Doug,

When I was considering installing matched length headers of about 18" length on my 13B, I wondered about adding support from the collector to the engine block.  The deflection recorder in the photo was constructed to see how much the headers deformed between cold and at operating temperature.  The fingers on each side of the collar were made of annealed copper to avoid springback.  As the headers moved, the shaft and collar would slide in the frame and permanently bend one of the fingers on either side of the collar while the other finger on each side didn't contact the collar and indicated the starting condition.  The shaft is 1/4" diameter for size reference. The photo is of the recorder after the test.  In the other photo, the installed recorder can be seen through the center of the coil of brown thermocouple wire.

The results of the test indicated that when hot, the collector moved upward as expected from uniform heating around the diameter of the headers and the longer length of exhaust header on the outside of the bend.  What I didn't expect was the deflection exceeded the distance available in the recorder and the bent finger was smashed between the collar and the edge of the rectangular hole to the right of the collar in the photo.

My take on this was that adding rigid support from the collector to the engine block would cause more stress than it would relieve.  Rather than a rigid support, you might consider installing a damping element to prevent or minimize resonant vibrations.  Ball joints close to the header flanges might also be worth considering in which case, rigid support at the collector might be effective.

At least, that was what I saw.  The experience of others may be different.

Steve Boese