Machining castings

Hi,
I'm interested in learning about manufacturing process but there is a subject I cant seem to find resources online or at least cant find resources to clearly explain the steps involved in machining castings with multi operation setups. Like how datums are selected, how can variations in casting rough surface not lead to out of specs final part ..etc? Can anybody provide links to good videos or books about the subject?

Usually you plop it on the table and line it up by eye for op 1. Subsequent ops use whatever geometry was machined from op 1.

If the part doesn't have enough meat on it to hold tolerance from op 1, usually grabbing the designer and hitting them over the head with the part in question is the best way to avoid future headaches.

DMF_TomB posts a lot of photos of large castings being machined along with explanations; searching through his post history will give you more arcane knowledge on the subject than any book would.

In general:
Castings are clamped to the table and their oversized surfaces are face milled and machined in roughing passes first, then finishing passes. Large 5 axis gantry mills are best for large castings because they allow most of the reference surfaces to be done in one operation. Moving parts between operations induces inaccuracy and is bad for process reliability due to operator error.

For small and medium size castings, 30 taper machines (Robodrills and Speedios) with 4th or 5th axis setups are generally the fastest for high volume stuff that doesn't need a ton of material removal on each part. Horizontal mills are for when you need to remove a lot of material on a lot of parts.

Ive machined thousands of castings, as said above your eye is the best sometimes. But always check that if you machine a surface first will there be enough meat on other surfaces to cleanup. And depending on material if I take 30 lbs. of material off of a feature will it remain tru when its done.

And if you machine a feature first will I be able to safley hold it after, make it your head before you start.
I used to line up the non machined areas as best I could so there wasnt interference with bolting or installation problems.

Ive also seen more than one casting come out of a machine because where it was chucked it was literally being held by one tooth on a serrated jaw.
Sometimes you have to rough machine a feature just so you can hold it safely.
On Stainless casting we would rough them out and actually remove from the machine and let it rest for a day then finish it.

Sam,
Welcome to the forum. Castings are not difficult to machine and there are some tricks of the trade that do apply. Look the rough casting over and measure and compare what you got to what you want. Getting a starting surface is usually required to work from and sometimes a wooden wedge or so can help you hold it in a vise to machine that first cut. I almost always use carbide to machine cast iron. I try to get several features cut without moving the part. Then maybe making a simple fixture with a couple pins and clamps can make life easy. Often a flat and a reamed hole will give you a few options as to what to do next and how to locate, machine and measure the rest. I have 2 vises made especially to hold odd shaped castings and use them often. Also using a coarse file to remove any high spots and draft can help with holding the raw casting as you get started. A few pics.
spaeth

Thank you guys for your input. Please forgive my ignorance but if I'm correct I think I heard someone mentioning that casting designs usually imploy special features like small tangs for fixturing and possible to use as datum for op #1. Is that must or it's just a way to improve mass production?

If you will be doing a large number of castings check the dimensions of several castings, preferably spanning several days of foundry output. Castings can vary considerably in dimension. Don't base your fixturing on the first casting you pick up. Been burned by failing to do that.

Large castings and large fabrications are usually approached in the same way. They are set up on the machine and scaled in, then a centerline is established. They're laid out with a spindle mounted tool using the readouts. If there's anywhere that is found short of stock, centerline is shifted or the part shimmed to provide it. Often boards will be jammed into burned holes to give a center point for dividers to scribe a circle so that the burned holes can be checked for sufficient boring stock too. Castings are usually more regular so subsequent castings off the same pattern can be set up the same as the first, but should still be checked as noted above. Fabrications are considerably more varied so usually need to be laid out on every subsequent one unless they're produced by fixturing.

sam77 said:

Thank you guys for your input. Please forgive my ignorance but if I'm correct I think I heard someone mentioning that casting designs usually imploy special features like small tangs for fixturing and possible to use as datum for op #1. Is that must or it's just a way to improve mass production?

Click to expand...

They do, yes. When possible, all ops are done with but the one clamping, even if that means a whole array remain clamped, and the fixture itself is flipped or rotated.

Often the reference is just a boss with a flat surface, ELSE "the plan" is to make use of an existing feature that is reasonably stable and not "biased" - wrench flats on a water valve, etc.

Go have a look at cast Brass plumbing fittings in the open bins, any Big Box, and you should "get it in one" at no cost and very little time spent. See also IC engine blocks or transmission cases - or just about anybody's gear cases, but those take more effort to find and lay eyeball onto.

Many such starting points still benefit from at least a "light" hit with an angle-grinder or similar to assure no fins or surface bumps.

Nothing "new" here. Castings are about the first goods as ever WERE "machined" in any way or volume, even before gunpowder and cannons. Mega-millions still are, every hour of every day, most continents on-planet.

sam77 said:

Hi,
I'm interested in learning about manufacturing process but there is a subject I cant seem to find resources online or at least cant find resources to clearly explain the steps involved in machining castings with multi operation setups. Like how datums are selected, how can variations in casting rough surface not lead to out of specs final part ..etc? Can anybody provide links to good videos or books about the subject?

Click to expand...

.
.
not that complicated. usually need to
.
1) stock divide part. that is check extreme ends or key features to see if centered at expected position. not unusual to have a azimuth rotation needed or a side shim at one end cause longest length is not parallel enough to usually X axis. usually use spot drill as a pointer and use a big ruler or scale to get within .020". this can be automated with a probing routine of course. part often sits on 3 points and uses 2 side points to establish parallel to X and Y position and 1 point to establish X position. the 1st op machines surfaces later used on the next ops for location. height of part is sometimes checked but if extra what if big passes added its usually not necessary unless drilling on the rough casted surface
.
2) most castings use extra what if big passes. that is all rough milling starts a extra distance away encase casting is big in spots. ideally they should make castings within .100" but occasionally that are not. usually the bigger the casting the more potential error. over 10 ton and over 8 feet i normally watch all rough milling in case extra big in spots. not as easy as it appears cause big part easily not see the far end being machined.
.
3) castings are not usually that precise. thats is think of it as rather than a cube rectangle of metal weighing over 100 tons the casting already has 90 tons removed from the rough shape of the casting so its not 100 tons but 10 tons and maybe only needs 1/2 ton machined to get it to final shape. obviously with over 90% already removed by casting thats a lot of metal that doesnt need machining and certainty would be cheaper too that making 90 tons of chips to throw away.
.
iron/steel is about 500 lbs per cubic foot. when parts are 2x2x2 foot cube thats 8 cubic feet or 4000lbs obviously as size gets bigger the weight gets bigger. a casting 2x2x2 foot might easily be only 400 lbs or less cause rarely is part a simple cube shape

Biggest problem we've had with castings is the machinist trying to treat them like another saw-cut piece of metal stock. Then when things come out wrong, the excuse is it was because the casting is off or it must have moved!

Old casting drill jigs will often hold the casting from several points that average out the discrepancies in the casting. There isn't really a home position on the part, but there will commonly be one on the fixture to set your tool/machine too. You would then drill or mill a number of features without removing it from the fixture and those features become your datum points for any follow up jigs or operations. In the end, whatever variances there are between castings don't matter because all of the machined features are positioned correctly.

Depending on the fixture, there will also be some areas that the operator has to ensure that the first cut is cutting material, and if not adjust the program or re-set the casting in the fixture.

sam77 said:

Hi,
I'm interested in learning about manufacturing process but there is a subject I cant seem to find resources online or at least cant find resources to clearly explain the steps involved in machining castings with multi operation setups. Like how datums are selected, how can variations in casting rough surface not lead to out of specs final part ..etc? Can anybody provide links to good videos or books about the subject?

Click to expand...

Sam with all of 2 post's. Oregon here He seem's to be from Libya, over at the Zone.

https://www.*******.com/forums/general-metalwork-discussion/372356-manufacturing.html

Centrifuge manufacturer, if I've even seen one. Go ahead and blab state secrets Tom.
Of course the link doesn't work, Abene stealing pack of monkeys.

Try https://www.cnc zone.com/forums/general-metalwork-discussion/372356-manufacturing.html
And just take out the space between cnc & zone

actually its not much different that if you got a badly torch cut part where length and shape might be 1/4" off in spots or badly wavy and tapered, nothing with square or perpendicular straight flat sides so part could be different 0.500 in length, width, height in spots.
.
you need to stock divide or center part so you machine roughly equal amounts from all sides rather than .000 on one side and 0.500" on the other side. and a wavy torch cut part you cannot clamp easily in a vise with long straight jaws due to waviness.
.
often clamping is done by using 6 points. part sits on 3 points (to not twist it). uses 2 points for azimuth rotation lock and set Y position and a 1 point for X position and work offset is adjusted to confirm the far ends are centered at expected position

eKretz said:

Large castings and large fabrications are usually approached in the same way. They are set up on the machine and scaled in, then a centerline is established.

Click to expand...

As a patternmaker who has many times been called to a machine shop because. "There is no machining on this side". The Establishing a center line is the most important thing. Building patterns we almost always work from centerlines rather than edges. Modeling patterns to be CNC cut is often the same thing. The reason for this is because we have to add draft and machining. This meant that an edge is no longer a reliable reference point as the top of an edge is different from the bottom.

Sometimes the missing machining is our fault sometimes the drawing is bad and sometimes the casting shrank or warped unexpectedly. But most of the time I have gone to a foundry the part is in the casting and someone has machined too much off one side. Not saying we are that much less error prone but usually the layout or model has been checked then the pattern has been checked and any missing machining has been corrected.

For what it's worth, there's a decent introduction on per-machining layout for castings, with lots of photos, in K.M. Moltrecht's Machine Shop Practice, volume 1. So while they apparently don't teach this in CNC operator school, it's not like forgotten lore from ancient civilizations or something.

Pattnmaker said:

As a patternmaker who has many times been called to a machine shop because. "There is no machining on this side". The Establishing a center line is the most important thing. Building patterns we almost always work from centerlines rather than edges. Modeling patterns to be CNC cut is often the same thing. The reason for this is because we have to add draft and machining. This meant that an edge is no longer a reliable reference point as the top of an edge is different from the bottom.

Sometimes the missing machining is our fault sometimes the drawing is bad and sometimes the casting shrank or warped unexpectedly. But most of the time I have gone to a foundry the part is in the casting and someone has machined too much off one side. Not saying we are that much less error prone but usually the layout or model has been checked then the pattern has been checked and any missing machining has been corrected.

Click to expand...

Pretty sure you meant to say machining stock, but absolutely correct otherwise. I have absolutely seen this happen in many shops as well, and that's one reason I was taught to leave "witness" marks when rough machining surfaces - places that only just don't clean up - whenever necessary. That way it's clearly apparent that there was only just enough stock removed to clean up the surface. This is another reason that it's absolutely critical to do a full layout on anything bigger than mid-sized work. (By mid-sized I mean anything bigger than would fit in a 3' cube). Very helpful to work on a turntable if possible also.

That full layout will also mostly prevent the "extra big" problem described by Tom. If you mark out the stock everywhere you can see where there's more or less and touch off on the high point. Sighting along the surface to a reference edge (like the front of a table or a t-slot edge) will tell you if there's much of a bulge in the middle, and then you can touch off there instead.

I fear we have tread far outside the bounds of the topic that the OP is originally worried about though. Shops with equipment large enough to do that kind of work are not too common these days.

I spent a few years on marking out and inspection.
Raw castings would be put on the marking out table. We painted the relevant areas with " Marking Out Whitening " and marked out the finish cutting lines.
It made more sense economically to mark out castings away from the large machines when even then the rates for these machines could be £100 per hour.
The machinists then set up the castings to our marking out lines.

You had to evaluate where the cut lines needed to be. That wasn't always straight forward and you could start out working from one datum and part way through the job you'd have to paint out your lines and start again because other faces on the castings wouldn't clean up. Sometimes you might restart 3 or 4 times. When you were happy with the finished item you centre punched the lines about every 2". Bores you marked out from the table and your centre lines. Normally you would knock a piece of hard wood into the bore and scribe the circles with dividers or trammels from the centre marks on the wood.

You were expected to put every cut line on the casting not just centre lines so marking out a big casting could take a day or two, especially if the casting needed turning over a few times with the overhead crane.

We used very large height gauges, scribing blocks, conventional squares and sheet metal squares plus a 1/2 lb hammer and centre punch, rules, tape measures etc.

I remember one time we were really busy in the shop so some machining was subcontracted to a really large company who actually had a marking out machine. These were steel castings weighing about 4 tons ( 8,000 lbs ). When these castings came back finished they were a total cock up. About 1.25" had been taken off the front face but parts of the back face spigots weren't even machined ! Oil pockets to bearing housings and oil seal recesses didn't line up etc.
They were a real disaster. The two castings had to spend a few days in the welding shop being doctored. They eventually passed inspection but they didn't look great. You really had to know what you were doing to get things right and marking out machine or no marking out machine the guys in the big shop obviously hadn't spent enough time marking out the job in the first place.

I enjoyed my time on marking out. It was very interesting work without being hard graft and you weren't getting really dirty either.

As an aside work that was marked out by one person had to be inspected by another person to avoid the risk of guys making the same mistakes in calculation.

Regards Tyrone.

never make layout lines on casting cause not needed with cnc. if probing not used to center or stock divide part than manually using a spot drill on the machine as a pointer at the far edges or sides and using the DRO on the cnc to show if part is centered or stock divided to expected position
.
actually thats the easy part. a big casting can have 100's of features and often bigger castings the pattern and or mold is assembled from pieces or sections. i have seen accumulated error cause some spots of casting to stick out a extra 1/2"
.
you can machine 99 features no problems then obviously a 6"facemill taking over 0.5" depth at over 50 ipm feed going to choke on it. cannot always easy to see the sides and back of big castings. often its best to stop walk around and see if depth of cut looks in the ballpark correct amounts. and learned the hard way just cause 1st casting was ok doesnt mean the 2nd or 3rd casting is correct. i have seen where casting core floated or went higher in the molten cast iron. then the open space or core was shifted. obviously big 4" or 6" carbide insert mill expecting nothing in the way makes a bad drill plunging at 500 ipm rapids into whats suppose to be no metal in the way cored space. not a good ideal to press the cycle start button and not watch the rough machining and to not be within arms reach of the control panel.
.
actually often its the 1% of castings that often cause 99% of the problems. complex parts not easy to see if something wrong. can you walk up to a car and see if all parts are in the correct locations ? you think with a car you can use a tape measure and check every single part of the car ?? complex parts cannot be measured with a tape measure most times

Lots of collective wisdom in the above posts.
I would add only that in preparation for op. 1, it should be realized that the smaller the casting the more important surface bumps/grains/protrusions of a given size can be in setting up "square". The broader the surface, the less influence, in many cases, a single nasty peak will have. So with smaller castings, or with castings with particularly pronounced surface irregularities, it can be helpful to start with some handwork: thermite mentioned use of an angle grinder, spaeth mentioned the use of a humble file.

-Marty-

Marty Feldman said:

Lots of collective wisdom in the above posts.
I would add only that in preparation for op. 1, it should be realized that the smaller the casting the more important surface bumps/grains/protrusions of a given size can be in setting up "square". The broader the surface, the less influence, in many cases, a single nasty peak will have. So with smaller castings, or with castings with particularly pronounced surface irregularities, it can be helpful to start with some handwork: thermite mentioned use of an angle grinder, spaeth mentioned the use of a humble file.

-Marty-

Click to expand...

.
.
using a crane to lift part out of fixture to hand grind a bump on casting is rarely done. sure i have done it before only cause bump was touching fixture where suppose to have clearance causing it to not go up against stops where it had too.
.
in general even a small casting only weighing 40 lbs a crane is normally used. not worth getting hurt lifting it and possibly dropping it on foot. also crane when rigged properly you can balance part for correct orientation to more easily put the bolts in to clamp to fixture
.
normally just shim part in fixture and adjust work offset 99% of the time. its rare to not be able to shim part