Casting a Gingery Lathe

[Randy Rue]

You can buy bentonite online. I tried screening home depot sand using a screen sieve and a massager and after about two hours went and found aquarium filter sand ready to use.

I also eventually bought petrobond online. I use it when I want the finer resolution.

Look on yootoob for a channel called "Paul's Garage" if you haven't been there already. Over the years he's progressed from melting cans with charcoal to pretty sophisticated stuff and is slowly progressing on a Gingery lathe.

Stay away from "pure" aluminum like you get from extrusions and cans and whatnot. What you want is the blue-ish stuff you get from heat sinks and motor mounts, engine blocks, alloy car wheels. It has silicon in it. Lots stronger and less trouble with dissolved gases. Or even better, try one of the zinc alloys that melt at much lower temps. Paul's Garage has a good video on that.

[Nobody Special]

So, I haven't been locked away in a Turkish prison, just got busy with work. A quick update: I've been trying to learn and relearn things as fast as I can go. Closed green sand castings are more complicated than the farting around largely with open faced molds that I did when dinosaurs roamed the earth.

Reading foundry sources online, green sand casting books from antiquity, and the old standby - the 1958 Navy foundry manual. I bought 2.5 front end transmissions for about 20 bucks each and broke them apart. Frosty suggested a splitting maul, which I have to admit is a little rough on the maul but does wonderful things to the transmission. Gonna be about 60-70 lbs of aluminum plus some random rods, ball bearings, and bearing races to play with later.

Made some new, fine grade green sand which works about a hundred times better than the medium grade - may still save the courser stuff as backing sand for projects down the road. You can see the difference in the ingots below.

I've got about 16 lbs of clean metal now and intend to start the bed pattern this weekend. Reading machinist forums online - a lot of them say this design has simplicity going for it but that the lathe has issues with chatter, partially due to the thin bed tending to twist. They suggest several improvements for the bed, which is one of the first and most difficult castings. 1. Use wider and thicker ways, i.e. the steel rail that sits on top of the bed. This means making the bed wider. 2. Thicken the bed and use countersunk screws down the length of it rather than four large bolts through ellipsis shaped holes. 3. Use x-shaped braces instead of straight braces. 4. Don't feed directly onto the middle of the pattern to avoid turbulence. Instead of a "pop gate", use a regular sprue, runner, and feeders, bottom feeding the mold at the parting line.

I've gotta say, this makes sense, but I'm dumb enough to want the "full Gingery experience", so I may make one of each. This means almost double the metal needed for the mold though. I'm attaching notes showing the general difference in the amount of metal needed, and kind of roughing out the two designs.

Things learned or relearned the last two weeks: Don't use sand that's too wet. Or too dry, dummy. Fine sand good, not-fine bad. Runners and gates are easier if you use a pattern. That pattern should be round in cross-section. But not a full round or you can't get it back out without collapsing the mold. Oops. Also don't leave the patterns where your pit bull can get at them. Swab carefully. Double check your alignment pins so you know the flask closed fully. Remember to flux. Clean the crucible immediately after use. Top feeding large ingot patterns causes turbulence and blowholes from trapped air when the sides chill the metal, even when the top and bottom look great (see two sideways ingots). Don't forget the vents!

 

[Frosty]

I think the cats are wondering why you put non-food on the table they can't command your attention by sprawling on. 

Let the casting cool before breaking them out or they'll form al oxide like the dark ingots on the table. 

Frosty The Lucky

[Nobody Special]

So, I ain't dead yet. Just been busy playing Army and helping a friend move. Between yesterday and today, I got the pattern made for the hardest casting, the bed, and made a flask for it. I'm still figuring it out a little, but there are a few changes from Gingery's design, mostly the dimensions and the x-braces. It's a bit taller, a bit wider, and parts are a bit thicker. Also left room on the bottom for a wider foot under the headstock. Doing the fillets was much more of a pain than I anticipated, largely due to the bloody x-braces, but if the braces cut twisting/torsion, they'll be worth it. If everything works out, hope to cut a molding board and pour tomorrow evening (Changing that a bit too. Too much clean-up with a pop gate straight in the middle and no risers).

 

 

[Nobody Special]

Failure is always an option in experimentation, I just wish it wasn't so frequently the main option. Just tried to make the mold for the bed for the third time and failed the third time. Gingery had a pretty simple, but lousy system for feeding metal into the mold; he put a big honking sprue right on top of the pattern. That way tends to cause turbulence, poor feeding, and shrinkage that creates a lot of extra finish work.

In my infinite wisdom, I thought I'd save myself a lot of finish work and put a regular feed system in with a sprue, runner, two risers, and four gates. Looked fantastic, but because the flask was so narrow to keep the weight down, I put it too close to the pattern and it was making a weak spot in the mold, with the end result that part fell in by one of the risers...three times, dang it. Got it figured now, longer gates so everything is further from the pattern.

The other problem I'm a little more stuck on. Gingery has two or three braces going straight across the bottom. The pattern is put in the cope, which is filled, right side up, the whole thing is flipped over, the drag placed on it upside down, and the cores in the bottom of the pattern get made when the drag gets rammed. The last part of his process is flipping the whole thing over, lifting the cope off of the pattern which is left setting on the drag, then pulling the pattern off of the cores, which are left setting on top of the drag.

That's bloody brilliant, but I found out in a hurry that when you try it with X shaped braces, to cut down on the Gingery's tendency to have the bed want to twist, the cores get stuck in the bottom. Not a little stuck, like I have to dig the smaller ones out with a pocketknife. I'm thinking I may have to rebuild the pattern from scratch and use either do it Gingery's way with thicker braces, or with much longer "X"s to create bigger spaces. Open to suggestions though.

[Frosty]

I don't know if I've said it already but the second pic above sure looks like the bed form is straight sided. If so pulling it WILL be a serious PITA and depending on the sand maybe impossible without breakage. 

It is always safer to lift the pattern than the flask. If the pattern board flexes reinforce it with a couple hardwood boards, say 1" x 4" with little screw tabs you remove to ram up the mold but install to make the pattern board inflexible to remove and maybe even put convenient hand holds in them and skip the coffee.

If you use sodium silicate bonded sand it will HARDEN on contact with air in a while, day or two maybe, a lot depends on how much sand is exposed to the air. Orr if you don't want to wait put it in a relatively tight container and either give it a squirt of CO2 or toss in a piece of dry ice and close it up. 

You WILL need to wax the mold and board or sodium silicate (water glass) will stick in a permanent sort of way.

Frosty The Lucky.

[leos.mike]

Haven't made any of these steps, but did read the book about making the lathe.

One thought I had during that read was "I wonder why there is an insistence on casting the bed instead of fabricating it?"
Is there a reason my newb-brain doesn't understand yet for why casting with all this hassle and the custom flasks is easier than just making them out of steel c-channel (or something)?

[Scott NC]

  Have you read The Charcoal Foundry?  Hassel = new skills and satisfaction in the long run.  Just like everything else.  Have you hand scraped before?  Also nothing is wrote in stone, how about a hybrid?  

[leos.mike]

1 minute ago, Scott NC said:

  Have you read The Charcoal Foundry?  Hassel = new skills and satisfaction in the long run.  Just like everything else.  Have you hand scraped before?  Also nothing is wrote in stone, how about a hybrid?  

Well, that's pretty much what I was thinking, not "Why bother casting the lathe" because building skills like this is just awesome. It was the much more specific "is the one, hardest, casting step (the bed) even necessary or can you get a working lathe easier by not casting that one oversized piece"?

And yes, I read The Charcoal Foundry first.

[Scott NC]

Okay.

[Frosty]

And there's your answer, why indeed. A person can mow lawns and pick up dog poop and earn enough on weekends to buy a used lathe more quickly than learn to and cast the Gingery lathe bed. Have you fabbed something this complex requiring a lathe's degree pf precision? 

The question and answer is the same regarding blacksmithing, there are very few things you can' buy or order cheaper than pursuing the blacksmith's craft to make one yourself. There is only one positive answer. Why not?

Frosty The Lucky.

[Nobody Special]

Good evening, It's got draft, more so on the long sides than the short...think the short sides could have used a hair more.  For the falling in, I'm pretty sure I badly screwed up by getting the runner, and especially the risers too close to the pattern. I totally agree about removing the pattern from the mold, but I'm not sure how to go about it cleanly with the way Gingery has the cores sitting on the drag, and considering they're rammed concurrently, almost as an extension of it. Built that way, the tolerances are kinda tight for baking the cores and inserting them.

The x-braces are clearly too small and don't have enough draft, which I'm not sure how to solve easily. Possible solutions, a. leave them out altogether and just use thicker braces going straight across, sanded to give them a slight draft, b. use two, wider X-braces to make the cavities and thus the cores larger (and thus come out easier), or c. put no braces at all on the pattern so it makes one giant core, then attempt to cut "X"s into it without collapsing the core. A. is probably the simplest. Or I could just go ahead and do one the tried-and-true Gingery method, but heck, I can't even make mac and cheese anymore without playing with the recipe.

To speak to Leos's question, two reasons. Part of it is the challenge of doing things the bloody hard way and developing a skill, in this case casting. The reading I've been doing has been giving me a much better understanding of metallurgy and it is a fun, if sometimes exasperating challenge, not dissimilar to other craftwork like blacksmithing. I could crank out a functional wood lathe at various levels between a few minutes and a few hours, and I could work half a day and pick up a far better metal lathe than I can manufacture, used. The difficulty level goes down when I listen to good advice, which I don't always, but a dramatic failure is fun too, if not always at the time.

The other reason is that the reason I'm casting a lathe to begin with is that my wife asked me to teach her to cast metal, which is hard to do from fabricating (she already welds better than me, if her old neck injury lets her). If she can pick up half of what goes into this, it will be a very nice foundation for whatever else she wants to play with. 

[Frosty]

Yeah, I'm not fond of the X braces as you made them, too many acute corners. I  think I'd make a zig zag to make maybe 3 right triangles instead of Xs. Heck, maybe just two. All they are are stiffeners to prevent your ways from wracking and it doesn't take something complicated to do that. Maybe just a couple obtuse diagonals. That prevents the sharp inside corners to make it easier to pull the form. 

Or maybe just make the long sides of the box thicker. Aluminum doesn't like flexing anyway so making the box a little thicker will stiffen it up quite a bit. Just another 1/4" would make a significant difference. Being thicker will retain heat to help the aluminum stay liquid as it fills.

I'm no fan of your X bracing, it'd be hard to ram up and harder to pull the form. 

Hmmmm, to fill the Xs I just thought of maybe cutting the open edges of the long sides of the box into runners and sprue it at one end, riser at the other. You could just leave the long runners as stiffeners and reduce the X bracing to one or two diagonals. No draft to worry about, just cut them when you pull the form.

Frosty The Lucky.

 

[Nobody Special]

Finished the new pattern for the bed, rammed it up...and a quarter of it fell in. Did it again...same result, on the other side. Moved the risers, rammed it up...and it fell in again. (turns out the ribs that Gingery suggested for the cope (top part) make a weak point because they don't go all the way down). Then I got smart.

Gingery originally has you put the cope on a board, right side up with the pattern in it, ram up the mold over the outside, then flip it over, put the drag (bottom part) on, and ram the sand in it AND the cores for the inside of the pattern, still in the cope. Then you roll it over one last time, take the cope off, and pull the pattern off of the cores before putting it back together. It's brilliant, but it kept falling in, no doubt from me varying from his sizes and putting risers on it.

Then I got smart...I figured the whole reason for flipping it over so much was to create the cores...but when the top fell in, the cores were still inside the mold and perfectly good. So instead of starting from scratch, I rammed the top, which put less stress than ramming it over and flipping it twice...and part of the mold stuck to the pattern. (see attached pics). Dang it.

Today, I cut the side of the ribs short, and only had the top fall in once before getting it right. I think if I'd added some kind of support at the bottom under the ribs, it would have survived flipping, but I got it when I left the bottom alone and re-rammed the top like last night. Poured 1.5 hours ago...waiting to break it open and see if the aluminum gods have smiled on us.

 

[Frosty]

Cool beans. Give it more time to cool than you think it needs!

Frosty The Lucky.

[Nobody Special]

The aluminum gods did not smile, although they nearly cracked a sardonic grin. It looked gorgeous when I opened it up...and then not so much.

The first attempt got minor washes and runs in the front...almost good enough, but a little bit of misfeeding. The top has one small inclusion and a shallow concavity about the size of a silver dollar from shrinkage in one spot. The back is a mess of hot tears though from it shrinking, but only on the one side. The interior also had rough sections on one end, probably trapped gas.

On the plus side, I got a LOT of practice making molds and patterns and learned a good bit about handling it. I know how much metal it needs (just over a quart and a half), and I think I know what I screwed up...metal not hot enough and the risers needed to either sit behind or on top of the pattern. I was surprised that despite the cracks in the back the shrinkage was much less than I expected. The final piece was maybe a third of an inch shorter than the pattern, if that.

If I try again and it fails, I may also make cores with a molasses binder and bake them hard. It would make things easier and I wouldn't have to roll the mold.

[Frosty]

1/3" isn't bad shrinkage over that length. Do NOT make the al much hotter, a LITTLE bit should do it. If you switch to molasses or other hardening sand binder you can preheat the mold SLIGHTLY too. A couple few burning charcoal briquettes along the length under it can make a good difference. Think 100f+/- a few preheat.

Getting there.

Frosty The Lucky.

[Nobody Special]

Some updates - The baked cores for some reason expanded and wouldn't fit in the pattern. Who knew? I made a second attempt and screwed up the pour, which was maybe a bit too cold, and the sprue froze/clogged and I get a huge mess. Third attempt - my ramming has gotten much better, I used incandescence to judge the temperature (at night), I've taken to hand packing the cores into the pattern to avoid flipping it, and I added a touch of bentonite to the sand and tempered it a bit further, which helped tremendously...and then I shorted the crucible by maybe 20 ounces or so...bottom of number two, and the aftermath of number three in the pics below.

[Frosty]

You're progressing pretty well for not having hands on instruction. 

when you say you're judging the al temp by incandescence, incandescence of what? Certainly NOT the aluminum, it should barely show color in night time dark. Judge it's readiness to pour by how it feels when you stir it.

You DO stir it in the crucible, at least when mixing the flux, yes? Molten aluminum has a distinctive feel when it's the right temp to pour, getting it hotter doesn't help and often makes things much worse. 

Next time melt at least 50& more than you think you'll need. Weigh the most complete one you've cast and add at least 50% more to the melt. That is why you make ingot molds after all. BECAUSE there is always extra melt and you'd be crazy to let it freeze in the crucible. Yes?

Make ingot molds from angle iron in lengths about 1/2 the depth of your crucible. Weld a bunch touching sides between a couple pieces of strap stock. You don't need to weld the ends solid the al will solidify before much can leak through a tight contact fit. Once the ingots are cooled flip it over so it bangs on the ground and the ingots will fall right out. 

Muffin tins work but pressed steel ones don't last very long but cast iron ones don't take to being banged. Then again I've seen corn bread pan ingots in the shape of little cobs of corn. Soooo.

Frosty The Lucky.

 

[Nobody Special]

For the incandescence, I checked it with the lights off for a couple of seconds, poured at red, not incredibly dark, but not really sure yet if it wanted to go light red to orange. Say about 1400, and I know that's less than 1500...at least in steel, and I spent a long afternoon once hunting down articles on incandescence - what it looked like at the time is that most materials, and certainly metals glow once you get past about 1000 degrees, with about the same color at similar temperatures.

Of course, I've slept since then, and I may have read it wrong, and I certainly wouldn't want to trust that as a method in the daylight, but my little laser pyrometer doesn't go nearly that high, and the texture/resistance also felt right.

[Frosty]

That's Way too hot. Did you look up the melting temp of aluminum? 1220f is pure al, casting alloys tend to run lower say 1150f +/- depending on % and alloying metal.

You should just be able to see red in a dark room, at orange temps it's oxidizing as soon as it starts to pour and the flux is no longer keeping air off it. 1250f is pushing too hot but should be okay.

It isn't filling fast enough, hotter is not the answer. You can see how it was solidifying by the swirl patterns in the casting. You need to improve the flow rate so it fills the cavity before it cools. Aluminum has a low specific heat, it loses temperature fast, the greater the temperature differential the greater the change. That is why making the melt hotter doesn't do any good.

You have a flow problem. Look to your runners, gates and fill cup to increase the flow rate. You want the fill cup to gulp the melt, not drain nicely.

Frosty The Lucky.

 

[Nobody Special]

Failure 4.0. So, this time I put parting powder on the drag before adding the pattern and cores...big mistake. When I pulled the pattern after ramming the top, I got it a little bit off to one side, which meant the sand cores (i.e. the negative space) ended up where walls are supposed to go. If I had waited on the parting powder, that would have been immediately obvious and fixable. Whoops.

Also, because where the backside walls existed at all they were maybe 1/8" thick if not less, I got heat tears in them. That said, the rest came out decent - risers worked well, very minor divot on the brace in the middle from shrinkage. Better, but still no cigar. Also, the bloody dogs tore into and were eating my lump charcoal. Sigh...

 

[Frosty]

I'm running our of thoughts about your castings, I think it's learn as you go. A spritz of vinegar water on the charcoal will keep the dogs out of it. Do NOT use ammonia unless you want the dogs peeing on it!

Frosty The Lucky.

[Nobody Special]

Success at last! Kind of! I think it may be a case of good enough for government work, but I've finally got something I think is maybe not perfect, but functional for the bed, the hardest casting. Call it maybe a C-.

Mild heat tears on one end, but not nearly so bad this time and not bad enough, I think, to make it non-functional. I think it's likely the pattern I'm using, the change in cross-sections is a bit too much. Heck, maybe I'll fill the cracks with JB weld and paint it! Last photo is four of the five attempts together.

[Frosty]

Congratulations! Did I miss it when you told me you were pouring the mold in that orientation? Or did you try it in the reverse without success? 

If you try it again flip it over so the runners, gates and pour cup are on top so the al doesn't have to flow to the bottom and be forced up into the casting like that. The melt cools in the runners, risers, gates, etc. and you want the melt flowing downward in the mold, not upwards from below.

To get it to flow upwards you need to have the greater mass at the lowest point so it doesn't start to solidify before the entire mold is full.

Frosty The Lucky.