JB's 3D CAD & Printing bench

[John Baker]

Graham, Andy - many thanks for your combined input. Absolutely fascinating. As I've said before, this is very much a learning curve for me, this being my first go, so this information is invaluable. It hadn't crossed my mind about putting smaller and larger items next to each other. Interestingly, it's not the finer components that are the problem - more the larger ones. Maybe, as you suggest, I need to take more of a multi-sprue approach, rather that having just one central sprue. Or maybe I need more feeds into the larger components to allow for drawing in more metal as it shrinks. The benefit of 3D printing the masters is that they can be changed easily. Having said all this, almost all of the castings are really quite good - much better than in some kits I've seen. Just the safety valve columns and the smokebox door to sort out. The lettering on the sprues themselves would be nice to have perfect, but obviously has no bearing on the model itself. Keep the advice coming everyone - I'm all ears (eyes?).

 

[AndyH]

No problem John,
Graham did all the hard work.
We use a specific sprue system for our flasks and machine, but as you can see here http://www.hewitt-impex.co.uk/wp-content/uploads/2017/03/Neusprues-leaflet.pdf there are some ideas you could incorporate into your own printed sprues. We’ve done this on certain items we know will require a larger volume of metal and to make room in the flask, and use the proprietary sprue as the base.
As before, getting to know your caster’s equipment and process is part of the challenge, then you can build the requirements into the sprue design. Then you can even save more space and metal!

 

[GrahamMc]

Interestingly, it's not the finer components that are the problem - more the larger ones.

Your caster is casting into a flask held at a high temperature so the metal reaches and fills all the fine detail before solidifying. The problem is that this is too high a temperature to get good results for the heavier items, they'll have problems with porosity.
Components of less weight and cross section need to be on different sprues so they can be cast in different flasks cast at different temperatures, as described in my earlier post.
There's nothing the caster can do other than what he has done.
Have a look at the pictures of the wax trees shown in the links on both mine and Andy's posts and see how the items are arranged around the central main sprue. You'll see that the items are not large numbers of components off a largish sprue that is then attached to the main central sprue but individual items attached directly to the main central sprue.
That's no good for what you want, the components in the pictures are larger than most of yours, but what you can do is arrange say, five or six small components in a sort of flat fan with each item on a short sprue linking at the base of the fan. The caster can then arrange these fans around the central sprue. It's a more efficient way of filling the flask and so should be cheaper.
Larger items like the springs benefit from extra sprues but arrange them carefully, for example with the springs have a sprue to the centre of the spring but then one either side going from this central sprue to the end of the springs. If there's more than one spring on the fan you can also take them back to the central sprue of the fan. Just have a think about what will allow the metal to get to the end of the component quickly.

The principle is simple, thick to thin, to thinner by the shortest route.

Not always possible with the larger components but that's OK as long as the rest of the feeds are arranged on that principle.
What can be impossible to cast well are fine details, like thin handles, on heavier items. It's not possible to get the flask temperature right for both the thin handle and the body of the component so they need to be made separately and joined later (if possible).
This sort of work can be a bit of a pain for casters. Normally if something goes wrong the caster just creates another wax from the rubber mould and does it again, at little expense to the caster and the customer is none the wiser. With 3D printed patterns the cost of the patterns can be relatively high for the customer and the caster only gets one go at it. If it goes wrong you get an unhappy customer so the caster might want to cast at a higher temperature than he/she would like to make sure the mould fills.

 

[John Baker]

Apologies in advance - a long, rambly post coming up! But potentially interesting to some

It's been a while since my last post on here - I've been busy doing test builds for the Manning Wardle Old class I project which is now nearing completion. I'm on my 3rd and final test build, with just the wheels and inside motion to finalise. Also been sidetracked by a foray into radio control, and a making a start on a display diorama so I've got somewhere nice to photograph my builds.

Previously in this thread, I have shown how I 3D printed the castings for the loco in a castable resin, and then had them cast in brass by a company. Since then, I took on board the advice of Graham and Andy above, plus that of others and redesigned the sprues a couple of times. Problem was, this was getting expensive, as a full set of castings was costing upwards of £50+ each time. I'd been following a few YouTube channels of people who do their own castings (jewellery rather than modelling, but the principal is the same), and wondered if I could vacuum cast my own models. Professional casting equipment can cost many thousands of pounds, but there are hobbyist machines that would do for my purposes, although my budget was pitifully small, so some compromises had to be made as can be seen below.

The basics needed were:

• 3D printer (which I already had)
• A kiln for burning out the 3D printed waxes
• A small, electric metal melting furnace
• A vacuum casting machine + vacuum pump (a cheap 2.5cfm single stage from eBay several years ago when I first floated this idea in my head!)

The latter 2 were combined in one machine, purchased from Facebook market place and advertised as not working. For £50, I thought it was worth a punt - the furnace part of the machine was working fine, but the vacuum half didn't pull a vacuum when attached to a pump. A quick bit of tinkering found the problem to be a simple fix, so a bit of a bargain given these machines cost £500-600 new! It has an integral metal melting furnace, a vacuum investment table for degassing investment plaster before & after pouring, and a vacuum chamber to draw the molten metal into the tiny nooks and crannies of the models.



Next up was the kiln - again from Facebook market place and not working as the controller wasn't working. No problem, as I had bought a programmable controller as I needed to set specific burnout schedule, so as long as the elements were fine - which thankfully they were - this would be perfect. It must be a good 30 years old so looks a bit brow beaten, but it's doing the job!




So onto the castings themselves. I've experimented with various resins, burnout schedules, temperatures of metal and flask, and finally think I'm getting close - not fully perfect yet, but much better than when I started. There are so many factors that are at play that affect the quality of the castings, one has to take quite a scientific approach to changing the variables each time in search of improvement.
Previously, @GrahamMc has written about the need for different temperatures depending on the size of the pieces being cast, and he's spot on - the castings pictured below are from 2 different trees - the tree with the larger pieces on is pictured below (the design of which isn't ideal, but it seemed to work OK), with a metal temperature of 960 C, and a flask temp of 500 C. The smaller pieces were attached to a separate tree in a different flask (forgot to photograph these), and cast with a metal temperature of 1080 C and a flask temperature of 630 C.

Apologies for the really poor photos, taken on my phone in pretty poor light. Cruel close-ups etc etc...... they look pretty decent in the flesh. Some of these parts are only a couple of mm, so are magnified quite a bit! Also, they are either snipped or cut straight off the tree, with no clean-up, so they are still raw where removed from the sprue. All in all, they're not too bad for a home-brew attempt!






You can see some ash deposits on the buffer head at the bottom, which I think indicates I need to get a better airflow into the kiln.







Including the spigot on the back, these blowdown valves are only 2.5mm long, so it shows that if you can print the required level of detail, you can cast it.



Next steps:

• refine casting temperatures further - I'm led to believe that a difference of just 10-15 C can affect the surface finish.
• Redesign the buffer bodies - The wall thickness is really thin around the 'neck', leaving small holes where the metal either hasn't filled fully, or has been drawn into thicker parts as they cool.
• Change where I place certain parts on the main sprue to aid surface finish
• Print the parts at a layer height of 0.02mm (rather than 0.05mm as these were) to reduce build layer lines in some parts.

I'll post on here again once I've got a reliable, repeatable system!
John