If just past your exhaust ports you open the exhaust big and wide, your exhaust velocity stalls. You are then more susceptible to reversion. That's why I stepped my headers - my exhaust velocity might slow, but it keeps moving, and since the next step is bigger, there's still room for the exhaust to expand toward the exit. That's also why megaphone headers are so effective. It's most effective if you can keep the exhaust flowing straight for at least two inches before you have any bends in the primaries. I didn't have that option on the Cadillac headers.
You must be registered for see images attach
They look like hell, but they flow like the mighty Mississippi. No, I do not have a tubing bender. Mandrel bends are fairly overrated, and this was a Lemons car, so I wasn't going to spend big dollars on the exhaust.
Years ago there was an engineer named Paul Yaw who built rotary engines in his shop on the west coast. He did a lot of bench testing with an airflow meter he built, and discovered a lot of this and wrote really good technical articles. Mazda contracted him to do some work on the Renesis motors, but when the rotaries faded out he started working on LS motors. I wish I still had those articles saved somewhere.
Doh! I just remembered his name, googled it, and found his old articles!
http://web.archive.org/web/20051222093839/http://www.yawpower.com/techindx.html
He now runs Injector Dynamics.
http://injectordynamics.com