In the early ’60s, Detroit had a brief flirtation with aluminum engines, but high costs, manufacturing problems, and warranty headaches soon drove automakers back to cast iron, taking advantage of new manufacturing techniques that allowed an iron engine to be almost as light as an aluminum one. In the February 1962 Motor Trend, writer Roger Huntington explained these “thinwall” casting techniques, which helped to make cast iron more competitive with aluminum in cost and weight.
The new short-stroke OHV American V-8 engines that began to appear in the late ’40s were much lighter than the engines they replaced, but with typical dry weights between 600 and 700 lb, they were still quite heavy in absolute terms. For example, a 1949 Cadillac OHV V-8 was 188 lb lighter than the 1948 L-head engine it replaced (and saved an additional 33 lb if you count radiator, oil, and coolant), but the 331 cu. in. (5,425 cc) OHV engine still weighed a hefty 699 lb dry. An early Chrysler FirePower V-8 was around 30 lb heavier still.
The 1949 Cadillac V-8 displaced 331 cu. in. (5,425 cc) and weighed 699 lb dry, 772 lb with radiator and fluids
The 1951 Chrysler FirePower V-8 was a bit heavier than the Cadillac engine — by perhaps 30 lb dry / Scott Moseman
To understand how Detroit engines could be made lighter while still retaining cast iron, it’s important to first understand how engines are made using sand-casting. As Roger Huntington explains on the page below:
A pattern is used to form the shape of the desired part in a mold of fine green sand. Then the pattern is removed to leave a cavity the shape of the part. Molten metal is poured into the sand mold — and when it hardens we have a metal piece the shape of the pattern … If we want a hollow part in the casting, we must make up secondary sand molds that can be supported in the original mold cavity, so the molten metal will flow and harden around them. We call these “cores”—and, of course, any complex casting like an engine cylinder block or head will have an elaborate set of cores that set in the mold. (They break up and shake out after the metal hardens.)
Forming these cores involved several practical challenges:
Since the pieces have to be suspended in the mold cavity they have to be quite stiff and firm — and yet brittle enough to easily break up and shake out when the casting hardens. Normally, these cores are made up of a mixture of fine sand with cereal and a synthetic oil. The core is formed by injecting this mixture into a metal cavity under high air pressure. After the mixture sets up, it is removed from the metal core box and baked in an oven for up to four or five hours to harden it.
Since this curing process required the moving the cores before they were fully hardened, it created many opportunities for the core to become distorted. Gray iron could also become porous and brittle when cooled, which could also cause weak spots. It was difficult to prevent either problem, so the design of the cores had to allow for these things:
In black-and-white figures, foundrymen actually expect core distortion of .020 to .040-inch-—-and for this reason, wall sections cannot be designed thinner than about .140. This means unneeded weight.
So, to sum up: If the cores could be made more precisely with less distortion and porosity, the engine wall sections could safely be made thinner, saving a significant amount of weight.
On the above page, Huntington explains that in the early ’50s, Ford developed a new crankshaft casting technique using preheated sand and thermosetting resin to form a smoother, more accurate mold cavity. This technique wasn’t originally used for engine cores, but Ford later realized that it could be.
1960 Ford Falcon Tudor sedan / Mecum Auctions
During the development of the all-new small six for the Ford Falcon, Ford engineers were under enormous pressure to save weight, with weight analysts studying the mass of every individual part. To keep the engine as light as possible, Ford applied the resin binder/hot-box technique to making the cores so that the casting walls could be unusually thin. As Huntington explains:
They now make their cores of sand and resin, and this mixture is blown into pre-heated metal core boxes that cure and harden the core in 20 to 30 seconds. The core is never removed from the box and baked in an oven, as previously. It’s formed and cured in the same cavity. The reduced handling and quick-curing have cut distortion to almost nothing. Casting walls can be thinned down to .100-inch. And it’s turned out to be quicker and cheaper, too.
The thinwall 144 cu. in. (2,365 cc) Falcon six weighed just 345.5 lb dry / Mecum Auctions
Using these techniques, the dry weight of the original 144 cu. in. (2,365 cc) Falcon engine was limited to 345.5 lb, only about 25 to 30 lb heavier than a 97 cu. in. (1.6-liter) BMC B-series four or the Volvo B16 engine despite having two extra cylinders and almost 50 percent greater displacement. Ford then refined this process further to hold the dry weight of the new 221 cu. in. (3,620 cc) Fairlane V-8 to around 450 lb (not including flywheel or clutch).
The 1962 Ford Fairlane introduced the long-running thinwall Ford V-8 / Ford Motor Company
Other automakers were not as quite aggressive as Ford was in reducing casting wall thickness (although Buick came close with its iron-block 300 cu. in. (4,923 cc) V-8), but they immediately recognized that these “hot-box” core curing techniques could save weight and money. Reducing the amount of cast iron going into each engine saved dollars as well as pounds, but that was only part of the equation: Hot-box cores cured in seconds rather than in hours and required less handling, which saved time and labor as well as materials. More accurate casting, in turn, reduced scrap rates during manufacturing.
Ford’s original 221 cu. in. (3,620 cc) Fairlane V-8 — the photos at the bottom of the page below show the cores used in its casting / Ford Motor Company
Ford and GM also found that they could safely reduce the minimum wall thickness by using a more ductile, less brittle iron alloy. The fact that automakers were willing to pay International Nickel Co. a $6-per-ton royalty to use their patented ductile iron-magnesium alloy speaks volumes about its value: Detroit in this era hated paying royalties to outside companies if it could be avoided.
Huntington offers some weight comparisons for the cylinder blocks of similar engines, cast both using conventional techniques and hot-box (thinwall) casting:
Block Weight, Conventional Casting
-
- Rambler Six (196 cu. in.): 155 lb
- Lancer-Valiant Six (170 cu. in.): 135 lb
- Lark Six (170 cu. in.): 118 lb
- Rambler V-8 (250 cu. in.): 161 lb
- Chevrolet V-8 (265 cu. in.): 147 lb
Block Weight, Thinwall Casting
-
- Chevy II Six (194 cu. in.): 123 lb
- Falcon-Comet Six (170 cu. in.): 83 lb
- Buick Special V-6 (198 cu. in.): 105 lb
- Fairlane-Meteor V-8 (221 cu. in.): 120 lb
These were not night-and-day differences, but they were significant, and, more importantly, the new techniques involved weren’t prohibitively expensive to implement. Aluminum castings could save even more weight, but they used costlier materials, and manufacturing them required significant investments in specialized production equipment. Huntington points out that the thinwall iron block of the 170 cu. in. (2,780 cc) Falcon/Comet six was only 18 lb heavier than the significantly more expensive die-cast aluminum block offered for the Chrysler 225 cu. in. (3,682 cc) Slant Six around this time.
Given how much trouble U.S. automakers had with their early aluminum engines, lighter cast iron engines seemed like a better compromise, offering some of the weight savings at much lower cost, with fewer headaches.
The open-deck die cast aluminum block offered for the Chrysler RG Slant Six weighed 65 lb dry (76 lb with its cast iron main bearing caps), but was costly to make
Thinwall casting was not an either-or technique. The actual amount of weight that could be saved by using hot-box core curing depended on the number of cores a given engine used, the type of iron alloy used, the structural architecture of the engine, and how wary the manufacturing staff were of making the walls too thin. Designs that mixed hot-box and conventional cores, like the Oldsmobile 330 cu. in. (5,404 cc) V-8 of 1964, were inevitably heavier than ones that were all-thinwall. Existing engine designs that switched from conventional to thinwall casting (like the bigger Oldsmobile V-8s of the late ’60s, or the BMC C-series six in the MGC and Austin 1800) saw smaller weight savings.
Despite thinwall casting, the MGC’s 178 cu. in. (2,912 cc) BMC C-series six still weighed a hefty 567 lb dry/ Bring a Trailer
Huntington was overly optimistic about the opportunities for further weight savings using even thinner casting walls, at least in the short term. Even Ford’s thinwall engines gradually got heavier in later years, and the late ’60s trend toward bigger and bigger displacements made the use of hot-box core curing techniques more of a holding action than a weight-saving revolution. For example, the thinwall Cadillac 472/500 cu. in. (7,734/8,94 cc) V-8 weighed 680 lb dry, and the thinwall Ford 460 (7,536 cc) was a few pounds heavier than that.
However, improved thinwwall casting techniques — eventually allied to more sophisticated computer-aided engineering and finite element methods — have kept cast iron engine blocks commercially viable for decades after many observers (including me!) would have expected them to be written off in favor of lighter aluminum designs.
Related Reading
CC Vintage Video: Ford Cleveland Foundry & Engine Plant, 1953 – My Father’s Job Of A Lifetime (by George Ferencz)
The second gen Olds V8s from 1964-on were not simply redesigns of the first gen conventional castings. Other than a handful of dimensions, the second gen engines shared nothing with the first gen Rocket V8s.
I didn’t say they were redesigns of the first-generation castings. I said they used a mix of hot box and non-hot box conventional cores, which they did. The 330 block used six cores, three hot box cores, three baked sand cores. ETA: As for the bigger 425, while it shared much of the same architecture as the 330, it was NOT initially a thinwall engine. Oldsmobile eventually switched to some hot box cores, but they didn’t do that at the outset, so it was a change made after the engine was already in production.
The first phrase is doing a lot of work there, as the carryover dimensions had a very strong influence on the design of the second-generation engines, which resulted in their being bigger and somewhat heavier than they might otherwise have been in the interests of reducing the manufacturing investment.
Perhaps that’s why, imo, the Olds 330/350 V8 was one of, if not the best, V8s GM ever built. Gil Burrell and his Olds engine designer peers Ball & Gill were somewhat conservative in designing and construction of their new V8, which wasn’t truly a “small block” in the sense of those of some of the other divisions, but was 154 lbs lighter than the original 303, had bigger bore and shorter stroke, all new (some thinwall, some not) castings, and differing water jacket, oil passages, valve angles, etc. yet some of the manufacturing could still be done with the 394, though it’s days were very numbered, as the 425 debuted in ’65 using thinwall techniques.It has also been reported also that Olds used a different/better ductile iron alloy formula than some of the other divisions, not sure if that’s been verified.
The 330 was, however, about 35 lb heavier than a Chevrolet 327, which was not a thinwall engine in this sense.
The related tall-deck Olds V-8 was NOT initially a thinwall engine. It adopted some hot box cores later in production (possibly not until the introduction of the 455 version), but Roger Huntington reported at the time that there had been some headaches with foundry setup, so the 425 had conventional baked sand cores.
The later ‘short deck’ Olds V-8’s (260-307-350-403) used a light weight block casting, identifiable by its ‘windowed’ main bearing webs. These blocks were even lighter than previous 350 blocks.
As the owner of a Fairlane V8 (’79 302) in my ’83 Ranger 4×4, I noticed no difference in handlin.g over the 2.3L four cylinder it replaced. Of course I have an aluminum intake and water pump. But I feel the water pump doesn’t count, as the original 221/260s came with aluminum pumps from the factory.
Too many popups to read, but I am of the opinion that the new iron was not as durable as the high nickel blocks.
Of course, the body and seat covers wore out and the wealthy Americans scrapped those million mile blocks to get the latest style car.
Not just to get the latest style, but the body rusted out. I never tore down an engine due to cylinder wear. The high nickel block thing I heard many times, but a non-issue in passenger cars or light trucks.
My 1984 Ford Bronco was ordered new with a 302 V8 and a 3 speed overdrive transmission. It was my daily driver for over 10 years until coolant was detected in the engine oil. Did my Bronco have a thin wall cast iron engine?
Aaron Severson, huge thanks to you for today’s great essay. This was “my Father’s world” at Ford’s Cleveland Casting Plant (Foundry), and I’ve toyed with doing such a CC—but didn’t know of the wonderful Motor Trend article. Hooray!
From Dad I learned about cores used in casting blocks, heads, water pumps—but not needed to cast crankshaft bearing caps, ‘cause they didn’t have any hollow internals. Cores used a very particular sand mix, and then there were the binder resins the sand was combined with, the production lines forming the cores and hot-box curing them quickly. (That half of the Foundry seemed almost as hot as the iron-pouring half!).
“Chaplets” were precise little metal button-spacers that helped locate the various cores in precise relation within a master sand mold (forming the outside of the block, head, etc.)—when various sand-cast protrusions on the cores couldn’t be counted on to fully do the job.
My collage below tells an important story: For the very last iternation of Ford’s Flathead, block-casting required 29 separate, fiddly cores, produced and located precisely. It was a small production miracle to get the Y-block’s block down to 14 (I think the Chevy 283 may be even fewer), and then just *8* for Ford’s early-1960s V8.
My own theory is that Ford started with a 221 CID displacement (3.5” bore) just to make sure the thinwall process would work reliably at production-line speeds, then gradually went to 260 CID (3.8” bore) and then 289 (4.0” bore)—I don’t know that the history/literature ever says things in quite that way, but it seems relevant.
Thanks again for a significant CC story!
My Dad bought a ’62 Fairlane 500 with a 221 V8 in April of ’62. My recollection is that it ran very smoothly, like the proverbial Swiss watch, but needed to be really revved up to get moving with alacrity. Luckily that little mill did like to spin, though clearly Ford realized that a bump up to 260 then 289 was needed and did so in very short order.
My Mom’s ’69 Mustang convert with 302 would move out much better if pressed and it had the right front end weight/power balance, unlike those with bigger.heavier blocks out front. A college friend had a new ’69 Boss 302 and that was one fun & fast car. Wish I could afford one now!
Thanks, George. Your memory is correct: The Chevrolet V-8 block used only 12 cores. While they didn’t use hot-box casting, Chevrolet came up with their own technique of pasting the cores together before moving them into the drag mold, so there wouldn’t be shifting or distortion from handling of the individual cores and the casting would be more precise.
For anyone interested, here are block cores for the new Chevy 265, 1955. I don’t know what the “other make” is—-but must it be GM if the cores are available & shown?
There is an article somewhere in CC about sand castings but I cannot find it. I believe it was written by a lady whose father was involved in the industry. As you allude to the science and technology of sand casting was extensive with all sorts of additives, sand types, etc. being used for the “recipe” for the core. And, the technology was always improving as competing vendors came up with improved mixes.
Maybe someone good a search can find the article.
I’m not remembering any such article and can’t turn one up from a search. Maybe somewhere else?
https://www.curbsideclassic.com/blog/video/cc-vintage-video-ford-cleveland-foundry-engine-plant-1953-my-fathers-job-of-a-lifetime/
Yes, that’s it.
Thanks — I added that link to the end of the text.
That was a great explanation, I’d always wondered what technology was used to cast thinwall blocks. Thanks!
Thanks for your article Aaron, what about lost foam casting? I believe Oldsmobile was using that process in the 1980’s. I seem to recall an article in Car & Driver stating that Olds was using the process for cylinder heads and the Olds Quad Four engine. I also believe Saturn used it in the production of their engines. An interesting article!
Iirc “lost foam” engine block casting was first developed by OMC in the mid-60s for use in their new 3 cyl 55hp outboard with loop charging/scavenging, a first in outboards and for engine casting anywhere I believe, introduced for production in 1968.
The second generation Buick large V-8’s that replaced the ‘Nailhead’ in 1967 were also quite light for their size. I don’t believe these engines shared anything with earlier large Buick V-8’s, but were all-new designs inspired by the Buick V-6 and small block V-8 designs.
Thanks for this excellent explanation of something i I have read about for years but never fully understood or appreciated.
Is the weight of the Studebaker 170 six correct? If so, it was amazingly light for its era.
Probably, but keep in mind that the listed weights are only for the block, not the complete engine.
This may be a little off the subject, but after spending part of a very cold afternoon chasing down an antifreeze leak on my wife’s 2004 Taurus with its “sideways mounted” Vulcan V6, I forgot how much room the old cars, such as that Ford Falcon in Picture 4 had under their hoods. You can actually see the wrench on the garage floor, or in my case the driveway, after you drop it.
I found the leak, it’s a loose clamp on the water pump which incidently mounts to the engine front cover, which in turn mounts to the engine block. So much for Ford having a “better idea.”
The best way to change the clamp is to jack up the car, remove the right front wheel, remove the inner fender, and then install a new clamp. That’s why we are going to spring for a 13.00 clamp tomorrow at AutoZone. It pays not to reuse a 20-year-old factory clamp when you replace a water pump on a 2004 Ford Taurus, which my wife and I did last year. Ask me how I know … I’m glad the water pump gasket is not the culprit.
By the way, I enjoyed your article, my dad owned a ’52 Chrysler Saratoga with the FirePower V8, and my aunt drove a Falcon with the in-line six, identical to the one in your article.
Terrific explainer! One thing the article got wrong, cited in your reference to their error on page 44 of the magazine:
Only the RG (225) version of the aluminum Slant-6 was marketed (from 1961 to early in the ’63 model year). That’s the block their 65-pound figure is correct for. It’s 76 pounds with the cast iron upper and lower main bearing caps and their stout bolts, narrowing the difference even more.
Aluminum 170s were experimented with, but never commercialized—though a (very) few of them escaped captivity and survived to tell the tale.
The discussion of Ford and GM getting licences for the Inco extra-ductile iron with magnesium is interesting. Chrysler added “elemental tin” to their engine block iron starting in ’63, for public reasons of greater wear resistance. I’m not a metallurgist, so I don’t know what-all various additives do, and can’t tell if that might’ve been a preparation for the Chrysler thinwall V8 (the LA block) of ’64, or if Chrysler also wound up getting an Inco licence.
Thanks! I defer to your considerably greater knowledge of the aluminum Slant Six, and I’ve amended the text.
This excellent post brings to mind an engine building article I read long ago. The builders used X-rays, if I recall correctly, to sort through factory blocks they deemed appropriate for building.
Cast aside blocks had coring issues like misaligned cylinders-fine for low rpm stock usage, but not worthy for pushing performance limits. This article details how such issues readily occurred. Fascinating.
Compact graphite iron has also made inroads toward futher lightening engine blocks.
I’ve heard a very experienced in Mopar engine builder state the original 225 iron block was heavier than necessary due to wall thickness sizing for aluminum. Post ’76(?) 225 blocks were thinned and lightened as they were iron only.
I’m not nearly as familiar with the Chrysler aluminum block as Daniel Stern is, so I’ll leave that to him, if he wants to weigh in.
The engine builder John.C heard from was correct on that subject. Here are two pages from Willem Weertman’s “A History of Chrysler Corporation’s Slant-Six Engine”. Weertman was Chrysler’s Managing Engine Design Engineer from 1955-1962; Assistant Chief Engine Design and Development Engineer from 1962-1976, and Chief Engine Engineer from 1976-1987.
I’m not seeing how that document supports the notion that the iron block was heavier than necessary due to there being an aluminum block.
In looking at the two blocks, it’s very obvious to me anyway that they were substantially different castings. The shape of the side walls, the open deck, and lots of other details all say that to me. They obviously required two quite different sets of molds and cores to cast.
What I get from that document (and previous reading) is that the aluminum and iron versions were developed simultaneously, but obviously separately too, due to the very significant differences.
There’s just nothing in what I’ve read, heard or can deduce from these two engines to suggest that the iron block was any heavier than it would have been if there had been no aluminum version. They are two distinct blocks, and they clearly did not share anything except conforming to key architectural points to allow them to utilize the same reciprocating parts and external attachments and such.
Oh, sorry! Those pages don’t really address that question, which I interpret not so much as suggesting the aluminum and iron engines used common casting tools (obviously not), but more architecturally. Greater block skirt depth than would’ve been warranted if no aluminum version had been in mind, for example.
I do think those pages pretty well put away the idea that the Slant-6 wasn’t designed to be made of aluminum and that the aluminum 225 was a fad-following afterthought.
Anyhow, here’re a couple more topical pages:
Thanks. That does explain it, finally. I had already come to understand that the two blocks were designed simultaneously, but did not realize that they had to make the block longer at the crankcase to compensate for the aluminum’s weakness.
I’m always happy to learn and change my mind in the light of the facts.
So, is the implication that the iron block was heavier because there had to be enough space to allow for enough wall thickness in the open-deck die cast block within the same exterior dimensions? Weertman doesn’t actually say that in that passage, although I guess that would follow.
I’m still not seeing how “that would follow”. Maybe I’m missing something.
If you’re designing an all-new engine that is going to be available both with a closed deck iron block and an aluminum one that will be die cast (thus requiring it to be open deck, with walls thick enough to provide adequate strength in cast aluminum), and both blocks must have the same or substantially the same exterior dimensions so they can share the same cylinder heads, I assume the block’s overall dimensions are likely to be bigger than if you were designing solely for cast iron (which can have thinner walls than with aluminum, all else being equal), thus imposing a greater weight penalty for the iron version than just the difference in material.
I’m still not seeing it. Here’s a set of pictures of the alu block and the iron block. They are very different, except of course in the key areas they have to be the same. If the basic casting/exterior shape/dimensions were the same for both, I could see your point. But they’re very clearly not.
Here’s alu block #1:
Here’s alu block #2 (other side)
Here’s iron block #1
And iron block #2 (other side)
It’s absolutely obvious that these are not the same basic casting. Which means they were not dependent on having the same interior wall thickness or so. I can’t prove it, but logic tells me the iron block was created and cast to be an iron block, without having to make compromises for the fact that an aluminum version (with very different external appearance) was also going to be created.
Except for the key architectural requirements, the aluminum block was quite different from the iron block visually, with its open deck and other aspects, so it obviously used different molds and cores.
The slant six wasn’t designed to be an aluminum block engine; like AMC, Chrysler thought it would be a good idea to make an aluminum block version as that seemed to be the hot new thing.
If you look at the two blocks, the visual differences are quite significant.
I’m quite sure the iron 225 blocks are essentially identical to the 170 blocks except for the taller deck height. So the 225 weighed a certain amount more than the iron 170 due to that factor.
Great article, Aaron. I never really understood what thin-wall casting meant until this. I thought it just meant they were casting in a tin shed.
I’ve no idea what Chrysler was doing in the iron-casting field in the ’50’s, but they bet their future on the slant six being aluminium, which is the way it was designed first with the cast iron one second, and at that, meant to be a kind-of stop-gap whilst they got the alloy version fully up to speed (figuring the greater expense of alloy would be off-set by a much kinder life to machining tools and much less machining from die casting anyway). Obviously, it didn’t work out: far too much wastage in production, and service issues too.
However, it shows how much that issue of weight was on everyone’s mind, presumably, inter alia, because of that ’50’s import boom with folk favouring cars using maybe half the fuel a trad-size plonker was guzzling.
Most curious to see the weight of that alloy Buick block. Same size as the thin-wall 221 Ford, but HALF the weight! I wonder. I’ve never much liked the standard Rover V8, not terribly smooth, not very revvy, and famously always prone to big leaks. Was it essentially a bit under-cooked for strength, as an early mass-made alloy engine? And had GM worked this out when they hoiked it off to Rover? Given their resources, you’d have to suspect it.
There’s nothing to suggest it wasn’t strong enough, as it was used successfully in racing (Repco) and Rover successfully punched it out to 4.9 L.
Smoothness? It was balanced in the same way as other 90 degree V8s. Not very revvy? It wasn’t designed to be a hi-po hi rev engine. It was a Buick, and they had no genuine sporting aspirations for it. The Olds version had slightly different heads that did breathe a bit better. And many folks were able to squeeze a lot of revs and power out of it. This would have been a factor of how Rover “tuned” their versions. Given its primary use in the RR, revvyness was clearly not in the brief. The whole point was a healthy torque band at lower revs than what typical smaller displacement European engines had.
GM didn’t exactly “hook it off to Rover”. Rover came asking for the rights and blueprints to build it.
The Rover V8 generally seems to have developed a pretty good rep, but it was what it was, a late ’50s design. By the ’70s and ’80s, it was pretty old-school compared to so many newer engines coming out of Europe.
The bare block of the original Buick aluminum V-8 weighed about 60 lb. A complete engine, minus exhaust manifolds, starter, generator, and flywheel, weighed 249 lb dry. Buick quoted the weight of a complete engine with accessories and automatic transmission flexplate as 318 lb dry.
Thanks Aaron. Great explanation.
I may have misunderstood the engine builder. Thanks for the info regarding the aluminum slant six, Paul.
For some reason I can’t reply directly to a comment.
You understood the engine builder correctly, and the engine builder was correct. See my comment further upthread.
See my longer response above.