Question: Problem with procedures

I created these two procedures, I want to introduce within a mla.
 But I have the following two problems that I need you guys help me 
solve:

A := Matrix(13, 1, [ -2.13534729, 3.1832472,  -1.80143597,  
.71727204,  .50344027,  -.61899395, -0.5332322e-1, 
 .28021362,  .10715224,  -.29302865,  0.4459872e-1, 
.11868632,  -0.5248134e-1]);

B := Matrix(16, 1[.183324722,  .240975303,  .209108771,  .190439972,  .142648498,  0.77993465e-1, 0.12475611e-1,  -0.32267127e-1,  -0.75291522e-1, 
-0.5647067e-1,  0.76201285e-1,  .123893204, -0.29201193e-1,  -0.91173542e-1,  0.1317696e-2,  0.26025526e-1]);

C := Matrix(10, 1, [ 2.78157254,  1.64650916,  -.1371439,  -0.649767e-2,  0.234444e-2,  0.511868e-2,0.187982e-2,  -0.204472e-2,  -0.46122e-3,  0.45724e-3]);

Q := Matrix(10, 1, [ 439.932854,  472.41802,  37.684494,  7.472018,  2.920828,  0.5184e-2,  -.963864,  -.188732,  .191203,  0.49025e-1]);

dWdt := proc (t::float) local DW, DW2, T; description "Se calcula dWdt  según el valor de temperatura";
T := t+273.15;
 if 273.15 <= T then
DW := add(((1/481)*i-1/481)*C[i, 1]*((1/481)*T+(1/481)*(-754.15))^(i-2), i = 2 .. 10)
else
DW2 := add((i-1)*A[i, 1]*((ln(T/273.16)+1.5)/1.5)^(i-2)/(1.5*T), i = 2 .. 13);
DW := Wr(t)*DW2
end if;
evalf[9](DW)
end proc:

dTdW := proc (W::float) local DT; description "Se calcula la dT9dWsegún el valor de Wr";
 if 1 <= W then
DT := add((i-1)*Q[i, 1]*((W-2.64)/1.64)^(i-2)/1.64, i = 1 .. 10)
else
DT := add((1/6)*B[i, 1]*((W^(1/6)-.65)/.35)^(i-1)*(i-1)/((W^5)^(1/6)*(W^(1/6)-.65)), i = 1 .. 16)*273.16
end if;
evalf[6](DT)
 end proc:

1  I do not understand because I get different values when   evaluating different  significant figures.

(D(dWdt))(231.928);
                              D(dWdt)(231.928)

seq(evalf(%, i), i = 1 .. 9);
      -0.000003, -0.0000030, -0.00000117, -0.000001170, -0.0000011600,

        -0.00000116000, -0.00000120000, -0.00000120000, -0.00000100000

(D(dTdW))(1.60980185);
                             D(dTdW)(1.60980185)
seq(evalf(%, i), i = 1 .. 9);
          20., 21., 21.5, 21.50, 22.000, 22.000, 20.000, 20.000, 0.

 2  I can not use these procedures with ScientificErrorAnalysis package

.with(ScientificErrorAnalysis):

T := Quantity(t, dt);
dWdt(T);
Error, invalid input: dWdt expects its 1st argument, t, to be of type float, but received ScientificErrorAnalysis:-Quantity(t, dt).

Tx := Quantity(231.928, 0.5e-2);
dWdt(Tx);
Error, invalid input: dWdt expects its 1st argument, t, to be of type float, but received ScientificErrorAnalysis:-Quantity(231.928, 0.5e-2)

 I need help from you, to solve these problems.
GraciasI need help from you, to solve these problems.