I'm looking for some guidance on how I can produce reliably strong one-off aluminum castings in a hobbyist foundry setup. At this point I'm fairly comfortable with the molding process and designing sprues/gates/runners to produce quality aluminum castings, but I've shied away from anything load bearing so far. I think in my case it would be better to overengineer things with generous safety factors than to invest in the expensive equipment used by professional foundries, but I need some baseline values to work off of. My biggest concern is the subject of degassing. I run a propane burner with adjustable forced air and I always do my best to keep a clean flame during a melt, but I happen to live in the exceptionally humid state of Florida. I'm not really interested in having a proper degassing lance made, so the question is how much strength is actually lost as a result of gas absorption relative to the humidity and melt time? Are there more cost effective means of mitigating this? I'm also curious about post processing. It seems like a T5 heat treatment could be done in a home oven, or if anyone has other recommendations. I would also love to hear what alloys might be most forgiving if I were to purchase virgin ingots. Lastly, I could use some ideas on a good testing rig for verifying the strength of my castings. I'm imagining pouring a few small uniform bars along with the actual part, and using some combination of a bottle jack and a load cell to get some values for the tensile strength. I figure this would also give me the option to design parts around scrap metal like car wheels, assuming I keep each particular alloy to itself. Thanks, Charlie
define overengineering: in the cast of castings, just making something larger/thicker DOES NOT make it stronger at least not in a linear manner as is mostly true of forged material as a matter of fact ESPECIALLY in the case of aluminum thicker material can get weaker if not very well designed and fed, In short thick parts get grain structure issues that make them weaker. go look at the underside of any machine tool table , look at all the design and ribbing, it's not just to save material and weight, it makes the table stronger for it's weight im in North Florida, Jacksonville so I feel you , I degass with sawdust without issues, http://forums.thehomefoundry.org/index.php?threads/degassing-question.2179/#post-45060 I have used pool shock in the past, with it you really need to buy a new bag for every pour, it absorbs moisture, and will add H2 to your metal rather then remove it if it is not completely dry using good clean metal Proper casting aluminum is probably more of an issue for most amateur aluminum casters then anything else, using forged scrap or heaven forbid soda cans trips up most newcomers, next is melt fast aluminum pics up hydrogen every second it is liquid so getting from melt to pour temp quickly is critical ... Really if your furnace kicks a$$ it will happen almost too fast. ( My furnace is scary I dont recommend it) , but if you are taking 15 minutes to get from molten to pouring, there is much more potential for Gas issues V/r HT1
A couple of thoughts: You need to be very careful if you're casting items which if they fail would cause personal injury or death. Not yours, but that of others. HT1 is on target as always about over engineering however if you are supporting a 100# load and you design for a 200# load and demonstrate it will hold a 400# load you are not grossly over designing. I don't degas very often and get a little porosity. It is unsightly when you try to polish a surface however fine scattered porosity in a structural member only nominally affects strength. There's a thread someplace where I etched some home cast sections and microphotographed them. I have some hot rod motor mounts from lost foam and they are holding up fine. Lost foam inherently generates porosity. You can cantilever a coupon horizontally from a clamp and slowly add weight until you get permanent deformation. You can then calculate the apparent tensile strength on the coupon knowing the total weight and cantilever distance. It won't be real accurate but you can get relative strengths of various coupons that way. As HT1 warned, you will only demonstrate the strength of the coupon as cast, for that particular pour and coupon thickness. You have to add weight then remove it to see if the coupon has taken a set. Then add more weight. Creep up on the weight needed for plastic deformation and you can do quite well. It's an easy calculation to get tensile strength from the load on a cantilevered uniform section.
By overengineering I mean taking a design that might be suitable for manufacture in a professional foundry and just adding an additional safety factor to account for inconsistencies related to the home shop setup. My research puts the fatigue strength of non-tempered A356 aluminum around 8000psi, so in theory a casting of this type should be able to hold 1000lb indefinitely if it has an area of say 1 inch x 1/8th inch in the load direction. The question is how much larger should my casting be to be considered "safe". That is just for example by the way, I certainly am not planning on suspending a 1000lb load from a homemade aluminum casting. I was not aware of the grain structure issues with larger castings, so I'm already glad about posting this thread. I'm curious what the maximum wall thickness should be for sand casting? I'm getting a range of answers online from 0.2 to 1 inch. In any case adding ribs and/or increasing contact area will be my main focus in design. I read through the thread on sawdust degassing, seems like a viable method and I'm looking forward to giving it a shot. I have heard about the pool shock method but I'm just not very keen on standing over a pot of molten metal that's spewing chlorine gas. After refining a mold I would have no problem purchasing some clean casting alloy for functional parts. The difference between even assorted scrap castings and regular extrusions is staggering.
This is a good point, luckily most of what I have in mind currently are things like brackets, bases, and other custom mounting solutions like your motor mounts. Nothing that will cause some catastrophe if it fails, but I'd like to do enough engineering to have confidence in my work. I'm not concerned about appearance necessarily, in fact a little evidence of the process can be a point of pride for me, I just want to confirm that some porosity isn't going to impede functionality. I like the cantilever idea for assessing tensile strength, especially that it will give me a value for yield strength, as well as breaking. I have a stock of casting alloy from a full set of wheels I picked up a while back, so I have plenty of material to refine the process. I've been wanting to get a baseline of this material as it should be plenty strong enough for low load applications. Of course once I have a good pattern for the coupon I can make 1 or 2 with each pour to check for consistency with the design. I'm really curious to see how close it will be to the published values.
Not a simple question to answer a lot of variables , I would discourage you from messing around with design too much, remember the engineer should know more then you ( i know now days with China, and extreme optimization , that might not be true.) true story in the Navy overhead lifting devices are safety tested to 200% , I have seen more then one chain fall fail its first test before use, and the manufacture refuse to replace it because they consider 110% to be an acceptable weight test, so it was very common to see chain hoists with their capacity downgraded, so a two ton hoist that you could only use to one ton V/r HT1
It's certainly a subject I've considered at length and have had the benefit of having friends that were professional foundrymen and ability to observe their practices. There certainly is no reason that industry best practice cannot be achieved by a hobbysist, but whether that is practical or not in that setting may be another matter. If you find degassing to be too cumbersome, you have a long way to go to emulate the pros. Knowing the strength of material is certanly valuable information assuming you can accurately predict the state of stress to begin with. Can you? You need to know the character and nature of loading and have a very accurate FEA model and analysis. Managing metal quality and H2 porosity in particular is certainly one of the prime issues, but the other, is one-off versus the development work typically done to put a part into production in commercial manufacturing enviroment. That will invlove iterative prototyping, destructive testing, and the abilty to reliably measure your results. You can make and test all the sample coupons you like, but that doesn't mean your castings will uniformly exhibit the same qualities. Often times, the design features of a casting will mean that it won’t freeze uniformly or in a directionally consistent manner (thick sections/intersections for example), and those areas are the ones most likely to exhibit H2 porosity and related shrink defects and experienced foundrymen go right to them when they section and analyze the casting. If those areas happen to coincide with a critically stressed location, it can be bad news or at minimum require some iterative development of the part and feed system to achieve satisfactory results. Then there are usually production sampling plans to insure that everything stays in control along the way. Most foundries will have the means of testing metal samples directly for H2 content and that would be done before any mechanical testing. There still is value to going to the effort of tensile testing or polishing and sectioning if nothing more as a means of verifying you can control the consistency of your mold media, furnace tune, and as far as using scrap metal, I'd say forget that unless you have a highly reliable source of information as to the alloy.......and don't assume all wheels are 356, because that is not always so. You might argue why do I have to know the metal composition if I know it's strength? If yield and tensile is all that is important, maybe, but then there’s fatigue, post heat treating, etc... If you have a fuel fired furnace, you will move a very large mass of combustion air through your furnace and there will be a large amount of water along for the ride, and if you live in a very humid enviroment, that will be a very large amount of water. You should do the calculation but you will be surprised at that mass of water. Commercial foundries will purge and blanket the melt in a holding furnace just to prevent H2 infiltration when there is no flow! I use a resistive electric furnace because it has no air flow. In fact, for reasons I'm not quite certain, it seems to have less than atmospheric levels of O2, because if I melt scrap and lift the lid, it will often flash when exposed to air. But I largely avoid exposing the melt to large amounts of water. If you have a fuel fired furnace, crucible hat and purge gas may help, but can also be cumbersome. I'm a lost foam caster and my mold media is dry sand. I dont have to control or worry about the moisture or other binder content, because it's not there. As far as lost foam being more prone to defects and porosity, I'd say that has not been my experienec, at least no more or less than conventional sand casting. The byproducts of decomposed foam are not soluable in Aluminum. In fact most things are not to any appreciable level. It's just H2 that is the bugger. However I will concede when the castings become thick, I do see more defects, but so do conventional sand castings. I have my coating permeability tuned to produce best results at 1/4" wall thickness. In the end, because of all the things that need to be controlled, most engineers avoid castings for critcally stressed parts and will opt for machined wrought/billet or forgings. If it's a one-off, it's almost always a CNC'd part from wrought unless you need the strength of a forging and ability to validate a design for production. It is generally excepted that the mechanical properties of castings are derated by process with die cast being the best, then shell/investment, and sand. Commercial aircraft production spars and ribs are CNC machined from large Aluminum billets. 95% of the billet becomes chips. One example I see from time to time is the triple tree clamp on a motorcycle. Why on Earth would you cast a small part tlike that when it will need finsihed maching anyway? Especially if you are making one? Best, Kelly
I used to repair CAT D8 bulldozers and one time I had a look at the torque converter castings which were aluminium castings. They were full of pinhole porosity and these castings have to take the full output of their diesel engine. They never failed but this was the only time there was pinhole porosity in CAT castings and CAT castings are usually the best you can get. Pinhole porosity in aluminium castings are unsightly but it does not effect overall strength. I have pressure tested my aluminium castings to 340 bar (5000psi) and that is when a thread stripped and it leaked. That casting had plenty of pinhole porosity and still got up to a very high pressure. Had that thread not failed it would have got to a even higher pressure.
I'd beg to differ on that point. Porosity as well other defects certainly do affect strength as well as other mechanical properties. It's really all a matter of the degree of poristity and degree that it affects the mechanical properties. The subject has been studied at adnausium by acedemia and industry, but in general, is why more generous safety factors are recommended for castings and as I mentioned above, there is a heirarhcy by casting process as to how conservative to be. Visually observable porosity is generally considered severe porosity but it exists at microscopic levels in about everything. Castings can have porosity and still be fit for purpose but it's just a matter of the degree to which they are (over) design to compensate. No foundry or manufacturer aspires to produce and use porus castings. Having operated a company that made hydraulic components for eight years, with CAT as it's biggest customer, I can tell you they were very intolerant of flaws of any kind and willing to reject parts more much lesser concerns. One of the mechanical properties of aluminum castings most affected by porosity is elongation......how much can metal plasticly deform before it breaks. Many aluminum foundries first analize the elongation even before strength because it strongly correlates with porosity and other defects and dictates a safer failure mode for a casting.....it bends (more with high elongation) before it breaks. For the original poster, there was a thread here within the last year where a member made an aperatus to measure porosity by comparing weights in displaced water to calculate density and infer porosity. I had my doubts about the ability to measure accurately enough but thought it was clever. I searched but could find it. Anyone remember that thread? Best, Kelly
I have some experience with design from college, but like Kelly mentioned, engineers prefer repeatable and reliable so we never touched on castings. 110% seems almost laughable as a safety factor, I was taught 150% at a bare minimum, and based on replies so far I'm thinking more in the realm of 300% for my aluminum projects.
Kelly, this is some really great info, thank you. I think I might want to clarify that what I'm trying to understand are the limitations of the material and method. In other words, at what point is cast aluminum no longer viable for manufacturing some idea I have come up with? I now realize just how blurred that line is, especially given the lack of control in a simple home foundry, but surely some rule of thumb can be arrived at that leaves generous room for error. That being exactly what I hope to approximate with some load testing. I figure if I'm getting far enough into the weeds that I'm setting up an FEA analysis its probably best to just start looking at stronger alloys or machining from wrought stock. I'm thinking more in the realm of knowing whether a decorative shelf bracket can hold 50lb or 150lb of stuff, things like that. Of course, this example only needs a dead simple failure test, but having a ballpark estimate to begin with would save a lot of pattern making, assuming I had a weight limit in mind from the start. Most of what I'm casting currently is in a variety of bronzes, and my propane furnace/greensand setup is more than sufficient for the task. Regrettably this does put me at a disadvantage for Al, but there is just too long a list of things I want for my shop to consider upgrading at the moment. 1/4" wall thickness should be suitable for a lot of the designs I have in mind. I'm going to get to work making and testing as much as I can, I think I'll post a new thread once I have some data to share. I'm also keen to try the water test you mentioned for porosity, I remember seeing something about that back when I was first getting started.
Makes me wonder how those castings got through their inspection department and ended up in one of their bulldozers. As I said before the torque converter castings never failed. On one of my many visits to olfoundrymans home he told me that he never degasses his aluminium when he pours gravity die castings. The reason he gave was the metal cooled so quickly the porosity did not have time to come out so you could not see it without a microscope. He also said that some porosity helps to eliminate shrinkage cavities which can make a pressure casting leak. I am sure that the casting I poured and tested would have leaked if there was no pinhole porosity.
I cant really comment because I have no idea what parts you are considering but common sense says to avoid parts that have a high consequence of failure and/or overdesign the piss out of them, and that's probably snesible regardless of how they are made. I've made quite a few critically stressed parts for myself, (not necessarily castings) mostly racing parts butif your racing, there's already a certain element of risk. Occsassionally I make them for a few friends knowledable enough to know what they were doing and assume the risk, but in general, I try to avoid casting critically stressed parts. But that doesn't mean I don't wan't good material properties. I wan't them to machine well with good surface finishes, be hermetic, and of course be as strong as possible. The good news is, with good design practice, a 1/4" wall aluminum part can be pretty stout and it's also a practical minimum wall thickness for sand casting parts. Look to similar cast parts by way of example for design. If you have the knowledge and ability to calculate and model stresses....by all means do so. As opposed to testing coupons, which can have its own challenges, I think I'd be more inclined to make and devise a way to test the actual part to failure. Besides that, work on optimising your casting processes, and be safe. Best, Kelly
Here is what the FAA has to say about castings: 14 CFR § 25.621 - Casting factors. These are in addition to the normal FOS of 1.5 for structural parts made from materials with statistically well defined material properties, such as 4130N and 2024T3. There has to be a very convincing argument from the engineer in charge to specify a casting where weight is such a significant factor - as it is for aircraft. For other devices cost may be the driving factor. In either case a well controlled process is required to achieve consistent results. The better the process control the less deviation from part to part and the lower your FOS can be. For safety critical parts FOS is only part of the equation. Fatigue life of an aluminum part is also important. Aluminum doesn't have a 'knee' in the SN curve like steel.
I don't know what "breaking" means. Surface cracks appear after failure on a microscopic level. As a mechanical engineer I would recommend you concentrate on yield for your designs. Ultimate strength is quite difficult to determine. Yield strength (when the material experiences permanent deformation) is easy to determine. When something stays bent after loading people generally consider it to have been failed, even if it doesn't break in two. The cantilever loading will help you analyze yield strengths.
By breaking I mean complete observable failure, ie a loose estimation of ultimate strength and ductility. I am indeed only concerned with yield as a design constraint, but as a matter of experimentation it will be interesting to see how the castings behave beyond yielding, and how close it will be to expected values.
