Automotive Rear Upright/Wheel Carrier

Discussion in 'Lost foam casting' started by Al2O3, Nov 25, 2018.

  1. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    201 does contain silver and 2xx alloys are definitely stronger. They also much more difficult to cast and heat treat to the condition where they exhibit those properties. I actually gave more consideration to 357. A356 T6 by comparison would be ~40kpsi ultimate and ~30kpsi yield with 5% elongation. These values are optimistic for (especially my) sand castings and probably better aligned with die and permanent mold. With my processing I'd probably stand a much better chance of achieving a higher percentage of published mechanical properties with 356 & 357 and JoeC was kind enough to provide me with grain refiners and modifiers for use with the silicon containing 3xx series alloys. There is the other practical matter; 356 is just a good all around casting alloy so I bought 500lbs. I wouldn't want to buy 500lbs of the others to cast two 6-8 lb castings. If/after I have success with 356 on this part, if I could find a low volume and verifiable source, I might have a go with a higher strength alloy. Which corvette years do you believe were 2xx? Problem is, those suspensions are such popular take-offs for kit cars and street rods, there's no reasonable cost scrap unless from a broken wreck.

    Ignore the "requirement" column. I just grabbed the chart from a web discussion on alloy selection.

    Alloy Tables 201 206 357.jpg

    Well OIF, like I said in post #1

    ....nothing ventured, nothing gained. Let's see how round two goes.

    Best,
    Kelly
     
  2. JoeC

    JoeC Copper

    Beware published tables of properties, I went down a rabbit hole with Phd folks from Alcan as to where the data cam from and what it meant

    Most all tables were from Alcoas work during/after WWI to promote Aluminum

    Their lawyers made them disavow the data for fear of lawsuits (of course)

    I wanted to know if they were Min, Max, Avg, or what exactly, since they never have a std dev or +/- quoted

    Turns out they were "best" the lab techs knew the marketing guys wanted good results, so they managed to find a way to exclude poor results

    There was a DOE study called USCAR that tested actual parts from actual foundries, and the range was from almost nil to table published, between foundries, between parts and even within the part itself, I may have a link if people want to get into the weeds

    In particular Yield and Elongation vary big time due to defects such as voids, oxides

    I have seen a very few rare 356 wheel test bars go 18% (edited to remove 28% typo) elongation and had one 0% when it fell apart during machining for a void in the gage length, average was only 3.3%

    So what? The zero % test bar was a void and a fail, but it was still a good wheel in my opinion. After all a pipe is very strong even though it is mostly void, the forces flow around the skin which is solid for the most part in a casting

    But note that this is tensile properties, not fatigue, casting defects have a profound impact on fatigue

    This has led to aerospace down rating casting properties making it almost impossible to sell castings into airplanes (only forgings)

    I would not be concerned about strength, but definitely worry about fatigue

    Maybe do some dye penetrant testing every so often to see if any cracks are opening up as cycles increase?

    I would happily drive the vehicle with your casting with all that said
     
    Last edited: Dec 26, 2018
  3. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    I did 30 yrs in aerospace industry. We used many castings in the early part of that period but by the end of my time they were mostly relegated to non-structural parts like enclosures, and mostly shell investment. Over that time the rapid increase in machining speeds along with the casting tooling cost, lead times, casting variation/scrap rate, and relatively low production volumes meant from a cost and mechanical properties perspective we'd rather turn a large block of wrought into 95% chips than use net shape processes. For very highly stress parts forgings still persisted but even those were being displaced by machining high strength to weight materials that didn't lend themselves to net shape process. The bulkheads on Joint Strike Fighter are machined from wrought Titanium and a good look at the size and machine cycle times of those would numb your mind.

    Yet....here I am....lost foam casting parts.

    Totally agree. -Castings in general and aluminum especially.

    Thanks Joe. The jury is still out. Not sure I can say that yet in respect to this part. Though not recently, I can say I have driven cars and motorcycles near limit with unproven parts.....still here, but don't recommend it and having made it this far, I'm better acquainted with my own mortality. I would like to complete the journey.

    Best,
    Kelly
     
  4. cojo98v6

    cojo98v6 Copper

    For the first attempt it turned out good in my opinion, I'd say it was a successful failure. You got a casting that answered several questions and at the same time created some new ones (part of the challenging fun). As others have stated its kind of strange how the cores floated, I would have thought they would have just broke apart. Maybe to keep them in place just take some aluminum TIG rod and kind of anchor it to the inside or just stab it through the foam, if that make sense. Use a 1/16" rod at somewhat of an X pattern, the sand surely would flow around it and then it would possibly hold that sand in place.

    I pulled out some old books from my engineering classes, one was kind of a generic metallurgy book, but could not find any real valuable information. I looked into my statics and strength of materials and found a little info, see picture. Not sure if any of this helps, but figured it would not hurt to post.

    IMG_1410.JPG
    IMG_1414.JPG

    You need to post a picture of your Pantera, they are sweet little cars. Last year my brother-in-law and me did the Hot Rod Power Tour and they had the Ring Brothers Pantera there on display, it was very nice. My goal some day would be to own a Ford GT, the first gen 2005, I prefer them as they look like the original gt40. Hopefully by the time I can afford one they wont be ridiculous collector price. The new carbon fiber one is too super car looking for me, plus the v6 in it kind of disappoints me and way out of my price range. In my perfect world every car would have a V8 and manual trans, lol.
     
    Last edited: Dec 11, 2018
  5. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    Learning is what I'm doing when I'm not succeeding:)

    I'm not sure that would work in this instance but it's a very interesting suggestion. Foundry lingo for that would be a wire chaplet. Basically re-rod for cores and core support. In bound cores it works. The bond in lost foam is just mechanical interference and friction between the grains of mold media. Grain shape has a lot to do with it but the bond is very weak between grains and I suspect it would be even weaker with a smooth wire....still an interesting thought though.

    Think I'm with JoeC on this regarding published data, especially when it comes to sand casting. Probably most useful to make relative comparisons between materials as opposed to banking on achieving some ultimate strength. I have no model nor FEA from which to predict the state of stress and even so, the boundry conditions and loading would be a guess.

    That's a bucket list thing for me.

    I'm with ya. Modern supercars are a thing to behold but my cars are manual trans and need to be driven; they don't drive themselves.

    Best,
    Kelly
     
  6. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    I spent some time in the shop today and got a good start on the pattern for Round 2.

    The pattern is similar to the first go except I made some design changes, assembly refinements, and made some additional pin router and assembly fixtures.

    20 Fixtures.jpg

    Most of the changes are subtle. I used the new machining fixtures to remove unneeded weight from less critical areas and since I was under my target weight, added to the thickness of the main structural webs. These are now closer to 3/8” whereas previously they were ~¼”.

    21 Web Thickness.JPG

    I refined the brake caliper mounts and where the axle bearing bosses interface with the body.

    22 Detailed 4.jpg 22.1 Detailed.JPG

    I added the circular core windows to connect the respective cored areas with interconnecting feeders as shown below.

    23 Core Connections.jpg

    I still need to add the core windows that connect the main cored areas to the rest of the mold. Below is my current thinking on where to cut those in.

    24 Top Windows.jpg

    Hopefully these mods will stabilize the cores. With these provisions, below his how I envision the mold sand flowing and packing. The red lines are the internal/external core windows.

    25 Sand Flow.jpg

    I’ll ponder the placement of core windows, complete the pattern, and prep it for casting.

    Best,
    Kelly
     
    oldironfarmer likes this.
  7. Al Puddle

    Al Puddle Silver

    But an action shot....!
    Just got my pictures back from Fotomart, I forgot I took this picture. Looks to me you hit the cup just fine and things started well.
    KellyPouring.jpg
     
  8. PatJ

    PatJ Silver

    I think the openings are a convenient way to support the cores; I guess this is obvious.

    Like the "freeze plugs" on auto blocks.
    They have nothing to do with water or freezing.
    Their sole purpose is to support the cores.
     
  9. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    Finished up the pattern tonight. Cut in the core windows, added the sprue & gate, and finished detailing it. I’m giving some serious consideration to dip coating it but need to think on that one a little more.

    26 Pattern.JPG
    27 Pattern.JPG

    Best,
    Kelly
     
    Mark's castings likes this.
  10. PatJ

    PatJ Silver

    I think this rendition is going to work.
    It has a good look/feel to it.

    Edit:

    Does this mean you are going to have to section another one to verify it?

    .
     
  11. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    I'm optimistic....we shall see!

    Most likely. Need to see what I kind of result I get. If it looks reasonable, I may do a little load and deflection measurement.....then section it. I can actually see most of the interior features through the core windows. -Hope I have the opportunity to feel conflicted about cutting it up because that would mean it turned out well.

    Best,
    Kelly
     
  12. PatJ

    PatJ Silver

    You could always test it to failure, and then section it.
    It would tell you the strength of a good part, and also tell you the strength of a bad part if the part were not cast exactly right, but either way you would get an idea of what sort of loads the part could handle.

    .
     
  13. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    Someone needs to teach that guy not to stand down wind when he pours lost foam.

    Best,
    Kelly
     
  14. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    Got the pattern ready to pour today and attended to some details.

    Internal Features and Stress Risers


    The external core windows do enable me to see how well all of the internal pattern joints are fitted. I’m very happy with the fit achieved on Pattern #2 largely due to the use of some additional pin router fixtures. There are some places on the cored interior of the pattern where pattern surfaces interface with sharp corners and transitions. These are stress risers and if there is even the slightest gap in a joint, those become big stress risers because the pattern coating and sand will fill those tiny gaps creating fishers in the casting……which would be a very bad condition for a part like this. With the external core windows now in place, I was able to drop in some wax fillet and access all but one of these areas with the fillet tool to apply wax fillet. Here’s shot of some of the internal fillet.

    29 Internal Fillet.JPG

    Addition of Bobs To The Pattern

    Some will disagree but I don’t think a bob can't hurt as long as they are located in the last portion of the mold to fill. If they are in the initial phases of fill, they do consume hot metal which may be undesirable.

    I say may, because when I have a fold flaw in a lost foam casting, they almost always present themselves in what appears to be the last part of the casting to fill with metal. I think this is the residual affect of the initial burning of foam and turbulence in the pouring cup at the onset of the pour or perhaps where two molten streams converge during the fill. Either way the fold seems settle into the last part of the casting to fill. I figure this gives the bob the opportunity to collect that suspect metal instead of the casting. The other benefit is the bob provides some gage of sufficiency in the heat in the pour. If I can fill the additional bobs at the extremities of the molten metal travel, I can potentially reduce the pour temp. If the bob doesn’t fill, I know I’m on the edge of the required pour temperature. In conventional sand casting bobs are usually introduced to prevent or address shrinks and/or tears. As long as you keep the gate and contact area and bob cross section small, it won’t suck metal back away from your part.

    I added three bobs at the extremities of the pattern.

    30 Bobs.JPG

    Pattern Coating

    I was thinking about dip coating the pattern. OIF opined in an earlier post about whether coatings provide some shell strength. They may to a small extent but I try to keep the drywall mud coating very thin in hopes it will yield some permeability. I do think to a greater degree coating may prevent localized sand erosion, and in conjunction with better surface finishes it prevents metal penetration into the sand (and vice-versa). I dislike the sand imbedded in the metal because it’s very hard on cutting tools and wouldn’t want sand granules shaken loose in the part when it was in use. I had planned to media blast and then use vibratory media to clean the casting.

    A dippable slurry will have to be much thinner which means much they are much wetter. This will make the internal features very hard to get dry. I kept thinning some drywall mud and dipping a sample part until I was getting the desired coating thickness.

    31 Dip Sample.JPG

    Since this is a large part, I decided I didn’t want to mix up a boat load of thinned drywall so I made some foam plugs and slush coated the interior of the core surfaces.

    31.1 Plugs.JPG

    I pulled the plugs, drained the cavities, and then brush coated the rest of the pattern.

    32 Coated Pattern.JPG

    I set the pattern over a heating vent in my home with a cardboard box over it. We’ll see what it looks like tomorrow.

    Unbound Core Buoyancy

    Thinking about my previous casting attempt at this upright and the previous unsupported core experiments, I was wondering what kind of buoyancy forces were being exerted on the unbound sand cores during the pour.

    Initially, when the sand is packed, the pattern experiences the weight of the sand. When the sand is vibrated it becomes partially fluidized. The pressure this exerts on the pattern increases, behaving similarly to a fluid, distributing pressure more evenly across all surfaces (inside and out), and as a function of the depth and density of the sand. When vibration is removed the sand is able to retain the state of pressure on the pattern due to friction and mechanical interference between the shape of the grains of the mold media…..aka sand.

    When poured into the mold, the molten metal evaporates and displaces the foam. The molten metal pressure at the bottom of the casting is also hydrostatic (I think some foundry guys call it metalastatic pressure). In lost foam casting with unbound sand, the metal must exert a higher pressure on the mold surface than the mold surface is exerting on the pattern or else portions or all of the mold would collapse. So I was wondering how these forces compare and potentially affect the buoyancy and stability of the an unbonded sand core such as in this project.

    33 Hydrostatic Pressure.jpg

    The density of aluminum is .1 lb/in3. Molten aluminum is probably a little less dense but close enough for this discussion. That means every 10 inches of molten metal depth creates about 1 psig of hydrostatic pressure. So with the distance from A-D in my rig being ~20”, the pressure created from the top of the pouring cup to the bottom of the casting is approximately 2 psi. This head pressure can increase the rate at which the metal advances through the foam. This can be further increased by additional sprue height or introducing differential pressure by creating a vacuum in the flask, which also can further immobilize and stabilize the unbound mold media.

    If the mold surface on the external surfaces of the casting are stable, the buoyancy force on the core is the difference in metal pressure above and below the core (E) times the projected area of the core as you look down from the top of the flask. The projected area of the core as shown below is 31 square inches.

    34 Projected Area.jpg

    The buoyancy force acting on the core is the differential pressure times the projected area of the core, minus the weight of the sand suspended in the core. The difference in height of the molten metal above and below the core (E) on this pattern varies from 2.5” to 4” so the lifting force may not be evenly distributed along the length of core as the core depth varies. However, the projected area of the core also decreases in the areas where the differential molten metal pressure increases, so I’m going to approximate the average differential pressure across the core as the average core height which is 3.25”. If Aluminum exhibits 1 psi/10” of column height, that means the average differential pressure across the core is .325 psig. Multiplying that times the total projected area of the core, which is 31 square inches, gives 10 lbs lifting force less the weight of the sand core, which for the larger core is about 4lbs. So the approximate buoyancy force acting on the largest cored section is about 6 lbs force.…..wow…..seems like a lot.

    Another way to calculate the buoyancy force would be difference between the weight of the core and the weight of the aluminum displaced by the core. The core is about 100 in3. I figure the density of packed sand is about .04lbs/in3 while aluminum is .1lb/in3. So the core would weigh about 10lbs if it was solid aluminum but only 4lbs as sand so 6lbs net lifting/buoyancy force.

    35 Core Windows.jpg

    So for the cores in this automotive upright, the contact areas in the core windows need to withstand those buoyancy forces and also to keep the core itself from sheering, breaking, and floating. Unbound sand is undoubtedly strongest in compression. So the core windows on top of the pattern (1, 2, & 3) will probably need to do most of the work to prevent core float, since they are loaded mostly in compression and slightly in shear as the core tends to want to buoyantly move upward and inward. The exterior core windows on the bottom (4, 5, & 6) probably do very little to stabilize the core because they are only in slight shear and tension and I suspect unbound sand has zero strength in tension, and some strength in shear. However, these core windows do help with sand flow and dense packing in the cored areas.

    The small internally positioned core windows connecting the cored sections (7, 8, & 9) are loaded in mostly in shear and probably contribute some small benefit to resisting the vertical buoyancy forces/float but more so stabilizing the cores laterally, while still promoting flow, filling, and packing to and from cores 10 & 11. Will it be enough? So, we’ll see if collectively the external and internal core windows can resist 10lbs of buoyant lifting force on the core.

    If it were possible to think the problem into submission, I’d have a good part by now. Hopefully I'll get some good casting weather over the holidays. You all know what I want for Christmas….

    Best,
    Kelly
     
    Last edited: Dec 22, 2018
  15. DavidF

    DavidF Administrator Staff Member Banner Member

    Instead of mixing water with the drywall mud, I wonder if it would be possible to use alcohol instead???
     
  16. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    Might have to try that on a small scale. With circulating air and heat, I think I'll get the interior to dry in air.

    I did notice that anywhere there was wax fillet on the interior would not wet by the (very) thin drywall mud. The brushable consistency has enough viscosity and surface tension that it will.

    Best,
    Kelly
     
  17. JoeC

    JoeC Copper

    Core Float

    It is hard to get my head around core float with un bonded sand.

    Seems inconceivable (I'm using that word wrong - best movie ever - Princess Bride) that loose sand acts as a piece, but clearly the sand acts like a sold piece and floats as shown in your sections.

    These windows should help acting like chaplets ?

    download.jpg

    It will be very interesting to see results.

    This is a very elegant and refined design from the start point, should behave nicely as loads are applied.

    We may be getting into lost foam / full mold at my day job in the New Year

    Wish I could share some of what we are doing, but I am under a serious non disclosure agreement due to past thefts of intellectual property by nations best left
    unnamed

    Happy cast in in 2019!
     
  18. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    Got a lot done today. Sunshine and nearly 40F.

    Drying out the pattern was no problem. It was dry after two hours sitting over a household heating vent capped with a vented cardboard box. I’ll have to start a separate post on dip coating with mud…..have some good things going on with that.

    So I molded up pattern #2 in my new rig. I partially filled and vibrated until the lower core openings were covered. I then added fresh sand to the central core and it fed the others very nicely. I also tilted the rig and vibed in different positions, then proceeded to fill and vibrate the remainder of the mold.

    36 No 2 Molded.jpg

    So the A20 was charged, and final prep for the pour performed.

    37 Stage Set.JPG

    Demolded and by golly, had a fully formed part for my trouble, bobs and all.

    38 Dumped Flask.JPG

    So I de-gated, removed the bobs, and did some wire brushing to clean it up and more closely inspect the result. It’s not perfect but I’m very encouraged. The internal cores remained stable and fully formed.

    39.JPG 40.JPG 41.JPG 42.JPG

    I felt and visually inspected the internal core surfaces. There was a slight leaker under the area circled in red below. This was the only area of the cores that was not perfectly formed. I believe this can be easily corrected by changing the height of the small internal “feeder” core window to the roof of this core. I don’t like the surface condition in this area. It looks like some fold flaws. Not sure how much it may have been affected by the leaker. Air voids in lost foam molds seem to produce lots of trouble.

    44 Leaker Area.JPG

    44 Bore Flaw.JPG

    …..and still under target weight. It will lose ~1/2lb and ~1lb when machined.

    45.JPG

    It seemed to me the sprue and gate on the previous part worked ok so that was unchanged. So what has changed from round one?

    1. Some of the pattern features, most notably thicker main webs and core windows. The brake caliper wings are more refined (though I may remove some of the mass in the bosses), and features are much more accurately located.
    2. I used fresh sand for the cores. I’ve been using my LF sand for two years.
    3. I had a full 80+ psi of compressed air available for vibration where previously I had an anemic 40psi. Wow what a difference.
    4. I tilted the flask in different orientations while vibrating in hopes of promoting better packing cores and especially the lower cylinder of the casting where I had a leaker. Probably helps all the cored areas.
    5. I slush coated the cored cavities. Not sure this had much affect.
    6. I did decide to add three bobs at the extremities of the casting. I did not notice any flaws present in the bobs but they did fully form.
    7. Since it’s still a trial part I didn’t bother to flux, degas, nor add Sr or TiB. It’s all re-melted parts and sprue stock. I poured at 1575F.
    8. I had a bucket of water handy to quench the casting and blow the coating off.
    I’m happy with the progress. More to come.

    Best,
    Kelly
     
    Mark's castings and Mach like this.
  19. PatJ

    PatJ Silver

    Greatly improved from the last one.
    That is good progress.
     
  20. I learn a lot watching you work, thank you.

    Of course I have a few novice questions and comments.

    Damned fancy gate.:cool:

    Mighty fine vibration rig too. I may have to break down and make a tilt bucket. I just did a casting which had recesses on both sides and a couple of tilts during compacting would have been the ticket. As it was I just lived with a bit of bleed through.

    When I first take a casting out of the sand the mud seems rigid before spraying with water. Although the dry mud can't be very strong, it does have some strength and it appears to me to gain strength with heat. I mix my mud thin and apply per your advice to let the color show through. I think it is very porous as I get lots of foam vapor in my sand. My mud appears to be thinner than your brushed on mud. I do use Dawn in mine as a surfactant (per your suggestion) and I get good coverage on the wax.

    Did you put surfactant in your dip solution? I wonder if more soap would get better coverage on the wax. I guess if you can get a tool in to work wax you can use bent brushes to brush the interior. I still want to be able to dip. And the mud is so forgiving you'd be able to keep a pot of liquid a long time.

    Did you intentionally leave the long seams unfilled because it was a test? Or am I not seeing it right.

    Surface finish on the top was a shame. Are you considering reducing the bore to machine out surface defects?

    Good math on the core float. Archimedes would have used your second method.:rolleyes:

    Won't alcohol damage the foam?
     

Share This Page