I have changed limit settings without solving my problem. If I can separate T-Stat/circulator circuit from the boiler temp control circuit I believe I can eliminate short cycling caused by T-stat/upper lim interaction. Then boiler control will be based on boiler temp limits only. I would live with variation of temp around the Hi Lim set point. The furnace appears to be oversized for the house at 100,000btu-h. Max burner on time last season on a 50 HDD day was 5 hrs/day. So recovery time to handle T-stat demand should not be a problem.
However, now the burner runs until either 1. the t-stat is satisfied, or 2. the hi lim cuts it off, or 3. all t-stats are satisfied and the temp drops below the hi lim diff. With five t-stats on line the probability of the temp hitting the upper lim before any t-stat is satisfied and another comes on line, etc. is much greater. With the t-stat out of the boiler circuit and T-T jumpered the burner on-off cycle will be determined by the constant factor of high lim setting and the differential. Without any circulation the time to raise the boiler temp from low to high is 6 minutes - enough time to get the stack temp up and minimize condensation. A 5 to 30 second cycle time is not long enough to do that.
My objective is to maximize efficiency within the system limits (minimum condensation and corrosion).