Prusa XL 3D Printer

Home Model Engine Machinist Forum

Help Support Home Model Engine Machinist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Here is a comparison of the No.06 flywheel with the No.02 Cretors flywheels.

There is significantly more mass in the No.06 flywheel for the same diameter, mainly due to a thicker rim.

I have what I call a BASC (big-ass steel crucible) just for those occasions when you need a large aluminum pour, and are not worried about steel/aluminum contamination.
I have never seen any problems with pouring aluminum from a steel crucible, but the recommend against it.
It may be due to the steel degrading over time, which it will do, but it is a slow process with aluminum.

Large Steel Crucible No.02:
9 3/8” inside diameter, 12” tall, volume = 826, max. AL capacity (brim-full) = 80 lbs.
Usable pour range approx. 56 lbs. of aluminum.

Aluminum 356 = 0.0968212 lb/cu-in
Ratio of weight of cast iron to AL = 2.68
Ratio of weight of AL to cast iron = 0.3727

So I can pour 50-60 lbs of aluminum if necessary in a single pour.

Picture of BASC below.


Super-Salamander A-shape

Size Top OD Bottom OD Height Brass Capacity CI Capacity AL Capacity
inches inches inches pounds pounds pounds

A0.5 2.625 1.875 3.125 2.2 1.98 0.7
A3 4.125 2.75 5.0 8.2 7.38 2.7
A5 4.875 3.375 6.0 15 13.5 4.9
A6 5.11 3.74 6.49 20 18 6.4
A10 6.29 4.33 7.87 40 36 11 (weight = 5.8 lbs)
A16 7.24 5.11 9.13 51 46 16
A20 7.75 5.70 10.23 66 60 21 (weight = 12.0 lbs)
B20 7.75 5.70 10.5 74 67 24 (weight = 16.6 lbs)
A25 8.26 6.10 11.02 79 71 26
A30 9.13 6.29 11.41 95 86 31
A40 9.12 6.25 12.5 120 108 40 (weight = 20 lbs)

The density of aluminum is about 0.098 lb/in cubed.

Brass is between 0.3034 and 0.3143 lb/in cubed.

Cast iron is between 0.2456 and 0.2817 lb/in cubed.

A10 - Morgan Salamander Super
8" tall
6.375" diameter at top
6.4 lbs
Brimful CI capacity: 31 lbs
Approx. Usable CI capacity: 15-20 lbs

A16 - Morgan Salamander Super
9.25" tall
7.25" diameter at top
9.6 lbs
Brimful CI capacity: 46 lbs
Approx. Usable CI capacity: 19-26 lbs

A20 - Morgan Salamander-Super
10.5" tall
8" diameter at top
12.2 lbs
Brimful CI capacity: 59 lbs
Approx. Usable CI capacity: 25-33 lbs

B30 - Morgan Salamander Super
11.5" tall
8.5" diameter at top
xx lbs
Brimful CI capacity: 94 lbs
Approx. Usable CI capacity: 45-61 lbs

"A" means an A-shaped crucible, and "B" means a bilge-shaped crucible.
The bilge crucibles have more capacity for a given size than the A-shaped, but the A-shaped are easier to source.

My largest Morgan "Salamander Super" crucible is a B30, and that is the largest that my furnace will easily accommodate.

I don't plan on using anything larger than a B30, but I know folks who use #70's with iron.

The usable cast iron capacities listed above are pretty conservative.
It all depends on how good you are with pouring from a crucible.
The fuller the crucible, the less you are able to tilt it down towards the spure before you begin pouring.
A nearly full crucible generally will have a lot of spillage when trying to pour with it.


The brimfull capacities listed above, if converted to 365 aluminum, would be about 1/3 the weight listed for cast iron (CI).

The beauty of a steel crucible is that the walls are much thinner than a clay-graphite crucible.
Comparing the steel crucible shown above with a Morgan B30, for approximately the same size crucible, you get over twice the capacity in 356 aluminum by using the steel crucible.

For gray iron, a clay-graphite ferrous-metal-rated crucible is required.

Last edited:
I found it easier and quicker to scape the surface of my pLA prints, then a light sanding, then a few coats of primer/filler, another light sanding, then fill any imperfections. Some PLA sands better than others. The primer/filler dries in a few hours.

To keep my resin printer warm, very warm in fact, at 80c. I use the cardboard box the wash and cure machine came in. I also drape a few heavy towells over the box, it has no front, and a 40w proper light bulb as a heater. My Ender3 is in the spare bedroom I use as a craft room. It's fairly draught free-ish. The flywheel is looking really good.
P.S. I print PLA at .2 layer height with the bed at 60 degrees. Not sure if that's F or C.
For scaping I use cheap stainless steel "dentist" tools. They will take an edge for a while. I also use a metal nail file. The curved end, for getting gunk from under what's left of your nails, is also good for curved surfaces. I would be interested to hear of your experience with the body filler on an fresh surface.
I will use the dog as a test for the two body fillers, since the dog has horizontal and vertical surfaces, and differing surface finished depending on the exact location on the print.

I will use the vertical surface and the horizontal surface filler on both types of surfaces, and see if the horizontal filler sags on the vertical surfaces.

I am wondering if the putty would flow with a carefully applied heat gun.
Aside from the fumes which should be avoided, I assume the filler is also flammable, so I will have to be cautious.

No sense putting the fillers on the flywheel without knowing exactly what they will do.
I have not used either type yet.

You might want to checkout the many videos on YouTube of people prepping 3D printed pieces for cosplay helmets, armor and such. It seems to involve filler and a lot of sanding...

Another thing to research is printing with ABS and then using an acetone vapor bath. Perhaps this is less work.
You might want to checkout the many videos on YouTube of people prepping 3D printed pieces for cosplay helmets, armor and such. It seems to involve filler and a lot of sanding...

Another thing to research is printing with ABS and then using an acetone vapor bath. Perhaps this is less work.

Thanks, I will check that out.

I hope to find a solution that is relatively easy and fast.

Just a follow on to an earlier post mentioning nozzle size and layer heights. It's an area with some misconceptions and things that are not immediately obvious until someone tells you about them. Then you say Duh, who didn't know that, it's obvious! I don't use a 0.6mm nozzle often, my focus is on detail for scale model use so 0.4 mm is about as large as I go on the nozzle front.

Normal layer height ranges for a given nozzle diameter are 25 to 75 percent of nozzle diameter. Layer height ONLY effects Z axis resolution, it has a very minimal effect on X/Y resolution. Finer layer heights are wonderful for reducing layer lines, but may or may not improve fine detail. I can print at 0.08 mm layer heights all day long at 120+ mm/sec on my Ratrig. You would initially expect that these prints would come close to those printed with 0.05 mm layer heights in my resin printer. For some prints this plays out, for detailed prints it does not. Quite often other than shallow upward facing curved surfaces there is almost no discernible difference between 0.08 mm layer height prints and those done at 0.12 mm. Either way, for a good model, some filling or sanding will be required. Less with the finer layer height, but it won't get the post processing down to just spraying primer and starting to paint in most cases. Your finest detail is also a function of line width

SIDE NOTE: alternatives to sanding and filling:

Printing with ABS or ASA opens the door to smoothing a print surface with acetone. As the specific applications being talked about in this thread don't have super fine detail the smoothing / loss of very fine detail probably is not a consideration. I tried PWB (polysmooth) which uses alcohol. Was not really impressed with the results. Using a filling primer may address layer lines well enough for some applications that minimal sanding is needed. I have tried XTC-3D epoxy made for smoothing prints. It has to be mixed very, very, did I mention very? well. As in 2 minutes of aggressive mixing in small mixing cup. It then fires off within another few minutes. So it's got a very short pot life, you must move fast. It does do a good job if you can get your print coated evenly before it starts to drag. Results were somewhat better than the same prints coated with 30 minute epoxy thinned around 20% with alcohol and brushed on. Some prints even came out almost glass smooth.


Some printers will go down to 0.05 mm layer heights with a 0.4mm nozzle, although many folks who want to print at this height also drop down to a 0.2, 0.25. or 0.3 mm nozzle. Smaller nozzles like to clog, and naturally they have to run slower as the volumetric capacity is reduced by the smaller orifice. Trying to run one at full speed usually just results in the extruder shaving plastic off the filament and skipping steps. Don't even think about trying anything other than normal filaments. Silks, carbon fiber filled, and glitter stuff will clog a small nozzle in a moment. White filaments are a problem due to the high solids needed to make the stuff look white. Heck, white filament is usually hard to get printing as well as other colors in normal boring configurations and profiles.

Line width and layer height are also related to print strength. Bigger lines with good layer and line adhesion tend to produce stronger prints. Other stuff comes to the party, but usually the other factors come along for the ride unless you really mess around in your slicer and get it really wrong.

You can't print a line width smaller than your nozzle diameter. Typically the printed line width with squish is 10% larger than the nozzle diameter. If you need sharp small detail, a 0.6 mm nozzle will round off the fine details more than a 0.4 mm nozzle. In many cases it does not matter, but it's just one of those things to keep in the back of your mind. If you want to print faster with a 0.4 mm nozzle, you need a high flow hot end or something similar to a CHT nozzle to allow the plastic to melt faster. The trick is keeping the plastic in the melt zone longer, not making the melt zone hotter. Raising the temperature in an effort to improve volumetric capacity works a small bit, then it all goes in the toilet from filament burning, sagging, clogging, melted looking prints, poor cooling, supports that can't be removed, basically it's just a mess. There is a certain amount of time required to fully melt filament to the center, can't cheat reality by more than a few percent :) CHT and the new REVO High Flow nozzles split the filament into multiple channels within the nozzle as it melts, then combines them back at the business end. Quite clever, in effect they make the filament into three smaller strands so it melts faster. Others like Volcano use a longer physical melt zone / heat block to get a higher possible flow rate.

In the case of this thread, Prusa's newer nozzles and hot end have a longer melt zone from what I've seen. I don't know if they are using the CHT / Revo high flow tech in their new hot ends / nozzles or just doing the volcano style long melt zone. Or both. Haven't had a chance to fondle the new nozzles in the flesh yet.

If you set the outer wall line width to around 105% of the nozzle diameter your outer wall will bulge less as it is only adhering to the layer next to it on one side. This is based on using a default print line width of 110% of nozzle diameter.

A larger nozzle can not only lay down a taller and wider line, it can flow plastic faster. Printing at 0.2mm layer heights, a fast printer with a good hot end can print a LOT faster with a 0.6 mm conventional nozzle than with a 0.4 mm nozzle as the print speed will not be limited to 150 mm/sec or so as with a conventional 0.4 mm nozzle. A 0.6 mm nozzle in a stock Ender 3 isn't going to up the print speed greatly, it still has the same melt zone size. If you can't melt it, you can't squirt it. BUT: A stock Ender class machine doing 0.3 or 0.4 mm layer heights will still produce a print a lot quicker than the same machine with a 0.2 mm layer heights, even at the same "printing speed". Most testing shows that normal boring consumer printers can melt around 8 to 10 mm^3 of filament per second, while a 0.4 mm nozzle seldom can flow more than 5 to 6 mm^3 per second, so some tweaking for larger nozzles is worthwhile for optimal speed.

First layer hints: I usually print lines at 0.44mm width, and first layers at 0.46 to 0.48 line width with an increased height to get good squish and inter line bonding in the critical first layer. With a 0.4mm nozzle on a model being printed at 0.2mm layer heights my first layer is at a height of 0.25mm. I usually print the first layer 5 degrees hotter than the rest of the model, and seldom print first layers any faster than 30mm/sec on most machines. Usually 20 mm/sec. The extra time spent printing an excellent first layer is far less than the lost time and filament when a print or critical tall support breaks loose 10+ hours into the run because the first layer was weak. My RatRig does good first layers at 50mm/sec so it's likely Prusa MK4 / XL can do so or faster as these are machines with better than "Creality grade" print heads and structure. It's still printing first layers at around 1/3 of normal print speed, seems pretty typical.

That ended up longer than I'd planned, hope it's of some use and not just boring.
Thanks much for the helpful information.

I will have to read it several times to soak it all in.
I am definitely very interested in such topics.

I am still learning about the 3D printing world, and don't really know that much about the technology.

Thanks for the heads up about white filament Stan, I didn't know that.
I am using Copymaster 3D premium filament at the moment. It scapes and sands to perfection. It's made in Europe, so it might not be available in the US. Try
You can't print a line width smaller than your nozzle diameter. Typically the printed line width with squish is 10% larger than the nozzle diameter. If you need sharp small detail, a 0.6 mm nozzle will round off the fine details more than a 0.4 mm nozzle.
This hasn't been entirely true since the new Arachne engine was released in the Cura 5.x series last year. It's been incorporated into the Prusa Slicer (ala open source) as well.
Slightly more click-baty:
That's the reason for the push to larger nozzles as your ultimate print speeds are usually limited by how much filament you can melt and then quickly cool again. If you can achieve the same detail with a larger nozzle then that's better yet.

Obviously there are limits to this. I was playing around printing 48 DP gears last year and I still noticed a nice improvement in detail from going from a .4 to a .3 nozzle while viewing under the scope.
With pattern making in metal and wood, I have tried a lot of filler materials just like Stan has.

I have tried sheetrock compound, spackling compound, wood puddy, bondo, etc.

The trick with filler is to find something that adheres to the surface well, is easy to spread, is not too hard to sand, and does not load up the sandpaper.

As Stan mentions, some of the two part filler/hardner materials take a lot of mixing, and then they set rather quickly, or if you back off on the hardener, sometimes they set too slowly, or never set hard enough (worst case).

The water-based fillers can also set too fast.

The Durham's wood putty is good material, but like the container says, it dries "rock hard", and so when you try to sand it out, you end up damaging the pattern.

I have found bondo also to be way too hard for easy pattern filling.

The last few wood patterns that I have made were filled with a water-based mix-on-demand wall filler, which is slightly on the hard side, but not excessively hard, and is workable to sand without too much force or effort.
Once I get the pattern filled, I coat it with shellac to give it a hard and more durable surface finish.

One approach that I considered with 3D printed patterns is to just use the patterns without any fill, and then buff the lines out of the casting with a sanding sponge and/or sanding disk.
The problem with this approach is that some of the lines in a 3D print are quite deep, and a good molding sand will mimic the surface exactly.

A good molding sand will mimic a single strand of hair in the casting itself (or even a fingerprint), and the defect will be clearly visible in the casting.

So I am contemplating filling a 3D pattern, but not sanding most of it, and then buffing out the casting.
This would prevent having to try and buff out deep lines, and would prevent the casting from getting noticeably thinner due to the buffing process.

A 2" diameter sanding sponge is a marvelous thing when used in a tool and die grinder, and it works wonders with buffing out castings.
The sponges come in three different grades, which are coarse, medium, and fine.

For areas that would be difficult to reach with a 2" sanding sponge, the filler would have to be sanded prior to casting the part.

And so I guess it begs the question, can the skim coat auto body filler be buffed out with a sanding sponge? ie: buff out the filler on the 3D printed pattern?
I will have to try the sanding sponge on skim coat.

I tried the fine sanding sponge on the surface of a PLA print, but it just melted the surface.

I have seen acetone used to smooth the surface of some 3D printed materials, but I have mixed feelings about the appearance the resulting print.

It is similar to going to a car show, and some cars have 20 (+) coats of transparent laquer, and so they look super smooth, but they often look so smooth and glossy as to appear unnatural.

Old steam engines generally had a painted surface, but not really a super smooth or super glossy finish.
Generally the imprint from the sand in the casting is still visible after the cast part is painted, and this gives the old engines a unique look, and a unique surface finish/texture.

The acetone smoothed printed look like they got too hot and melted (I guess the surface does indeed melt a bit), and if you are looking for a super smooth finish, then this is the way to go.

I also considered the powdercoated paint method, ie: spray on a powdercoat, and then heat the surface to melt and smooth the coating.
A PLA print won't take the heat required for powdercoat paint though.

And I have considered using a heat gun to smooth auto body skim coat before it sets, but one would have to be cautious, since I am sure skim coat is flammable, and you would not want to inadvertently find the auto-combustion point of the material as you were heating it.

I recall videos of the auto manufacturers using lead to fill auto bodies, and that is a technique that may be useful, ie: providing a little heat from a heat gun to cause the filler to run slightly and level out, but not overheating and melting or distorting the print.

I am going to take the 3D printed dog out to the shop, and see if I can find my chemical respirators.

I really need to find out what the skim coat and seal coat will do on horizontal and vertical surfaces.

If the skim coat and seal coat are too thick, it would be interesting to see if they could be thinned.
I really think sags could be easily sanded out, and better to have a few sags and have a smooth fill, than to have putty that is too thick with an uneven surface.

Pat you sasy Bondo is "too hard" what have you been using? There are many grades depending on thickness of application, rathe of drying, ease of sanding, etc Look for some of the "supersoft " ones for ease of sanding.

Clogging of paper is often down to poor mixing leaving soft pockets that stick in teh paper or just not allowing it to set enough.

I doubt your idea of getting it to flow with a heat gun will work as most of these two part fillers use a thermoreaction to set them off so adding more heat will speed things up. Just look at polyuester resins most will have % amounts of hardener for a given working enviroment temperature. You need to take the workshop temp into account when waiting for things to set or judging working times.

I use car body fillers quite a lot for work on wood that will be painted and keep a few depending on the job in hand, a rapid one is good for a few screw holes that need filling on site but an easier to sand one better suits repairs to old mouldings

Generally the surface of a model needs to be smoother than the original, not easy to scale sand when working on a 1/12th scale model but not such an issue on a 1:1 replica. Also the smoother the pattern the easier it will pull from the mold which is usually more important that the texture of the finished surface

As with anything take your time and don't rush bondo and the pattern are plastic so too fast sanding will just heat things up and melt or clog the paper if you have a cordless drill use the disc in that at slow speed rather than a high speed die grinder or put it in the drill press at slow speed and bring the pattern upto it. Seemed you rushed into printing this pattern without doing tests to see what surface finishes you got and made it worse by speeding up the second half so at least take your time with teh surface prep. I can do my CNC patterns far quicker but the surface finish suffers so I much prefer a finer cut that really does not need any hand work and can be cast straight off the CNC. It's faster in the long run and I can be doing something else while it cuts, you need to be there to fill and sand
Last edited:
Here are the results of the first try with the skim coat and the seal coat.

Jason is right, the thermoreaction at room temperature at the recommended amount of hardener is fast.

I started with the skim coat, which is the more dense product on the left.
I mixed in the hardener, smeared an area on the dog model, and then it set.
LOL, a little hardener goes a long way.

Next I tried less hardener with the seal coat, which is the more fluid product, and it went on better.
It did not seem to flow into the lines of the print though, which may or may not be a good thing.

About 15 minutes later I was able to start sanding, and the products do sand easily, and are not too hard.
They don't load the sandpaper either.
I could not find my plastic applicators, so I resorted to smearing on the seal coat, in various thicknesses.
Very crude method, but best I could do at the moment.

The sanding goes pretty fast.
One has to choose between sanding down to the top of the original PLA material, or just stopping just short of the PLA.
Sanding down to the PLA opens up the lines in the print.

Then I tried the sanding sponge in the tool and die grinder, but it was too agressive.

I put the sanding sponge in a variable speed tool, and had more control, but I think in the end it would be easier to to hand sand the pattern.

There are various approaches on ytube, with some thinning the putty down with acetone, and brushing it on.
Other spray on thinned putty, but this will clog a sprayer when the putty sets.

Some spray on multiple layers of "high-fill" spray primer.
The grooves in my 3D prints are too deep to be filled by high fill primer.

I think it will be sort of like applying mud to sheetrock.
Less is better, and better to use several thin coats than one heavy coat.

Using a plastic applicator would probably help too, especially on flat surfaces, but perhaps not much use on curved surfaces.

I think the flywheel pattern will require an initial coat of seal filler (the thinner product), then a bit of sanding, and then a high fill spray primer, or perhaps shellac which is a pretty good filler that dries very fast.

The water based filler I have been using is sheetrock repair compound by DAP called "Fastpatch 30".
The problem with water-based fillers is that they dry slowly compared to polyester products.

The beauty of the polyester seal coat material is that you can apply it and be sanding in 15 minutes, so that really speeds up the filling process.

I would anticipate the seal coat can be used to fill and sand the flywheel pattern in about an hour.
Sanding is easy with the fill coat material, and that is a real plus.

A 1 hour fill and sand time for a pattern would be an ok compromise for me as far as print time/print speed, and surface finish.


I imagine the seal coat and skim coat would scrape pretty well with a sharp blade, as someone mentioned, but I think I would have better control on curved surfaces with sandpaper.

I think a good technique will be too apply a thin layer of seal coat that does not quite cover everything, sand that, and then apply another thin coat that does cover everything.

Two or three thin layers will be easier to manage than one thick layer.

I would approach it much like I do a flat surface. Sand first to remove the worst of the high spps. Then fill and sand again, repeat as required.

0same within build filler primer, sand first and also sand between coats so you only build it up in the dips and don't build up the high spots.

Watch any watermarked fillers that Re not acrylic and are mixed with water like powdered drywall filler. Goes on easy and sands easy but is affected by damp so not good for patterns for future use in a bank