New castings

[Lawrence@NZFinescale]

I thought I'd post these as a new thread as it's a bit off topic from the layout saga.

I just got some castings in that represent probably some of the best I've done. Such things are generally acceptably, if not totally, reliable.

These are all 1:64.

The first casting is an early diesel locomotive cowcatcher done as a commission.  The second is combined cylinder/slidebar/motion bracket for the Sharp Stewart B class (and We class conversion). The sprue also has the eccentric rod on it (seen from the rear with a strongback added which will be invisible in the finished model). Rods on the cowcatcher are drawn and cast 0.6mm.  They should really be 0.5mm, but while that's possible, it can be problematic on longer rods so I was conservative here.

My procedure for these castings is:

• Draw the part up in CAD.  Generally I draw the casting sprue as well.  This will double as a print support, so there's a bit of juggling to get something that serves both functions.  Additional support is usually added in the slicer software (removed prior to casting). Most of the time I arrange things as half sprues as this makes printing a lot easier.  The L/R halves are welded together with casting wax prior to casting.
• Print in a 'castable' resin.  These are done at 35u resolution 35u layers.  Close inspection shows some layer lines, but after sand blasting and clean up these won't be significant.  More recent printers could probably do a bit better.  I'm using stock of a discontinued resin, and again newer resins may improve the result a little.
• The MOST important ingredient to success is avoiding damage to investment features from inrushing metal.  This means a) avoiding the creation of vulnerable investment features, and b) minimising sprue cross section to reduce metal velocity.  I'm no casting expert, but the approach is successful even if my understanding of the underlying process is incomplete.  Vulnerable features are thin fingers or fine edges of investment that are inherently weak and break away when the metal hits them.  Can be hard to get one's head around as they are the negative to the model.  Broken investment leads to missing details, porosity and poor surface to the castings.
• I use a contract caster to do the actual casting into silicon bronze. They don't always look quite this pretty as sometimes there's a bit of scale on the surface.  This cleans up on sandblasting, so isn't a problem.

This is not a particularly difficult process, but it does need experience and access to the machines, materials and contractors required.  For those with CAD models I'm happy to chat through converting them into metal.

[finescalerr]

Those raw castings are as close to perfect as is possible. Your "instructions" are invaluable. Most satisfactory. -- Russ

[Ray Dunakin]

Very nice. The layer lines are only barely visible.

[Bernhard]

I am impressed by how small details can be cast and how sharply they are depicted. Well done!

Bernhard

[Hauk]

I am very sorry, but my best efforts to find any faults with these castings failed miserably.
Pure perfection, in other words.

[Hauk]

My procedure for these castings is:

• Draw the part up in CAD.  Generally I draw the casting sprue as well.  This will double as a print support, so there's a bit of juggling to get something that serves both functions.  Additional support is usually added in the slicer software (removed prior to casting). Most of the time I arrange things as half sprues as this makes printing a lot easier.  The L/R halves are welded together with casting wax prior to casting.
• Print in a 'castable' resin.  These are done at 35u resolution 35u layers.  Close inspection shows some layer lines, but after sand blasting and clean up these won't be significant.  More recent printers could probably do a bit better.  I'm using stock of a discontinued resin, and again newer resins may improve the result a little.
• The MOST important ingredient to success is avoiding damage to investment features from inrushing metal.  This means a) avoiding the creation of vulnerable investment features, and b) minimising sprue cross section to reduce metal velocity.  I'm no casting expert, but the approach is successful even if my understanding of the underlying process is incomplete.  Vulnerable features are thin fingers or fine edges of investment that are inherently weak and break away when the metal hits them.  Can be hard to get one's head around as they are the negative to the model.  Broken investment leads to missing details, porosity and poor surface to the castings.
• I use a contract caster to do the actual casting into silicon bronze. They don't always look quite this pretty as sometimes there's a bit of scale on the surface.  This cleans up on sandblasting, so isn't a problem.

This is not a particularly difficult process, but it does need experience and access to the machines, materials and contractors required.  For those with CAD models I'm happy to chat through converting them into metal.

It seems that you have optimized your production chain pretty well, as your castings are some of the best I have ever seen. I really appreciate that you are willing to discuss how you achived these excellent results!

A couple of questions: 

-Do you always print the master for the metal casting  in a castable resin, or do you for larger runs print a master for a rubber mould and then shoot waxes the conventional way for the actual metal castings?

-Could you tell us a little about the cost involved? For instance, I am curious about how small batches that are economical to make. How small series are your contract contract caster willing to make?

-Were should one look for a local caster? Is the jeweller industry the place to start?

If I were a rich man, I would have made castings of all my 3D-printed parts, but I think I will have to save it for the very special projects like engines.

[Lawrence@NZFinescale]

A couple of questions: 

-Do you always print the master for the metal casting  in a castable resin, or do you for larger runs print a master for a rubber mould and then shoot waxes the conventional way for the actual metal castings?

-Could you tell us a little about the cost involved? For instance, I am curious about how small batches that are economical to make. How small series are your contract contract caster willing to make?

-Were should one look for a local caster? Is the jeweller industry the place to start?

Thanks Hauk

I have certainly had moulds done in the past and still use a few.  Printing and preparing resin parts is time-consuming and a major cost for a commercial operation, but no cost at all for a hobbyist.  But moulds cost too, so for short runs the print route is probably cheaper (and considerably more flexible).

Of more relevance to me is that many of these castings would be unmouldable, so the scope of what I can do is greatly expanded. Because I'm avoiding moulding costs it is more viable to have a couple of goes at optimising patterns and makes one offs viable.  You can pack more into printed sprues too. You also avoid the issues associated with shrinking wax (as it is the wax moulding step that shrinkage mainly occurs). Never any part lines in a cast from prints either - not that that is a major problem with moulds.

Appropriate casters are in the jewellery trade.  I've used the same supplier since day 1 (over 30 years), but I know others who have used a competitor.  It's not unusual for jewellers to want a single casting, but I tend to do batches of at least a dozen or so (primarily to economise shipping each way). My understanding is that these outfits will generally do some form of bronze as this is used for cheap plated jewellery. I'm a thousand km from my supplier, so no need for someone local.

For the cylinder casting:

• Printing costs. 3-4 hours on the printer and around 4ml of resin including support.  Resin is on the expensive side $200/500ml. Of course I print a lot more than one casting per run.
• Some time and materials to clean up prints and to weld sprues (wax + soldering iron)
• My caster is in another city so I incur a courier charge each way of around $8
• There's a unit charge that depends on casting volume.  $8 for the cylinders (which is large for what I do, $3.50 is more typical)
• I'm also charged on metal weight @$0.40/gram.  The cylinders are 19g, so $7.60 (the cowcatchers are only 7g)

So the cylinder casting cost me around $16-$18, the cowcatcher well under half that.

For reference, last time I checked mould making was $35 (but that was some years ago and size dependent).

I should add that I don't get charged for failed castings even if it is really my fault for attempting the impossible, or presenting something in a poor configuration. Generally QC is on me so it helps to be doing regular work so returns for credit are routine.

BUT my own experience is that the learning curve to get to great castings is the largest single cost.  Given that printing is no longer cutting edge and casters will be used to getting resin prints that's probably easier now than it was for me.

All costs in NZD excl local taxes (They look much better in USD).

I'm more than happy to do castings for others (it's part of what I do).  I prefer to receive CAD files as I need to arrange the sprues.  Obviously there is labour to prepare and print plus a margin on casting costs.  Shipping is significant for one or two castings as well.  If you cannot justify the cost and time of setting up yourself it's an easy option. If you need a short run of something it's typically quite attractive.

[Stuart]

Beautiful, crisp work. It's amazing the detail and sharpness of your pieces.

[Lawrence@NZFinescale]

A while back I took a bit of gentle ribbing about missing detail (leaf spring separation) in my castings.

Well technology moves on and I'm now using a different printer/resin combination that is producing better results.

Attached are all 1:64.  The scale in the generator pic is 0.5mm. 

At normal viewing, the benefits of the newer processes are not that apparent as castings were pretty good before.  However under magnification there is a significant improvement.

[finescalerr]

Satisfactory. -- Russ

[Stuart]

Extremely impressive.

Stuart

[Ray Dunakin]

Sharp enough to read the lettering on the generator!

[Bill Gill]

Like Staurt said: Extremely impressive. And you continue to experiment and seek improvement, equally impressive.