (careful as that stuff moves on you), cardboard. I've heard of folks sticking Bondo around a part and using clamping against the bondo. I've made mounts to hold the part (a brake shoe) out in the air in the right position, then centered the mount on a faceplate to make the cut at the right radius and taper.

Finally, just walk through the process from start to end in your head before cutting and you should be fine. If there's any fault in the MLA castings it's that there's a surplus of meat in them to remove (but really, that's a whole lot better than going the other way).

On edit: Forgot to mention this - even if you keep a vacuum going to take away the cast iron dust you're going to be spitting black boogers for a while. Get use to that and the taste of rust whenever you wash, it's part of machining cast iron.

SGW
01-27-2012, 03:24 PM
+1 to what rkepler said.

It's been my experience that Andy's castings are really nice to work with and have plenty of machining allowance. Take your time figuring out where the part is inside the casting and you'll be fine.

As suggested, I'd establish some sort of reference surface first to relate other dimensions to. Knock the high spots off the to-be reference surface, set it up for machining as flat as you can get it -- you'll probably need shims of some kind -- and take a skim cut so the reference surface just cleans up. You may have to take more off later, but leave some extra for now. Then set the casting on your surface plate on the reference surface and start measuring for the location of the other features of the part. It will end up being a set of compromises and averages.

Yes, holding a casting can require some creativity. As long as what you come up with works and is safe, anything goes.

projectnut
01-28-2012, 04:34 PM
I measured the main body casting on the bench today and it looks like there's plenty of material on all the critical surfaces. Everything is oversized by at least .125" and in some cases as much as .250"

I did take a stab at trying to determine the high spots. I chucked the main body in a 4 jaw and indicated each jaw. I was able to get the piece centered on the chuck to within an .002 difference between each of the jaws. The thought was that if the piece could be centered then the uneven readings on the surfaces needing machining would be due to the excess material.

After centering the part I started indicating each of the surfaces. If I'm doing it right it looks like all the high spots are roughly on the same side and range from .080 on the spindle nose to .120" on a non-critical flange.

Is my logic and measuring procedure sound, or am I missing something obvious?

As a note I did look at Pat Loops original article where he made the colllet chuck from round stock. His original is slightly larger in diameter and length than Andy's version.

SGW
01-28-2012, 05:10 PM
If I'm understanding you correctly, I think you are okay. Sleep on it for a night or two, then go back and see if it still makes sense.

1/8" to 1/4" machining allowance sounds about right for Andy's castings. He's generous.

lynnl
01-28-2012, 07:34 PM
Well, I've never machined a casting either, ....

I don't know how much 'meat' is available for waste on your castings, but recalling, as best I can, my experiences, I'll offer these thoughts.

1) First, get the backplate machined to marry up with your spindle. But don't yet cut the groove (race) for those balls to run in. Best to leave that until such time that you are prepared to cut that groove in all four mating surfaces. That is to insure all four races are on the same circle and concentric to the spindle axis.

2) Save the boring on the collet bore and taper until last. You want all the other pieces fitting together and fitting to your spindle by the time you do that.

3) Most, or maybe all, of that exterior machining would also best be done after you have it running on the spindle. (By "exterior", I'm excluding the mating surfaces.)

Actually, I guess if I had that casting kit, I'd go grab that Pat Loop article and study it carefully and thoroughly before proceeding.

(edited to add the following correction)
Re my item 3) above: Instead of "exterior machining" I should have said "exterior FINAL machining." As rkepler points out (below) you will likely (almost certainly) have to do some cleanup machining passes on some of the exterior surfaces in order to properly grip and align for other, subsequent operations. Initially you'll want to keep those to just minimum, skimming cuts; just enough to achieve a true surface.

I'm going to amend the thoughts that I offered before (above).

I would NOT start with the back plate, but rather the main casting (nose or front piece); that's going to be the most challenging and complex, AND by doing it first, That will leave you some latitude for any compensating adjustments you may need to make in the back plate and collet take up nut.

I would chuck it up, gripping on the back (larger) end, and get the imaginary line representing the axis of the bore running true and on center, ... calibrated eyeball would be sufficient here. This is in preparation for taking a uniform clean up pass on the necked down front.

I would then satisfy myself that the clean up of that front portion will still leave plenty of meat all around the final bore once it's bored out, AND THAT the back portion will likewise still have sufficient thickness all around, when I later reverse it in the chuck and machine the back end concentric with the front.

Once I'm satisfied I'm not headed for disaster, I would then take that clean up cut on the front, smaller end.

Then I'd turn it around (reverse it, end for end) in the the chuck, indicate that front end so it's running true, and commence work on the back end.
First a light clean up of the exterior, to include facing off what will be the rear surface that will ultimately mate up with the back plate. Then do all of that interior work, for the collet nut, collet key holder, and ball bearing race, ...ALL IN THE SAME CHUCKING, and to final dimensions.

So at this point I'd know my internal, mechanical features are concentric with that back end, and the exterior of the nose is at least PDC (pretty damn close)

I'd make it a point to keep track of the diameter I machined the ball race, so I get all four exactly the same. Though, as Pat Loop's article points out, if you machine that into a sort of "(" (parentheses mark) shape, rather than a pure semi-circle matching the ball diameter, then a couple of thou off probably wouldn't matter.

I welcome any and all to critique my thoughts here.

Please keep us all updated on how it's going with this Projectnut! I'm eager to hear.

(added) Oh yeah, I forgot one thing: When I reversed it to start working on the back end, I would drill and bore the through hole (but not to final dimension), before opening up the cavity and ball race, etc. Then I'd use this initial bore to indicate from when I do turn it again to work on the front end. Or I guess if there's already a hole there, skip the drill and just bore a true undersized hole.

RussZHC
01-28-2012, 10:23 PM
I treat all threads here the same way to start, as part of a learning process.


I welcome any and all to critique my thoughts here.

So a question...I have previously asked a few questions about boring and IIRC one of the replies was to do the exterior and then the "interior" as in finishing the bore...but my understanding is that there are few hard and fast rules...why would it not be "best" in this case to do something very close to the finish bore first and then set-up the rest, if possible on a mandrel, based on that bore?
With the idea being that bore is all important since isn't the whole purpose of a collet system to be as accurate as possible to the axis of the spindle...

I like the idea of doing as much work as possible in a single chucking since I think I can understand that purpose re: concentricity.

lynnl
01-29-2012, 12:31 AM
...why would it not be "best" in this case to do something very close to the finish bore first and then set-up the rest, if possible on a mandrel, based on that bore?
With the idea being that bore is all important since isn't the whole purpose of a collet system to be as accurate as possible to the axis of the spindle...

I like the idea of doing as much work as possible in a single chucking since I think I can understand that purpose re: concentricity.

Well in this case, for maximum accuracy, the final boring will need to take place after the front piece is mounted onto the back plate, which is, in turn, already mounted to the spindle. i.e. in situ, just as it's going to be operated on the lathe. Also, the final boring must include a taper at the mouth of the bore, corresponding to the 5C closure taper.

In other words, the final boring needs to be done from the front. Yet there does need to be a bore in place, in order to carry out some of the cavity machining on the back side. And it would be difficult (very difficult for me at least) to do that with it mounted on a mandrel.

That's my view at any rate. But I'm certainly not the ultimate authority. Maybe there is a better way.

(added) Russ, if your comment is aimed at getting the exterior turned true, then yes, the mandrel idea would work fine, but those cavities and ball races are essential to the accuracy, more so than the exterior. PLUS, the exterior can also be made as perfect as possible once it's mounted on the spindle, just as with the final boring.

dp
01-29-2012, 03:29 AM
Sometimes it is as easy as making a cardboard template of the first surface you intend to machine and smacking the center with a punch. Set it on an RT and find the center of the opposite face with another template.

Deciding what that first surface is going to be has to take you to the next surface - usually at 90 from the first, etc. Like in chess - know the next couple moves. The entire part you want to make is completely inside the casting - no exceptions.

NzOldun
01-29-2012, 03:29 AM
ProjectNut,

One thing nobody else seems to have picked up on!

In very likely you'll need carbide tipped tools to get through the 'crust' of the casting, particularly for 'skim' cuts. Once through the crusts, you can then switch to HSS, if that's all you have, but be prepared to have to sharpen your tool bits much more often than you would for steel!

NzOldun

projectnut
01-29-2012, 11:24 AM
I do have both carbide and HSS tooling. I did take a couple passes this morning with the HSS and I believe carbide will be the way to go. As far as taking my time that's not a problem. I'll get some time to work on it tomorrow but that's going to be it for a while. I have to be out of town for a few weeks so I'll be able to give it plenty of thought before I do anything potentially stupid.

My laptop will be with me so if I have questions or possible brain cramps Ill be in touch.

gizmo2
01-29-2012, 12:44 PM
I'm on my second steam toy, so take this as newbee advice. But here's the method that works for me. I figure to start on the centerline, start with the hole. Let's say the bore is supposed to end up at 3/4". I will knock off the flash and rat tail file that hole out until it will accept the next piece of drill rod. Let's say it will initially pass a 1/2", but not a 9/16". I will worry that out until the 9/16 rod will just make it through. Mount her up between centers (I made a pecial short center for this) with the part riding along, then chuck 'er up with the four jaw and shims cut from a Pabst Blue Ribbon Beer can. Other brands may work. Sneak up on the tightening to minimize deflection. If you did it right, the rod will now pull out and you can proceed to bore that center hole to spec. A new mandrel will keep things spinning around that center line, and keep the casting from distorting too much during future clamping set ups.

This is based on the faith that the casting isn't too wonky; maybe I've been lucky so far but this has worked better for me than eyeballin' the center line.

Bazz
01-29-2012, 01:03 PM
ProjectNut,

One thing nobody else seems to have picked up on!

In very likely you'll need carbide tipped tools to get through the 'crust' of the casting, particularly for 'skim' cuts. Once through the crusts, you can then switch to HSS, if that's all you have, but be prepared to have to sharpen your tool bits much more often than you would for steel!

NzOldun

Before the carbide and HS era every thing was made with cast iron and machine with forge carbon steel ,it amazes me the nice work they did back then with carbon steel tool so high speed tool are very capable to machine cast iron

Turn slow get under the crust with a big cut in 1 pass.
Yes carbide will go better but not a must

Mcgyver
01-29-2012, 02:05 PM
Before the carbide and HS era every thing was made with cast iron and machine with forge carbon steel ,it amazes me the nice work they did back then with carbon steel tool so high speed tool are very capable to machine cast iron

Turn slow get under the crust with a big cut in 1 pass.
Yes carbide will go better but not a must

+1, although don't always try to get under it in one cut. Until there are good datum surfaces, the casting may be held at a few points so set ups are less solid than you'd expect with bar stock. Use care before before trying a big cut. whether you can get under it in one pass will depend on the casting. If you can't, hss is also better at an interrupted cut

whats the worse that can happen? you use 1/100 of the hss tools life and sharpen afterward

SGW
01-29-2012, 06:06 PM
In this case, don't worry about it. When I got Andy's T-slot cross slide casting, it was clean and must have been annealed because it had no skin to speak of.

rkepler
01-29-2012, 06:46 PM
Even if there is no skin the little nerds from sand casting as well as draft, flash and parting lines will all make things king of slippery in the vise. Use cardboard or aluminum undeeer the gripping jaws to deform to the casting surface and give you that extra bit of grip.

Oh, yeah, I run a large coarse file over everything that might be machined later just to clean things up. Sometimes that gives you a better feel for what and where it's going to be gripped.