Casting Aluminum Pistons for Internal Combustion Engines

Here is an interesting video I ran across today.
I will provide some feedback concerning my ideas of what may be causing some of the problems in this video.
This feedback is not intended as criticism of what this individual is doing, but rather commenting on how he could possibly improve his process.

He is experimenting with various types of sprues, runners, pouring basins, etc., and slowly figuring out which method works best for him.

I have been down this same path, and so perhaps I can save a few folks from repeating some of my errors, and other's errors.



.

Another video from the same individual, casting pistons for a motorcycle.


.

Here is another piston casting video from a different individual, for comparision with the videos above.
This individual comes from a professional casting/foundry/metalurgy background, and so his success rate is very high.

There is another video, in which olfoundryman mentions a book by Steve Chastain called "Making Pistons for Experimental and Restoration Engines".

I found this book online in pdf format.
I don't know how to post the link here, since when I click on the link, it does not go to a website, but rather downloads a pdf of the book.

Steve Chastain has a BS in Mechanical Engineering and Material Science from the University of Central Florida, and has written several books about backyard casting.




.

Motorcycle cylinder die casting.
I think some of these molds are water cooled, to control and accelerate solidification.

Diecasting a piston.
This does not have all the risers and such that olfoundryman uses, so it makes me wonder if those dual risers are really necessary.

Some interesting casting and machining techniques in this video.
These pistons are aluminum, not "iron" as described in the video caption.

Seems like pistons are routinely diecast.
I have not found a video yet for commerical mold casting pistons.

I have not done any diecasting, since the mold making process would be very tedious, and this process would lend itself to mass production, not a one or two off hobby model engine part.

Here is a backyard piston diecasting arrangement.
I don't think I have ever seen an in-grade hobby furnace.
The interior die can't have any overhangs, so the casting is not quite as sophisitcated as it could be, but seems to have pretty decent results, with a few defect areas visible.

Here is a piston casting method without the risers seen in olfoundryman's video.
So it would appear the risers on the sides are not necessarily required.
I don't seen any noticeable shrinkage on the bosses.
The original piston was used as a pattern.

It looks like olfoundryman copied Steve Chastain's sprue/runner/gating method.
I am not convinced that Steve's method is ideal, but his book illustrates him casting multiple pistons for use in an automobile, so he is making usable pistons.

The shrinkage and solidification appears to vary with the exact aluminum alloy used, and so that could make things tricky.
Steve uses the term "micro-shrinkage", which I have not heard before.

I use aluminum 356 alloy, and I am not sure how that would work with a piston.
I would probably attempt to cast a piston in gray iron.

Steve Chastain's book has quite a bit of information about piston and piston ring design.

I am not sure exactly how I would gate a piston, but I would not use Steve's gating/riser method as a starting point, even though it does appear to work well enough.

From the video above (post #9), they use one simple gate into the top side of the piston where he wrapped the leather band, with a very short sprue, and that appears to work well, with no risers used.

Picture is from video above.
To cast a piston that has the same skirt diameter as the original, he really should have wrapped a piece of card stock around the skirt, to give it slightly more diameter, to accommodate for shrinkage.

There would be some waterfall effect using this method.

From olfoundryman's video, he shows a section of the riser, and mentions that the alloy he used has a high shrinkage rate.
The high shrinkage rate may explain why Steve Chastain uses large risers on either side.
The defects inside of the riser are most concerning.

This is a screencapture from olfoundryman's video, showing the riser section.
Makes you wonder if those problems propagated into the piston itself.

The checkerplate surfaced chill that olfoundryman used is most interesting.
That would have to be coated with some sort of sprayed-on ceramic coating to prevent sticking.
I have not used chills yet, since I have not had a need for them so far.

.

Haven't seen any new videos from Olfoundryman for almost a year now, I know he was suffering from major back pain. I enjoyed his videos.

I probably will not "enter the game", it is a rabbit hole. I am just "coffee table" TV-casting....
As for the gates and risers, I would think it is a big difference if it is for a mass or medium output production, allowing and justifying to minimize the gating and sprues and doing various different trials.
If only for a few pieces for whatever I am trying I would probably choose to play safe and I would not care if it could be made "better". It is like trying to get the run time of a CNC program down. After 2.5 hours tinkering with the computer machine time is down from 60 min to 45 min. For three parts the slow program would have been "good enough", not worth trying. For mass production, they will not stop after 2.5 h trying.

Questions: What is the difference between gravity casting and die casting? Is that the same thing? For a large operation, would they use Metal injection molding to make pistons?

How important is the temperature. I would guess a) temperature control garden vs. factory; factory wins. b) taking small amounts from a big furnace helps further.

Cooling times might be different between die and sand casting. What are the dies made of? How hot are they? (probably another rabbit hole)
My guess is that the gating and riser is not 1:1 comparable between the two methods.



Short video, they take the metal from big pot. And the die ist slowly tilted to get even flow.

One argument of old foundry man against CNC is that the pistons have somewhat complicated undercuts that would make things complicated. I can see in the die cast piston, that it does not have the undercut. (no idea if that is really important for the function, just spotting a difference)

Greetings Timo

I've had a few die cast parts in various kits and these were not particularly large runs, one there were only 5 kits made, the other a couple of 100 over a 10-15 year period. Also know someone else who has the casting dies for a number of small IC model aero engines and did small runs of those so certainly an option for home use and even a V8 would be worth making a die for as you would want a spare piston or two.

Picture is the simple die for making an internal gear and another part both in a zinc alloy.

Piston undercuts can be done but you need a multiple piece die for the inside shape so it can be removed a piece at a time. Though I suppose you could use a sand core for the internals and a die to locate it and cast the outside shape.

I cast the piston for my mini diesel engine. My technique was sort of based on Olfoundryman's but I just used one riser feeding a pair of gates.

The alloy used was a piston specific one with added nickel.

timo_gross said:

I probably will not "enter the game", it is a rabbit hole. I am just "coffee table" TV-casting....
As for the gates and risers, I would think it is a big difference if it is for a mass or medium output production, allowing and justifying to minimize the gating and sprues and doing various different trials.
If only for a few pieces for whatever I am trying I would probably choose to play safe and I would not care if it could be made "better". It is like trying to get the run time of a CNC program down. After 2.5 hours tinkering with the computer machine time is down from 60 min to 45 min. For three parts the slow program would have been "good enough", not worth trying. For mass production, they will not stop after 2.5 h trying.

Questions: What is the difference between gravity casting and die casting? Is that the same thing? For a large operation, would they use Metal injection molding to make pistons?

How important is the temperature. I would guess a) temperature control garden vs. factory; factory wins. b) taking small amounts from a big furnace helps further.

Cooling times might be different between die and sand casting. What are the dies made of? How hot are they? (probably another rabbit hole)
My guess is that the gating and riser is not 1:1 comparable between the two methods.

Click to expand...

When I started learning foundry work, I found a book on gating/risers/runners etc., and so I was using those methods with aluminum.
Then I ran across some videos that were made by Bob Puhakka, that were posted by a hobby casting guy (these videos still exist), and Bob said he learned his methods from John Campbell.
Like the other hobby guys who saw these videos, we said "Who the heck is Bob Puhakka ?".
Turns out he runs a very successful aluminum foundry in Canada, and makes large and small precision castings, sometimes all the castings for all the mechanisms in an entire ship.
Bob often has his castings x-rayed (perhaps all of his castings are x-rayed), and they don't have defects in them, externally or internally.

Bob's methods are totally different from the book I was using, so I changed to using Bob's methods, since what he said made perfect sense, ie: follow John Campbell's 10 Rules for good castings; avoid high velocity flow, and avoid air bubbles/entrained material in the metal flow; avoid bifilms, etc.

Now when I see someone illustrate how to cast something online, I have to wonder if they are using a good method or not.
Looking at that section that olfoundryman cut on the riser, that large amount defects is exactly what Bob Puhakka discusses, and Bob's methods avoid such defects.

I have watched some mold fill simulation programs on ytube, and dead end runners, like the ones Steve Chastain uses, are said to be not a good thing, because the wave front of flowing metal hits the end of the runner, and then bounces back, and causes the metal to spray into the mold cavity through the gates, which you don't want to do.

I you use an open-top spin trap at the end of each runner, then you don't spike the metal flow pressure.
The spin trap is a place for all the entrained air, slag, loose sand particles, etc. to accumulate without getting into the mold cavity.

.

Isn't that the idea of using risers. They hold metal to feed the actual casting as it cools so you would expect to see the shrinkage in the tops of the risers which means they have done their job

I have exchanged a few emails with Bob Puhakka, but the challenge is to scale down the methods he uses to make very large castings into something on a hobby level.
I think I have done that, and so far, the results are very good, with no defects.
I use bound sand, so in order not to waste non-reusable material, I have to get every casting right the first time.
There are no "do-over's" for me.

Bob's methods are constantly changing, and I think he uses a tilting method now, probably with bottom mold feed, to avoid having to pour down a sprue.
I don't try to keep up with Bob's methods, but rather have locked in on a runner/riser/gating/spin trap method that works consistently.
Due to what Bob has said, I can understand and explain why his system works very well.

I have watched olfoundryman's videos to learn about die casting, and it would appear in some cases that he heats the dies prior to a pour.
Some of the videos show the dies being cooled in water after each pour.
At any rate, the dies are generally coated with a sprayed-on non-stick material between each pour; probably a ceramic slurry.
I don't really know a lot about the die casting process, but the cooling and solidification is very fast.
Olfoundryman uses the diecasting process to make some superb automotive carburetor bodies, which are most impressive (he was a commercial foundry guy and chief metalurgist, so he is no hobby guy).

I would call "gravity casting" a method where you introduce metal into a sand mold by pouring or rotating the mold, letting gravity do the work.
In the die casting videos, they seem to just pour into the dies, and so that also seems to qualify as "gravity casting".
I have seen plastic injection into dies, that that is a forced material, not gravity.

Temperature is quite important, since that affects whether the mold will fill completely, and also affects the surface finish of the casting a great deal.
Some pyrometers are not very accurate, and it is not that important to know an exact temperature, but rather it is important to know what temperature reading on your pyrometer works for your application, regardless of whether that reading is calibrated or not.
I pour aluminum at 1,350 F, which is slightly on the hot side, but not excessively hot.
Some use a lower pour temperature with aluminum, when they know they can get a complete mold fill.
Scooping out of a large furnace with a ladle would help keep a consistent temperature, but a pyrometer works quite well with aluminum.

Cooling time with die casting is very fast.
Cooling time with a sand mold I think would be slower, which could be problematic, and as olfoundryman mentions, is probably the reason he and auto manufacturers use the checker-plated chills, to begin the solidification process from a known point, so that it progresses across the mold without having solidification on opposite parts of the mold cavity simultaneously (one of the things John Campbell mentions should be avoided).
You will get hot tears in the casting if you have solidification on opposite sides of the mold cavity, due to shrinkage.

The dies can be made of anything that will withstand the molten metal temperature, and withstand the wear and tear of multiple uses.
For small quantities, the molds could be made of graphite.
For large production runs, I would guess the molds are CNC'ed from high strength steel.

I have seen olfoundryman use a propane torch to heat his dies, and I think for a very thin intricate casting such as his carburetor body, the dies must be heated to allow for a complete fill of the mold.
The key is to spray on a commercial non-stick material on the dies with every pour, else you will just stick the metal to the die somewhat permanently.

Comparing the gates/runners between die castings and gravity castings is tricky because I am not sure if anyone has illustrated a good gravity casting yet.
Most commercial pistons seem to be die cast for mass production.
As shown in some of the videos, in some less sophisticated countries, sand casting pistons seems to be done with good success, although sand casting is a much slower and less efficient process.

Typically, the wall thickness of castings for engines should be uniform and consistent, and so I think one challenge with casting pistons is that the bosses create a much thicker area than the wall or top of the piston, and so if you are not careful, the walls will solidify first, and then they will draw molten metal from the bosses during shrinkage, and could lead to shrinkage defects in the boss area.
Some minor shrinkage in the boss area would probably be ok, as long as the walls and top of the piston were correct.

Shrinkage in the top or walls of the piston would not be good, unless perhaps it occurred at the bottom of the skirt.
.

Jasonb said:

Isn't that the idea of using risers. They hold metal to feed the actual casting as it cools so you would expect to see the shrinkage in the tops of the risers which means they have done their job

Click to expand...

Bob Puhakka mentions in his video series that with aluminum, it is important to have an unbroken laminar flow into the mold cavity that does not break the solid bifilm surface of the metal.

Turbulence breaks the bifilm, and churns it into the molten metal, with will cause defects in the casting.
Sometimes the defects are visible in a sectioned castings, and sometimes they show up under magnification, but defects are present if you break up the bifilm during the pour.
This is why it is critical not to stir an aluminum melt in the crucible when it is melting; you are just stirring bifilms into the molten metal.

The velocity of the metal must be controlled, else the metal will spray into the mold cavity through the gates, which causes many defects.

One way to control velocity is to use a filter like olfoundryman is using.
One good way to churn a lot of bifilms into the metal is to use a pour basin, which rolls the metal over onto itself, entraining the films into the metal.
You never want to churn or roll aluminum as it is flowing into the mold.
The bottom of the sprue should transition smoothly into the runner(s), without an abrubt change in section.

I pour directly down the spure, without using a pour basin.
While the initial surge of metal down the sprue will have entrained air and bifilms in it, once the sprue is full, it will not have entrained air in it, and will have minimal bifilms if the lip of the crucible is at or touching the top of the sprue.

The initial entrained air and bifilms travel down the runners, which are lower than the gates, and flow into the spin traps, where loose sand, bifilms, entrained air, etc. are spun around and trapped. The spin trap is open to the top of the mold.
Molten metal does not begin to enter the mold cavity until the runner system has been swept, has been preheated by the flow of metal, and until the molten metal reaches the gates, which are at the top of the runner.
I use V-shaped runners to enhance the sweeping effect, and to allow the runner to pull from the sand mold.

I don't use a filter, but rather use the gates for velocity control.
I use somewhat thin and wide gates, and try to feed the mold cavity from two sides if possible.
The mold cavity should fill as fast as possible, but without any splashing or turbulence (laminar fill).
You can see splashing and turbulence in some of the computer fill simulations, and that needs to be avoided.

One benefit of using the gates for velocity control is that the dimensions of the sprue and runners are not critical, since they are not for flow control.
I see a lot of people running all sorts of calculations for exact sizes for the sprue top and bottom, the runners, etc., and none of that is really required.
You just have to be in the ballpark with sprue and runner sizes, and be sure the sprue is not too large, else it will not remain full, and will aspirate air.

There should not be any basins, or abrupt turns or changes in dimension of the runners beyond the bottom of the sprue.
Everything should be very smooth to produce a laminar flow.

I lay out any required risers after laying out the sprue, runners, and spin traps.
The riser is generally connected to parts of the casting that may be thicker than the remaining casting areas.
I don't add risers if I don't think they are necessary.
The green twin base was a large thin casting, with a few bosses on top, and no risers were used when casting it, and there were no shrinkage defects.

I suspect that pistons can be successfully cast with no risers.
The fact that olfoundryman's risers are full of defects would seem to indicate that the pour basin is churning a lot of air and bifilms into the melt, which are settling into the riser.
The filter does nothing to remove entrained air from the melt.
While the spin trap looks like a riser, its purpose is to catch the initial sweep of the sprue/runners, including slag, loose sand, entrained air, and churned bifilms.
You should never seen all that material flow into a real riser, because if all that is flowing into the riser, then there is a good change it may be flowing into the mold cavity.
A riser is not designed to capture material, but olfoundryman's runner does seem to enter the bottom of the riser at a bit of an angle, which is how a spin trap is set up (runner is on a tangent to the spin trap centerline).
But a spinning vortex of defects should not be rising up next to the gates.
The fact that olfoundryman's riser happened to catch some trash just proves that he has a lot of trash flowing where it should not be.

.
(photo below is a screencapture from olfoundryman's video, and shows his runners, filter, gates, etc.)

Thinking off the top of my head, I think I would try casting a piston upside down, with two knive gates into the bottom (actual top of the piston), with a horseshoe runner system like olfoundryman, but with the runners extending byond the gates into spin traps, and omit the risers.

I would try it with no risers first.
If there was a bit of shrinkage at the bosses, then I would add a small riser above each boss.
I would not use a chill at the bottom of the casting, but for some aluminum alloys, it may be necessary.

The backyard casting folks I am aware of generally avoid using pistons for scrap, and it is probably due to excessive shrinkage and other problems that you may run into when trying to use that alloy of aluminum.
If you do use pistons for scrap, you may need chills, and more or larger risers.

One way that I avoid casting defects is to use consistent scrap (another of John Campbell's 10 rules).
I use approximate 356 aluminum ingots from the same supplier, or actual 356 aluminum ingots from a foundry supply house for critical applications, and electric motor end bells for gray iron scrap.

Changing alloys can open up a can of worms.
I wonder what the downside of using 356 alloy for pistons would be ?
The alloy used for aluminum pistons seems to be tougher than 356 alloy, and perhaps more high temperature resistant.

As I mentioned, I would normally use gray iron to cast my pistons, but it would be nice to be able to cast them in aluminum also.
.