Sunday, June 12, 2022

Casting aluminium wheels from melted down aluminium wheels

 This was a real pita to mold. It took several attempts, eventually I added a little more water to the sand, and pushed 3" nails into the centers of the high spots to hold things together.


Here you can see one of the many failures. It was fine until I flipped it over to close the mold. By now I've learned to flip the cope above the sand box, rather than above the drag, a lot less cleaning up to do.



The nails helped a lot, but so did adding just a little more water.
I used steel pins to align the two part of the mold, a bit of trial and error on a small sample helped me to figure out the right sized holes. 3d printed pins are ok, but the tend to catch more on opening the mold.


The pour went well, it was a lot of metal probably the second biggest pour I've done.

It's always fun to see what comes out, this was done the next day when it was cold.

Not a bad result, they will need some cleaning up, and I'll need to drill out the hole for the axle.


You can see how much shrinkage there was in the riser, without such a big riser direct attached to the part, it would have had a similar sized shrinkage defect on the hub.











Monday, June 6, 2022

3rd time lucky

 This was cast with a length of steel rod embedded in it. I used a smaller length to make up the drag.

Then once I flipped it over, I pulled the short steel piece using a magnet, and replaced it with one that was slightly longer.
This let me locate the second part of the pattern.

I added a fairly large riser between the sprue and the piece, and it did the job, no shrinkage.
It needs a little cleaning up. The circle over the steel pin is not shrinkage, it is smooth and level, but you certainly can see that the metal solidified a little differently there. If I was doing it again, I'd have a shorter pin, so it didn't come as close to the surface.








Sunday, June 5, 2022

Casting the gingery lathe Face plate

 I cast the face plate.

It was cast on a steel mandrel, which I had to drill out quite a bit before I could drive it out with a hammer and punch. 

After a few attempts at molding this, where the pattern lifted the sand in the slots, I tried just pressing 2 panel pins in to each of the slots and that worked.
That gave the slots just enough support to stay in place as I rapped and removed the pattern.

And yes this photo was staged after the event, I was in a hurry to get set up and cast, as the forecast was only good for the morning. The steel mandrel is missing from the pattern here.

I still have a fair bit of work to go before I am ready to machine the face plate.






Sunday, May 29, 2022

Pullies and lead screw


https://youtube.com/shorts/6r-MS5sE5aU

I've replaced another 3d printer pulley with a cast one, and also added the apron for the cross slide. The cross slide moves about 0.2 mm for each rotation of the lathe l.

The 3d pullies were starting to fail where the bolt was placed to tighten them on to the spindles.


Shaking out, still no biscuit



 So I added the riser, and the shrinkage is sorted, but I have a small hole in the casting, where the metal didn't flow.

hotter pour next time, I time from when the last piece melts.

I also checked to see if the pattern is wrong. It turns out that this section should be thicker, 3/16, not 1/16. Doh.

At least the riser seems to have sorted the shrinkage.







Saturday, May 28, 2022

Close, but no biscuit


The two halves of the pattern are held together by pins, if the pin is too tight, then the pattern won't separate when you split the mold, so you can get breakage around the edges. If the pins are too loose the pattern can slide to one side and things won't line up...

I think that I can clean this up well enough with a file, as it's for a slow speed pulley....


 

Shrinkage

I forgot to add a riser, whatever cools last tends to shrink, particularly if it's thick.

Next times I'll add a big riser.

The riser has to cool after the part so that it can keep feeding in liquid metal as the part cools and shrinks.