Casting a Model Westinghouse-Type Twin Steam Engine in Iron

[vtsteam]

Hi Tom, the core was coated with plumbago (graphite), per Ironman's practice, and the facing sand had coal dust in it like his @ 4%.

It will clean up nicely if it anneals -- there's a generous machining allowance, particularly in the upper and lower flanges, and the bores are well undersized. There are no casting flaws apparent at all.

[vtsteam]

Annealing worked. I simply put the casting back in my oil furnace, lit it up for a couple minutes until the casting was glowing a dull red, shut the furnace down, covered the exhaust port with a brick, and plugged the burner blast opening. Four hours later, I opened it up and took out the casting, which was still too hot to hold.

It filed easily, and the hacksaw, which skated before, now had no problem biting into the iron.

[tom osselton]

She’s a beaut! What size crucible did you use? I’ve bought a extra #6 clay graphite for when I give it a shot but am thinking I’ll get a larger one later on. Now to find some Ferosilicon I’m pretty sure I saw some in Dads chemical collection otherwise ebay.

[vtsteam]

Hey Tom, I've only used a #6 and not run into anything yet that I couldn't cast, including a 7" chuck backing plate. 10 lbs is no problem, and technically it should hold 20 to the brim, but that would be nuts as far as I'm concerned.

I think 12 lbs is about the limit I'm comfortable with. To put it in perspective, this casting weighs 2 pounds 2 oz. If you multiply by three for sprue, riser, losses, and some extra insurance to pour in ingots if everything goes well, that gives you 6.6 lbs. I melted 7, and had a fair amount in ingots. I haven't felt the need for a larger crucible.

However one advantage I can imagine with a bigger crucible is that for smaller melts, they wouldn't require adding more cold metal during the melt, because you could pack a full charge in it to start with. This would probably cut the time to melt by 30% -- judging by my last melt. -- and fuel, too. I was thinking about that very possibility.

Too bad an A8 costs proportionately much more than an A6 or A10. Must not be a common size....but it would be quite a convenient one for me.

As far as ferrosilicon goes -- I'm tempted some time to try a session without it, and then annealing the casting to see if that works. Ferrosilicon can't add more carbon, so it mainly must control crystallization on cooling. If you could get a similar result by annealing, that would be a boon.

I've also been thinking that the time to anneal, would be for a prior session's castings, placed in the furnace after the present one has been poured, and the furnace shut down -- they stay hot for quite a while after. No fuel needed, that way.

[tom osselton]

That’s nice to know about the #6 crucible I have my dads kiln in the basement so annealing isn’t a problem for me. I’ll have to dig out my furnace from the shed mind you first I’ll have to cut the grass and see if the quads battery charges. Patterns won’t be a problem I have dads in the basement for his steam engine.

[vtsteam]

Tom, what kind of steam engine? How big? Details, man!

Tom have a look at this video by Ironman using an A6 crucible -- he fills that thing to near the brim, 6.6 kg (14.6 lbs), and look how many castings he pours!

[vtsteam]

Well the Westinghouse Twin casting has some hard spots, discovered after chewing up a couple mills. Very disappointing. I blame the disk rotor metal. Maybe different rotors have different compositions, but the one I used for this engine was tough to cast with -- took a very long time to melt, and was chilled even with .35% ferrosilicon. A subsequent trial melt of radiator metal took half the time, and produced nice gray iron with the same amount of ferrosilicon.

I should factor in, though, that the Westinghouse engine casting is pretty thin. Maybe the rotor metal would have worked better for a thick casting. Though going by the ingot, that was completely white (I think, now -- the fact that it looked silvery didn't mean anything). So I might have needed more ferrosilicon even for pouring a thick casting with it.

I've got another half dozen rotors -- from mixed vehicles, so I'm not sure if they will all behave the same. I'm wondering if I mix half radiator metal in with half rotor metal if it will get something with reasonable melt times and better softer iron.

As for the Westinghouse casting -- probably the best for that job would be straight radiator iron -- since that scrap is in thin castings to begin with, that might be a good match. But I'm a little tired of making cores, etc at this point, so it may be a few weeks before I attempt it a fourth time.

ps... After praising the A6 above, I actually went ahead and ordered an A10. I want to cast a faceplate for my new lathe, and that's going to take nearly 20 lbs of scrap iron. I skipped an A8, because it was only 10% cheaper than the A10 but substantially less capacity.

[vtsteam]

Reboot on the third Westinghouse Twin casting:

I didn't have enough time before predicted thunderstorms today to do a full iron melt, so I thought I'd try annealing that last casting one more time. I was thinking maybe I didn't get it hot enough -- dull red, so this time I went for cherry red, and really let it have the full furnace blast. I also put a sprue and an ingot in there from that melt as test pieces -- to see if they changed.

When red hot, this time, I plunged them into a bucket of wood ashes for slow cooling.

This evening I'm really encouraged tio report that the ingot definitely went to gray iron, and the engine seems like it may have done the same. We'll know for sure once I get it on the milling machine. But I'm hopeful! 

[RussellT]

You have me wondering whether it might be possible to anneal some sash window weights.  I have tried machining them previously and found they were very hard.

Russell

[vtsteam]

I'd definitely give it a try Russell! 

[tom osselton]

I’ve been watching “ Windy hill foundry “ on YouTube and he seems to anneal everything that needs machining.

[vtsteam]

Yup, me too, just working up from the earliest videos so far. Though he calls it "stress relief", which would make sense for larger castings, but doesn't for much of the small stuff he does. I think the real advantage is he's probably annealing which is a legitimate benefit since he uses disk rotors as melting stock. I see he adds a lot of ferrosilicon, too.

Maybe he gets better in the later videos, but wow, IMO he over-rams stuff in the early ones, and then you definitely need vents, which means more cleanup and rougher castings. On the other hand he doesn't seem to get any side packing near the pattern at all so he gets a lot of flash and breakaways.

I really think Ironman has got the most impressive molding finesse, and beautiful looking castings as a result. But Windy Hill is good fun to watch, and he has some good and interesting ideas (I like the long screw for pulling patterns, and his disk crusher is cool) and he gives a good idea of what's required for a small production business.

[vtsteam]

Had high hopes for the second round of heat treatment for the third casting attempt, but milling showed it still had some hard spots on the thinnest part of the casting -- the lower crankcase around the base flange ( < 0.25", 6mm thick).

The upper flange machined nicely, and the exhaust boss did as well so there was improvement. But I have to face it, there's no saving this casting. For thin casting sections, disk rotor metal appears to have been a poor choice -- unless additional ferrosilicon might have made a difference. But .35% was already a substantial amount. I do think steam radiator iron might have been a better metal to use, and If I had done, the casting might have been successful.

But another possible approach would be not to core out the crankcase and bores at all, and just open out those spaces after casting. That might actually be easier anyway because coring leaves hard to remove inner surface sand inclusions and scale particularly on a small size casting. And those rough surfaces would requiring milling anyway. So there's not much savings in machining.

Also, small cylinder bores in a twin are unlikely  to be perfectly centered and aligned when cored, that makes a milling cutter the only way to adjust those locations. If machining a solid casting, however, the centers can be accurately located by drilling.

I think it makes sense to core larger cylinders because they can be more easily cleaned and bored afterwards, and you're eliminating a lot of excess metal in the casting, but these were only 3/4" in diameter. Better to drill that out from the solid.

So, one more lesson learned....well several lessons learned!    I may try a solid casting of the twin next. 

[vtsteam]

Fourth attempt:
1/3 Disk brake rotor iron
2/3 Radiator iron
increased ferrosilicon to 0.45%
no cores
About 8 lbs charge
Melt time 40 Minutes to pour

Slag was quite fluid and tended to reform quickly

Pour went well, 1 small ingot left over.

This ingot broken in half shows very fine grained gray iron throughout.

[tom osselton]

That ingot looks good but the ferro silicon seems way high. It would be interesting to see how the two types of cast turn out with the regular amount of ferro silicon and I wonder if the fluidity of the slag is a indication of pouring temp or ferro silicon needed.

[awemawson]

I’d be very interested to find out the composition of the brake disks you’ve used, obviously there’s something in them that’s making them slower to melt so presumably raising the melting point as well as influencing the final casting properties. I wonder if they are actually a semi steel ?

[vtsteam]

Agreed Tom it's high. But I'd like a good casting first, then start reducing the ferro, and increasing the brake disk content gradually to see where the middle ground is:

Also remember, we already have tested single iron types with normal additions of FeSi:


1.) 1.34" dia iron bar, 100% radiator iron, (0.35% FeSi)  --> 100% gray iron.
2.) Westinghouse twin #3, 100% brake disk iron (0.35% FeSi) --> white iron ingots, and chilled spots in casting

Andrew, I wish I had your metal analyzer -- might explain a lot! And for sure the metal is longer melting (higher temp) than the radiator iron. The brake disk melt took an hour, the radiator metal took a half hour, and the latest pour of a combo took 45 minutes.

Of interest, not 15 minutes ago I broke up another disk rotor, and the inner rim bent for several blows of the sledge before breaking -- much more bending than you would normally see in thin cast iron. Ductile iron, perhaps???

What might also complicate things, guys is, I'm pretty sure these rotors (from different vehicles) probably vary in composition. So it may be hard to come up with a hard and fast rule for rotors.

[vtsteam]

Well the operation was a success, (gray iron), but the patient died  , Faulty riser position allowed a major shrink depression.

I should have known better. This particular engine pattern really needs a riser bang in the middle of the casting, right on top of the exhaust port boss. I've seen this particular problem before, and solved it. Not sure why I tried a different riser location -- I guess I just thought a really big one off of the casting would preserve the boss detail.

Here's a photo of the new 4th casting on the left with shrink depression (looks much worse than photo shows) and the third casting on the right, which had the riser on top of that boss and no shrinkage. The riser did kill the detail of the boss, but I could have whittled it back into shape. I'm going to guess that radiator metal also shrinks more than disk rotor metal -- I did get some shrinkage in the iron bar I cast. Oh well.....  More to keep in mind.

Sometimes I long to go back to aluminum casting. Iron is tough. But there is the thrill of really hot scary iron and a roaring furnace, and the reward ......some day....... of seeing what you want finally accomplished.  Uh...I hope.

[vtsteam]

Nope, probably not ductile iron. According to online sources, ductile has a lower melting point than regular gray cast iron. Probably not semi-steel either -- that's an old low grade of iron scrap mixed with steel, according to online defs. Unlikely to be used in rotors

One possibility is a modified gray iron with Titanium. Another possibility is "Automotive Malleable Cast Iron", which has the highest melting point of the common cast irons, lowest carbon, and produces a solid white iron that must be annealed @ ~ 1650F. At least in characteristics, that sounds like the particular brake rotor metal I had. If brake rotors were frequently of that type, it would also explain the Windy Hill foundry always kilning the castings.

Brake disk compositions do vary, so that one rotor may act completely differently than another. In fact it was the second rotor I've tried and I don't think that the first one I melted behaved similarly.

Modifying the rotor metal with radiator metal would raise the carbon, silicon, and phosphorus contents, and lower the melting temp. I could also try adding carbon itself -- similar to the way Ironman made cast iron from steel.

Much to experiment with and think about.... 

[vtsteam]

Andrew, if I sent you a small sample of that disk rotor, would you be interested in giving it a go through your metal tester?

[awemawson]

Steve I'd happily do it, but my tester is intended for aluminium alloys so I don't think that we'd learn much sadly.

[vtsteam]

Didn't realize that, Andrew!   

I think on today's menu, we will re-melt the last casting, sprues and riser and try again. I will have to make up the roughly 20% loss with new scrap. Choices are:
radiator iron
left over disk rotor iron
or a combination

I'm favoring using more of the disk rotor iron attempting to reduce shrinkage. But that probably will also reduce carbon (already reduced from a second melt, as well) so maybe add some plumbago (graphite) to the charge?

The big question is how much ferrosilicon to add.
Normal is  .25%,
disk rotors (Ironman) .35%,
last pour (gray iron) .45%

Since re-using the last pour metal, it may have a reasonable amount of silicon in the iron already, but I've also heard somewhere that ferrosilicon's benefit only lasts about 5 minutes after addition. And re-melting the last pour will cook it for a lot longer than that. So do I trust that It is okay now to add none, or 25%, or .35%  (I'm guessing .45% won't be necessary.) 

Well, let's pick a number..........ehhhhh  .35% ..........let's go for it

[vtsteam]

eh, no good. Really bad melt. Crucible tipped to one side of the furnace, and I didn't notice until near the end. Also lots of clumpy sticky slag that stuck to the walls of the crucible. Maybe it was sand in the sprues, maybe it was the plumbago, too. Anyway, slagged it twice, but then watched bits of slag detach from the walls of the crucible while pouring and slide down the sprue. I'm sure this casting will be scrap, even without breaking the mold open. Bummer!

Well, I ain't quittin'! 

I have to check the condition of the furnace because I think I'm getting some leakage through the lining in one area. Anyway, next time, all new metal.

[vtsteam]

Post Mortem:
Yup to the plumbago being a problem. It was found in a clump in the bottom of the crucible -- still powder ---which is odd because it is far lighter than iron and should have floated. It was at an angle and the only thing I can figure is since the crucible was tipped over against the wall of the furnacethere was a cold pocket of slag and colder metal trapping it.

I cleaned out the crucible thoroughly.

Also yup to the furnace having leaks. I found some cracks in the hotface near the bottom, and oil under blower pressure was building up in the blanket insulation behind. This led to copious long lasting smoke through a few small holes in the stainless steel jacket after the furnace was shut down. It also meant difficult to control mixture settings while burning. And loss of fuel and heat.

I patched the interior with ganister and then satanite, and did a small wood burn to set them.

Finally late today I decided to re-melt the metal from the last pour to see if the problems were external, and not the actual metal itself. I put sprue, casting and riser back in the newly cleaned crucible, and added 2.5 more pounds of mixed disk rotor and radiator metal to bring it to 7 pounds again. I added 14 g of ferrosilicon and poured into two vertical bar molds, one round (1.335" dia.)  and one rectangular (1"x2" section). The melt went much better, slag was manageable. Melt took 1/2 hour. The furnace hardly smoked, and seemed to heat more quickly.

I'll be interested to see how much shrinkage I get in these test pieces and how hard the resulting metal is, and how much slag got in. Yesterday was a major disappointment, but today, i feel like I'm back on track, and there's the possibility of not only salvaging the metal from last time, but getting some more useful stock out of it. Fingers crossed!

ps. On the off chance that one more try might soften up the hard spots on casting #3, after the pour, I put the casting in the empty crucible and fired up the already hot furnace for a couple minutes, then shut it down and covered the openings. Couldn't hurt!

[vtsteam]

Well great day in the mornin'! June 1st and all the castings did well, but especially Westinghouse Twin #3, which was fully annealed by the last (and third) go at it. I was able to machine all mating surfaces with absolutely no chill hardness anywhere.

What was different? Well putting it in the furnace immediately after another iron melt, inside the crucible, and then giving it more blast for about 5 minutes really heated it up. Then shutting the furnace down and plugging the exhaust with a brick, and the ituyere with a rag. Then leaving it overnight to cool.

That was a real demonstration of the furnace's blanket insulation's efficiency, because when I took the casting out it was still quite warm after 20 hours of cooling! There's not a large amount of thermal mass in this furnace. It's really the insulation doing the job of heat retention.

What was different than the other attempts at annealing this same block?

Well, the in the first annealing attempt I added it to the furnace after a melt, but didn't fire up again, thinking that residual heat would be enough.

The second time, I heated the part up in the furnace from cold -- it was a brief firing to get the part cherry red, and hadn't been preceded by a melt, so the walls of the furnace hadn't stored much heat. Instead, I buried the part in a bucket of wood ashes to slow cooling.

So the third time was the charm, and lots of initial heat in a sealed furnace overnight seems to work the best.

As for the other two castings using "bad metal" from the fourth Westinghouse casting attempt....well the metal wasn't all that bad after all. Must have been the operator. in the photo are the resulting two bars. The round one cleaned up nicely to 1.25" with only a few tiny slag imperfections. Little slag at the very top. No apparent shrinkage. Nice gray iron easily machined. Certainly useful as a post for a Drummond style QCTP, etc.

The rectangular bar shows a little more shrinkage mainly because the center cools slower than the faces and corners. This results in slightly concave sides near the top, reducing as you get to the bottom. Not a lot of shrinkage but some. Little slag near the top. I haven't machined it yet, but expect it will be similar to the round bar, if not better. It would probably be fine to make a couple of tool holders to fit the above Drummond style post, or any other similar purpose.

Anyway a red letter day -- where everything that seemed to be going wrong last week did an about face, and went right!