My first idea on how to run the gates and venting system was to just use large enough vent risers to help fill the part with liquid metal during the solidification process. The thicker parts (mounting points) need to have metal flowing into them up till the time they turn solid. If not the shrinkage would most likely cause distortions in these thicker areas of the part as they cool and shrink. I also was thinking about running the Sprue down the center of the mold and sending the runners and gates to each of the mounting points. I'm sure the pattern would fill up just fine that way but, since the walls of the tank have such a large and relatively thin area throughout the mold, I decided to go with a bottom fed gating system. By feeding the largest section (front of tank) from underneath, its my hope to have very little turbulence in the metal as it flows threw the mold. I'm probably over thinking this but I know turbulence is bad, and it will probably be easier to gate the mold this way anyhow.
So then I started reading about directional solidification. What that is, is where the metal first starts to go from a liquid to a solid and continues to solidify from that point on threw the part. By adding chunks of metal to the mold you can decide where the solidification process begins. In my gas tank mold I would want to add the chill plates to the tank mounting points. The advantages in doing this are, the way the metal sets up with a chill plate makes for a more mechanicaly strong area. It has to do with the way the metal molicules lighn up during solidification. Thats another topic for another post though. Also, by chilling the larger parts (not close to the gate. Wouldnt want to stop metal flow threw the supply gates.) first, there is less of a chance that they will deform from a lack of liquid metal in their center during silidification due to shrinkage. Another plus for using chill plates, and one of the biggest is in the amount of metal needed to fill the mold. If I were to just put big ol risers on top of the mounting points, I could probably kiss a pound of aluminum good by for each mold I fill during a melt. That would have me melting enough extra aluminum to make the large riser concept not such a good idea. More melting time, more metal, more slag, etc, to deal with over all. If I were only planning on casting one part per melt then it wouldnt matter much. However I plan on spending a good five hours pouring as many molds I can make each time I fire up the furnace.
I would use more vents with the chill plates, but they get poked in with a coat hanger size wire, so its not a big deal. I'm also planning on using Petrobonded sand. The pertobond uses oil rather than water to hydrate and does not create steam. Since there is no steam, less venting is permittable. The finish with Petrobond is finer as well, cutting down on finish work to the part after casting. The sand is a smaller grain because it doesen't need to be as permiable as regular water based green sand.