What's new
What's new

Can software calculate casting shrinkage?

kaiwinthrop

Plastic
Joined
Apr 26, 2019
I am making a set of molds for pressing ceramic water filters. The molds will be sand cast in 5052 aluminum and then machined to spec. They are roughly 16" in diameter.

Can anyone here use software to take an .STL file and expand it to account for cold shrinkage so I can have a wooden pattern CNC machined from the file?

the file is here: https://drive.google.com/open?id=1L5_wj5uw66nf6KYgo09DgCEDAxwSBPQy

I work with Potters Without Borders, a nonprofit technical assistance charity, these molds will be distributed all over the world to help purify drinking water in at-risk areas.

Am I going about this in the right way?

thanks.
Kai Morrill

Capture.JPG
 
In my opinion casting shrinkage calculations are a black art best left to old pattern makers that have many years of experience making
mistakes. You see. it isn't linear.it depends greatly on cross sectional thickness and that usually varies throughout the piece. So it ends up being a WAG.
 
I have zero mold experience. Would all surfaces shrink/expand the same amount? If so can you just scale the part by a percentage?
 
Well, the question is, can I scale it by 2% and call it good or will I be wasting a lot of effort and material.

k.
 
Why do you need a wood pattern?

Machine it out of aluminum...

Or wax would be easier if you are doing lost wax casting.

It just seems pointless - if you want precision to make a mold then don't rely on wood.
 
why wood

Why do you need a wood pattern?

Machine it out of aluminum...

Or wax would be easier if you are doing lost wax casting.

It just seems pointless - if you want precision to make a mold then don't rely on wood.


It is the foundry that recommended wood, Im not doing the casting myself. Ive had the molds cut from billets before in the UK but here in Halifax its not feasible. anyway, a 16" diameter billet wastes a lot of material.

The blank which we are casting can be machined to adapt to a variation in receptacle diameters. I have already designed a specific amount of waste material into the pattern, the final step is to correct for the shrinkage.

several mold sets will be made from these blanks. The foundry recommended wood because it will be re-useable. Their process is to pack green sand around the wooden pattern and then remove it. We will be producing three or four mold sets per year. The wooden pattern will last for a few years.

I would like to get within a millimeter of accuracy in the casting.

kai.
 
No old pattern makers on here?
I guess if I was an old pattern maker I wouldnt be trooling the forums looking for work. we will see.
 
I see, that makes sense then.

What about an alternative wood-like material like MDF? It would maintain dimensional accuracy better I would suspect.

Also you could just ask them what they recommend material wise?
 
No offense to the sidetrackers, but please ignore the people diverting the thread into a discussion of whether wood is a suitable material for making patterns.

Unless you have the proverbial grizzled patternmaker available, a simple proportional enlargement of the pattern (as in post #4 above) is a fine and satisfactory procedure. That's why shrink rules (scales, not laws) were a useful and widely used tool. It would be worthwhile checking with the foundry on the expected shrink ratio for the alloy in question.

Do not neglect to add machining allowances to the pattern. One rule of thumb is to allow 1/8" on the rough casting everywhere that's to be a finished surface.

Somewhat related: In the image attached to the OP, there's a fairly large change of section between the rim and the wall of the upper pot. You should add a substantial fillet so that thickness transition is less abrupt. You can always machine off the fillet, if it would interfere with holding/mounting the pot.
 
Last edited:
Well, the question is, can I scale it by 2% and call it good or will I be wasting a lot of effort and material.

k.

5052 is not commonly cast (if at all?). 300 series are.

You can add shrink to the whole model, but there are a ton of gotchas. If you have any critical location features, you need to plan to adjust them after the first couple parts are cast and measured.

I'm not sure what the shrink percentage for sand casting is, but your foundry will work with you on that.
 
IIRC, the number for aluminum is more like 5% or 6%. All the info you require should be in Machinery Handbook.

PM
 
The answer to the OPs question is yes, you can apply a shrinkage to a CAD model to acheive a reasonable size for a wooden pattern. You will have a machining allowance as well as the shrinkage on features that need to be machined. Make life easier for yourself and don't make that too tight. You will also need to apply an appropriate draft angle in the right places so that the pattern can be drawn from the mould. CAD can do this as well. Cores may need to be worked out as well but will often be simpler. This isn't quite as simple or automatic as I suggest - there is certainly lots of black art in pattern making and casting but you can get the job done quickly in this way.

Just to show that this is not just theory you can take a look at the result of a fairly recent such job on this site:

Bolton Steam Museum - Home | Facebook.

Scroll down to the entry for 8 November 2018 and you will see the casting (68kg of cast iron) and some of its machining. I did the original CAD model for this fairly complex part, then with advice from our resident pattern maker planned the pattern and made the shrinkage and draft modifications. He then made the pattern from the full sized paper sectional drawings that I produced. Normally he would just make a pattern directly, but the gear profiles and especially the helical ramps made it easier to do this as a joint effort.

The whole job turned out well and is in use as a replacement barring gear for our largest beam engine. The hardest part was finding a local foundry willing to take on a one-off like this. As it turned out the foundry did a really superb job and I can't say enough good things about their service.
 
Last edited:
Ask the foundry where you're going to have them cast. Pattern makers are given the shrink factor by the foundry. Each foundry has it's own unique process that may effect the shrink. There are just too many variables that would effect the shrink unrelated to the material. Foundries have a set-up charge to engineer the process for a reason. I specialize in Wax Investment molds (lost wax process) not sand castings but I'm sure it's no different. From my 45 years experience in the casting business I can safely say that no two similar castings will shrink the same at different foundries.
 
Who's kidding who?

If AL shrinks 5/16 of an inch per foot, how is that 5 or 6 percent?



See post #2
I work for a factory that had foundry and
ordered molds for aluminum, and i was doing the mold drawings
I started using old existing drawings, to verify shrink.
but all castings came wrong all the time,because engineers went 7%
all over as std, but before that i was in inspection (CMM) and ispected
molds dimensional against mold dwg. for mold vendor acceptance.
but they sample before full production, after machining found out the you
can not use std. percentage, measured after casting and machining, and they always did
corrections, since i kept a book of mold behavior, when moved to engineering i applied
what i found in my past QA. position and saved the place lots of money because no more
mold rework to the cores and shells in their own tool room, i got a engineers pay,
they figured i was smart, but was not the case,it was a broken communication from production
and design, so they always applied std values, but since i moved from QA, and machining
dealt with discrepancies,
 
I work for a factory that had foundry and
ordered molds for aluminum, and i was doing the mold drawings
I started using old existing drawings, to verify shrink.
but all castings came wrong all the time,because engineers went 7%
all over as std, but before that i was in inspection (CMM) and ispected
molds dimensional against mold dwg. for mold vendor acceptance.
but they sample before full production, after machining found out the you
can not use std. percentage, measured after casting and machining, and they always did
corrections, since i kept a book of mold behavior, when moved to engineering i applied
what i found in my past QA. position and saved the place lots of money because no more
mold rework to the cores and shells in their own tool room, i got a engineers pay,
they figured i was smart, but was not the case,it was a broken communication from production
and design, so they always applied std values, but since i moved from QA, and machining
dealt with discrepancies,

So what did you find worked well as a scaling factor? Were there 'extra shrink' and 'lesser shrink' cross sections that you could apply a rough rule to?
 
Short answer: Yes, software can calculate shrinkage factor in cast parts. The real problem: Can you get your hands on that software?
 
So what did you find worked well as a scaling factor? Were there 'extra shrink' and 'lesser shrink' cross sections that you could apply a rough rule to?

it was that time no cad software was popular.
was expensive as hell, like 100k including computer
only huge companies could afford, all drafting table a cassio
calculator, I would apply in thicknesses as previous expierences
because,we were allowed only 0.050 inch, for machining stock, molds
were 17x10 round, and aluminum was expensive so lots of pounds can be saved
or wasted,i always sometimes add shrink or minus from dimensions, and got so
good to maybe 0.01 error,also gates can play a role how it shrinks,so I would
say foundry is an art,to make castings with limited leaks, they had to be air tested 100%, also mold paint, paint is insulator from heat, think can remember
like 1200 deg, melt was lot of fun working that kind enviroment, making 60K
parts a month that size, machining side was big, also new workholding tooling
had to be made per part#, I was for auto oe product.
 








 
Back
Top