I did try the upside down orientation but with not very good results.
I placed an 0.8mm wide by 0.5mm high strip around the top of the body - wide enough for a medium sized support to attach to but also leaving enough of the "proper" top surface to provide a good guide to file back to. I also avoided covering the corner brackets in the hope that I could support and print them on their own.
Medium sized supports were placed around the strip and the corner plates were supported with light supports.
The result wasn't too good. The body had come adrift from the supports at the end opposite the door. It looks as though this happened when the printing was around the support/body join since the body seems to have printed out not too badly above to the end of print.
From the end view it looks as though the supports in the middle might have kept in contact, the ones at the extremities being the ones that have separated. They have separated now but that could have been my handling during the taking off the plate and the washing/curing. The dimensions for the medium supports used were exactly the same as used for the "right-way-up" orientation and I had no problems with them separating.
At the other end of the body, the supports worked very well with no hint of separation, so I'm not sure what's going on.
And just obvious in some of the other shots, the body has bowed in quite noticeably about half way up.
I ran another print and exactly the same thing happened. The medium supports are exactly the same as the ones used on the "right way up" version and I had no separation problems with that orientation.
So I want back to the "right way up" orientation and printed off some more for the SSMRS AGM. At least these prints are now predictable and all printed well.
I might go back and see if I can get any better results with the upside down orientation. I could beef up the supports to see if that improves matters or I might try just upping the exposure time to see if that does it. I've done that before if I have had support separation problems and it has worked. As for the bellying of the sides, I note what Mick said earlier about thickening the sides but I'm a bit reluctant to do that since the tops are overwide as it is.
Jim.