Community Tip - Learn all about the Community Ranking System, a fun gamification element of the PTC Community. X
Hello,
The Maximize function does not work properly in the following case - I need to calculate the worst case error in a current measurement circuit where x, y, z, y, v, w are the sources of error (resistors, opamp gain, offsets, temperature dependance coefficients) and a, b are the current and the temperature which I would like to be parameters so that I can plot later the worst error versus current and temperature. The Maximize function gives the correct answer - all variables saturated at their maximums only for CTOL=10^-9 - this is completely strange and unacceptable behavior!
For CTOL=10^-10 the solution is erroneous and also for any other CTOL up to 10^-21:
I guess these Maximize and Minimize functions are really nor working well with somewhat more complex examples which unfortunately are present in the real engineering practice!
For CTOL=10^-10 the solution is erroneous and also for any other CTOL up to 10^-21:
You are dealing with numeric approximations and setting CTOL to a value as low as 10^-21 sure is far too optimistic, given that Mathcad just uses the standard IEEE format to store its numbers with an approximate accuracy of about 14 digits.
Usually when you set CTOL to 10^-10 or lower chances are very high that you run into a "calculation does not converge" error and get no result at all.
Looks like Mathcad is not the suitable software for your specific tasks.
On the other hand having to play around with guess values and accuracy settings is not unusual when dealing with numerical approximation algorithms and it looks like you found a setting which gave you the desired results.
if you find a software which does it automatically and more comfortable, you sure should use it. Mathcad 15 is not sold and not supported anymore, so you can't hope for any changes or improvements and have to use it as it is.