Greensand Mold Attempt #1 = FAIL

I had a little bit of time today and the weather was relatively nice so I thought I'd try making the greensand mold for the taillight housing. The first thing that was apparent is that my flasks are just a bit too small for this. Originally, I was sizing them specifically for this part but my last minute changes of adding a strip of molding (too better hold the sand), plus some minor tweaks to avoid wasting wood, led to the working area being a little smaller than I intended. I always seem to do that.

Here's the form in the drag (which is upside down) ready to be packed with a light dusting of talcum powder to help release the pattern. My initial thought for the best way to cast this part was to have the parting line at the rim. That way, with the housing opening upward like a bowl, gravity can help the aluminum fill in the back. In addition, if I happened to use too little aluminum or have any other voids or casting defects, it will be much easier to fix on the back than say on the edge part of the housing.
	The sand has been packed and the drag turned right side up ready to receive the cope. It packed in nice and tight.  I used the screen from my last greensand post to "dust" sand over the pattern until it was sufficiently covered before dumping the remaining sand on top.
I dusted the inside of the pattern, added the sprue and vent tubes, and packed the cope tightly. I used 1" PVC pipe for both tubes because that's what I had handy.
	The tubes were removed and I tried (as gently as possible) to separate the flask. FAIL! The portion of the mold for the interior pattern sheered off a the parting line, not what I was hoping for.
I pulled the pattern with its payload out of the drag though not being terribly careful as I knew the mold was already ruined.  I tried tapping on the back of the pattern to dump out the interior sand but it held tightly. This is going to require some reconsideration of my method and placement as well as some adjustment to the pattern if I intend to used greensand.

There are a couple of things that I immediately know contributed to this failure. Number 1 is that before making the pattern, I knew nothing about draft angles. For the those like me who do not know, the pattern is supposed to have a slight angle (less than 5 degrees) on any surface that contacts the sand and is not parallel to the parting line. I.e. if the pattern has to rub against sand when being removed, it should have a slight angle to reduce friction, making removal easier. As I said, I did not know this and built the pattern with the walls perpendicular to the back and rim.

The other item, and which further compounded the lack of draft angle was the interior surface of the pattern. Since the pattern was cut in layers using the laser cutter and then glued together, there are slight irregularities between layers. On the exterior, I was careful to sand and fill them until perfectly smooth but on the interior, I wasn't as thorough. This led to an uneven surface (albeit slight) which gave the interior sand places to grip.

My plan is to rework the pattern to clean up the irregularities while keeping in mind draft angles. If that doesn't make the mold come out any better, I'll most likely switch from greensand to the lost-wax method of casting.

It seems like a jump but I'm already working with the wax method on another casting project I'm doing. I haven't made any posts about it though because it's a test for someone else. If it turns out okay (and I get their permission to share), I'll do a write up on my process.