Semi-Lost-PLA Casting Method

[GreenTwin]

Still seems a lot of extra work and I thought your aim was to cut down time?

My aim is to make it to retirement, alive.

Since I don't really have time for any serious shop work at the moment, these posts are basically just musing about various "what-if's"; pondering the new methods and materials that have come along recently, and wondering how that may factor into what I do in the foundry world.

The feedback from others who may be trying similar things is quite useful too.

As member "creast" mentioned, it can be quite expensive to try a new product/method, and often there is a lot of hype around a new product, but then you try it and find out that perhaps it is not ready for prime time.

There are few things that are more discouraging than spending a lot of time making a pattern, ramming a mold, firing up the furnace, pouring some very hot metal, and then producing a bad casting.

I have basically reached a 100% success rate with iron castings these days, and I don't want to go back to the early days where my failure rate was perhaps 70% or more, because either I was using the wrong material, or I did not know the correct procedure.

.

 

[GreenTwin]

Here is a fellow who I think tried to cast the same gear that I tried to cast, and he did heat his PLA pattern in the oven, but at 340 F, the gear pattern is still quite stiff. I think this was the 2nd attempt by someone to cast these gears for a windmill.

Judging from the video below, I think a PLA gear pattern will have to be heated to somewhat close to the melting point of PLA, in order to make it easy to remove from a bound sand mold.



And a back story about this casting fellow:

He does a lot of iron casting out west, and makes some very long straight edges in iron, perhaps up to 48" long.

An iron furnace with a diesel burner makes quite a loud roar, along with quite a bit of low frequency rumble, and this guy's neighbor complained about the noise.
He ended up putting a stack over the lid opening, with the stack being perhaps six feet long, and lined with lightweight refractory.
He was successful in reducing the noise of his furnace enough to keep it from disturbing his neighbor.

I think he came up with the latch mechanism for his lifting tongs himself.
I am not sure if they are fail-safe (I suspect that they are), but they could be made fail-safe if they are not already.

And I think he has since changed out his rubber pouring cart wheels to cast iron wheels.
Using rubber or plastic around an iron furnace is not really a good idea, and should be kept to a minimum, or avoided altogether if possible.

Be aware that if you are using a lifting hoist like he is using that swings out away from the furnace, it is possible to get too much weight too far from the furnace, and you can tip the entire furnace over on its side.
I would suggest outriggers at the base of the furnace that protrude at least 10" beyond the furthest reach of the hoist.

I have used a crane and pouring cart with larger crucibles and iron (when using a #B30 crucible), and one person can easily handle 200 lbs. using a crane and pouring cart.

I think he uses a ceramic blanket lined furnace, with a coating on the ceramic blanket.
While this type of furnace heats very quickly, I prefer a 1" thick hot face of dense refractory, such as Mizzou.
A rigid hot face and domed dense lid takes a bit longer to reach pour temperature, but will really take a beating with just minor touching up with plastic refractory from time to time.

He uses a combination of sodium silicate and greensand mold components.

I ended up putting dense foam under the wheels of my furnace to prevent much of the low frequency noise from being transmitted into the ground, and that quieted my furnace down quite a bit. I don't use a stack for a muffler.
I will probably remote mount my combustion air blower inside the shop, because that makes a lot of noise by itself.

I think this guy has his blower mounted in a box, which would be the long rectangular box to the left of the furnace.
You can hear the rumble sound in his video with the furnace lid open.
The burner is basically operating like a pulsajet engine, which is also very noisy.
And the rumble and noise are pretty much completely stopped when he closes his furnace at 0:36.

 

[Jasonb]

It would be so simple to do it by dividing the mould up into say 4 quaters as I suggested early on. There is nothing to say you have to pull the pattern out vertically. Alternative is to lay the gear on it's side and use three flasks, splitting the middle one vertically which again gives you the four quaters, lift off top one, slide middle two sideways and then lift the pattern out the bottom one. Same sort of thing that is sometimes done for decorative cast iron like heavily fluted columns on lamp posts, bollards etc but just a lot shorter.

If cast with gear vertical using bound sand you could do two opposite 1/4s and then do the other two before the first has set so would not even add much to the time. Certainly less that it would to heat the pattern and the mould that it would be in upto 340F plus.

I have heard people say that the 3D prints shrink slightly as they cool, I assume the opposite could therefore apply when you heat then causing them to expand in the mould and get tighter not looser?

 

[GreenTwin]

It is really the helical gears I am considering, which I think will be challenging.

I am not sure if I should waste my time trying the lost-PLA process with the Frisco Standard gears, or just go straight to an investment process.

I would have to find an investment process that was iron-rated, and not just plaster or something similar.

We know the mini Wilton vices were cast in steel using the investment process.

.

 

[ddmckee54]

Jasonb:

3D printed parts will indeed shrink as they cool, and they will expand when heated. Clough42 demonstrated this quite nicely when he was 3D printing a test rig for his ELS PC boards. The line-up pins in the 3D printed board holder didn't match up to the board when the holder was printed at a 100% scale factor, the holder was too small.

That's also why when 3D printing a pattern to be used in Lost PLA, it is recommended to use minimal wall thickness and minimal or no infill. The theory/hope is that during the burnout cycle, even though the pattern is trying to expand as it heats, it is also softening. With the minimal wall thickness, and minimal or no infill, the walls hopefully will deform into the hollow spaces within the pattern instead of expanding outwards cracking the mold.

Don

 

[Jasonb]

Thanks, that should save Pat barking up the wrong tree and trying to heat his pattern.

Pat the lost PLA Dip & Burnout) method would be the easiest to try, you only need buy a suitable material to dip them in. If going full investment then you have to sort the waxes, flasks, investment and ideally vacuum to suck the bubbles out. So a lot to sort when the dip & burn may do the job.

 

[GreenTwin]

If the helical gear patterns have a removeable core, then it is unlikely that the teeth part of the pattern will expand so much that it will damage the bound sand.

Getting the bound sand packed tightly into the gear slots may be a bit of a trick.
I will try to get out to the shop this weekend and at least give the bound sand / heated PLA method a try, to either validate it or rule it out once and for all.

Obviously the bound sand method is very simple and quick, if it works, and does not require me to purchase some short-shelf-life material that I may not get time to use again before it goes bad.

Thanks much for the feedback from everyone.

.

 

[Jasonb]

Might try one in the freezer

 

[GreenTwin]

Might try one in the freezer

If the bond for bound sand can be broken, then it is easy to remove a mold.
Usually just a quick sharp small impact will do it.

The trick with bound sand is to remove the pattern from the sand during the "strip time".
If the pattern is left overnight in the mold, then generally it is permanently glued into the mold, and cannot be removed without damage.

I may try the freezer thing; it has potential to work.

.

 

[Jasonb]

do you use talc or some other type of release agent?

 

[GreenTwin]

Typically paste wax, ie: as in floor wax.
Coat the pattern, let it dry a few minutes, wipe it off, repeat a few times for a new pattern.

Edit:
Bound sand does not contain clay, and so you don't ram it in the mold as you would greensand.
The technique is to press the sand in place with the fingertips, then work out and press with the heels of the hands.
Ramming bound sand just fractures it and shifts it away from the pattern, since you cannot really compress bound sand very much.

 

[creast]

I still think that some form of investment casting would be the best path for Pat's gears.
Shell casting is certainly the way for large castings but I have never tried this and from looking at the process it seems a longer and more costly process than the investment plaster method on smaller parts.
There are all manner of slurries and sand types for the shell process all of which are expensive and the slurries need to be maintained carefully. Also the drying time between coats means several days of building up the layers before casting.
The investment plaster method only requires the plaster which if stored correctly in its powder form can last well over a year. In fact my 25kg bag which is now almost depleted is over 2 years old and still performs.
The process of burnout is quite lengthy but I have got away with burnout to 550 deg C over a 4 hour ramped control quite reliably. Often I allow to cool and blow out gently with air and maybe pre-heat again next day before casting.
The added benefit with this method is the vacuum assist during filling with the molten metal being forced into every detail. It can be done on a budget as I do personally all the time!
I can't say if my plaster which is rated for Gold etc is suited to Iron as I have never tried it, but it is great for Zinc, Aluminium, Brass and Bronze. Aluminium tends to come out dull grey with this method and I am unsure why this is.
Obviously this method gets a bit more cumbersome and costly as the part size increases.

 

[GreenTwin]

If I remove the center from the helical gear print, and only leave a thin rim, just enough to support the gear teeth, then it may be possible to use a hot knife, or a sharp tip in a soldering gun, and cut the pattern between the gear teeth.

This would break the print down into sections, sort of like Jason has been suggesting, but with the cuts following the curve of the gear teeth.

I think this is the only chance at removing the gear pattern from the sand without sand damage.
I may come down to hot knife cutting between each tooth.

Damage to the sand mold between the teeth would be ok, since this would protrude upwards between the cast teeth, and could be ground off.
Ideally the sand mold would not be damaged at all.

Edit:

If the pattern is cut into separate teeth, then a slight tap on the end of the tooth should break it loose from the sand mold.
This would be much better than trying to pull the tooth pattern radially out of the sand, since the mold is not going to have any strength in the radial direction.
.

 

[Nerd1000]

If you just need one pair of gears you could possibly have them 3d printed in metal. Lots of commercial services doing that these days. The other thing to consider is whether some stock gears will suit your application. I did get quotes on some a while ago, and they were expensive but perhaps worth it to save the trouble in this case.

Unless of course your objective is to take the journey of casting them, which is totally valid.

 

[GreenTwin]

I have never found a source for 2:1 ratio helical gears of the same diameter.
If I had a source, I would consider purchasing them if they were close to the size I needed.

But with JasonB showing how to design these types of gears, I have really learned a lot, and am determined to make them myself.

I have gear cutters for standard spur gears, and I have an indexing head, so have no doublt I could make those, and thus I purchase those types of gears for IC engines when I need them, and they do come in a wide variety of sizes.
My dad made gears all the time, including gears for the differential for his Stanley Steamer replica auto.

For helical gears, they can be cut, but since JasonB has been kind enough to work out an excellent design that can be 3D printed, I think it is worth a try to cast them.
I don't have good machine tools, but I do have some first class foundry materials and furnace, and so I have a better chance of casting them than trying to get a helical gear cutting apparatus set up and functional.

And another benefit of casting helical gears is that you can make them the exact size needed for whatever scale engine, just by adjusting the scale factor on the 3D printer program.

And since I started casting engine parts, I actually enjoy the foundry side of the model engine hobby as much or more than the actual machining and assembly of an engine.
Pouring molten iron into a mold and making engine parts is just the cat's meow in my opinion; nothing quite as exciting as that as far as my other hobbies go.

.

 

[Nerd1000]

I have never found a source for 2:1 ratio helical gears of the same diameter.
If I had a source, I would consider purchasing them if they were close to the size I needed.

But with JasonB showing how to design these types of gears, I have really learned a lot, and am determined to make them myself.

I have gear cutters for standard spur gears, and I have an indexing head, so have no doublt I could make those, and thus I purchase those types of gears for IC engines when I need them, and they do come in a wide variety of sizes.
My dad made gears all the time, including gears for the differential for his Stanley Steamer replica auto.

For helical gears, they can be cut, but since JasonB has been kind enough to work out an excellent design that can be 3D printed, I think it is worth a try to cast them.
I don't have good machine tools, but I do have some first class foundry materials and furnace, and so I have a better chance of casting them than trying to get a helical gear cutting apparatus set up and functional.

And another benefit of casting helical gears is that you can make them the exact size needed for whatever scale engine, just by adjusting the scale factor on the 3D printer program.

And since I started casting engine parts, I actually enjoy the foundry side of the model engine hobby as much or more than the actual machining and assembly of an engine.
Pouring molten iron into a mold and making engine parts is just the cat's meow in my opinion; nothing quite as exciting as that as far as my other hobbies go.

.

Ahhh I see, they have different helix angles to create the ratio. I thought you were doing normal screw gears, which can be had as stock items:

https://khkgears.net/new/screw_gears.hml

 

[100model]

I have never found a source for 2:1 ratio helical gears of the same diameter.

I think you are talking about skew gears, if so industrial sewing machines have them and are large enough for model engines.

 

[GreenTwin]

The nomenclature gets pretty confusing, and I am not a gear "expert" by any stretch.




This source says "crossed or skewed configuration", so I guess crossed helical gears fall into the "skewed" configuration ?

https://en.wikipedia.org/wiki/Gear

.

 

[GreenTwin]

The sewing machine gears seem to be plastic, and I don't see any two of the same diameter, but different angles on the teeth.

For a 2:1 speed ratio, with gears the same diameter, crossed configuration, the teeth angles need to be approximately (but not exactly) 30 and 60 degrees.

These gears can be made using a 3D program, as JasonB illustrates.
I was able to get an approximate gear made in Solidworks, but the model must be exact, not approximate, for the gears to mesh and work correctly.

I printed JasonB's helical gears, and they do have exactly the correct tooth form and angle; they do mesh perfectly; they are the same diameter, and they give a 2:1 speed ratio.

.

 

[Jasonb]

Pats gears are also quite large for "model size"

Pat rather than moulding a hollow gear and then trying to cut it up what about the other method I mentioned of making the pattern in parts rather than the mould. Would want a bit of analysis in CAD but that is easy enough in F360 to check the draft but if you printed it in 5 or 7 parts so the middle can be withdrawn vertically and then each remaining segment inwards then up and out that would save destroying the pattern by cutting. It basically collapses in on itself once the middle has been removed.