For point (1): Rest supports are close by.The program arrives at the spring hot load based on dead wt. distribution of the system at " no temp" condition.So if two rest supports ( one of them or both of them may be springs ) are close by , the dead wt. distribution can result in a zero or near zero load on one of them.Solution: space the supports apart.This can also be solved by using user designed springs at location, but that in my opinion is not generally a good design as the system dead wt. distribution should not be disturbed much.
2) The most general reason for this flag is a variability greater than 25 % ( or whatsoever is the value mentioned in the spring spreadsheet)or if the load could not be met by a spring based on variability and space ( if mentioned) within the working range i.e. the load falls in the "lift off" or "bottomed out" areas of the load range. .Solution: try a variability higher than specified value.I have spoken to many spring vendors and I feel that if they agree , although we should try to control the variability between 5-15% close to strain sensitive equipments, (just my opinion , I cannot provide any data or published work to support this), we can exceed the variability beyond 25 % ( 25 % comes from an MSS-SP recommendation)and solve the problem.But I would not recommend exceeding the variability beyond 35 % and suggest that if the variability exceed this no. , better go for a constant spring.
The other reason can be to use the user defined spring ( a variable spring if constant spring is not desired )
Regards
_________________________
anindya