VERY BASIC Question about running ACE model

I am now faced with another question -- before I ran the model, I used tetrachoric twin correlations to estimate heritability using Falconer's formula, as well as c2 and e2.

Using the script from the website, and specifically this code at the end:

A <- mxEval(A, twinACEFit)
C <- mxEval(C, twinACEFit)
E <- mxEval(E, twinACEFit)
V <- (A+C+E)
a2 <- A/V
c2 <- C/V
e2 <- E/V

I come up with a number for a2 that's very similar to my original estimate. But c2 and e2 are REALLY, really, really different. Before I delve into additional resources attempting to understand why this is the case, could anyone think of any obvious places where I could have made a mistake? I made literally no modifications to the code on the website. Thanks again.

I'm working on applying the script cited here for another project -- I'm still estimating ACE estimates for a binary variable. When I run it, I'm getting the following error message:

> univACEOrdFit <- mxRun(univACEOrdModel, intervals=T)
Running univACEOrd
Warning messages:
1: In model 'univACEOrd' while computing the lower bound for 'ACE.A[1,1]' NPSOL returned a non-ze
ro status code 6. The model does not satisfy the first-order optimality conditions to the require
d accuracy, and no improved point for the merit function could be found during the final linesear
ch (Mx status RED)
2: In model 'univACEOrd' while computing the lower bound for 'ACE.C[1,1]' NPSOL returned a non-ze
ro status code 1. The final iterate satisfies the optimality conditions to the accuracy requested
, but the sequence of iterates has not yet converged. NPSOL was terminated because no further imp
rovement could be made in the merit function (Mx status GREEN).
3: In model 'univACEOrd' while computing the lower bound for 'ACE.E[1,1]' NPSOL returned a non-ze
ro status code 6. The model does not satisfy the first-order optimality conditions to the require
d accuracy, and no improved point for the merit function could be found during the final linesear
ch (Mx status RED)
4: In model 'univACEOrd' while computing the upper bound for 'ACE.E[1,1]' NPSOL returned a non-ze
ro status code 6. The model does not satisfy the first-order optimality conditions to the require
d accuracy, and no improved point for the merit function could be found during the final linesear
ch (Mx status RED)
5: In model 'univACEOrd' NPSOL returned a non-zero status code 1. The final iterate satisfies the
optimality conditions to the accuracy requested, but the sequence of iterates has not yet conver
ged. NPSOL was terminated because no further improvement could be made in the merit function (Mx
status GREEN).

And then when I do:
univACEOrdSumm

I get:

The final iterate satisfies the optimality conditions to the accuracy requested, but the sequence
of iterates has not yet converged. NPSOL was terminated because no further improvement could be
made in the merit function (Mx status GREEN).

Even though the summary says Mx status GREEN, the fact that I have some interspersed Mx status RED means that I cannot trust the estimates, right? I realize I haven't provided many details and will work on posting my full code up later -- I was just curious if anyone knew off the top of their head where am I likely to have made a mistake? Thanks.