There's no locking around the OACTIVE flag set/clear, right? Is it possible that multiple TX threads are fiddling with OACTIVE and then it's not being properly cleared and tx kicked? Adrian