Convolution of the cube having different units: Jy/beam and Kelvin

[renovatio]

Dear Miriad users/developers,

I have a little question just to make sure that I am doing the convolution and Jy/beam---Tmb conversion in an correct order.

Since I will work on molecular line ratios, data cubes of both molecular lines have to have the same resolution, so the same beam size. I also like to have the unit of the final cube in Kelvin, not in Jy/beam. Thats why I am first convolving the both cubes to the same final beam size using 'convol' task with 'final' keyword. The default for the 'scale' keyword of 'convol' task to determine an appropriate scale factor to make the output in Jy/beam. To do this, it looks at the header for the unit and the beam sizes. I think convolve does its job correctly for the cube in Jy/beam. 

I tried the following two different ways to get the final cube. Do you think which one has the correct order:

Cube(Jy/beam)-----convolution to the final beam A"-----Jy/beam conversion into Tmb (cube in Kelvin)
Cube(Jy/beam)-----Jy/beam conversion into Tmb----convolution to the final beam A" (cube in Kelvin)

Same pixel in the both resultant cube (both in Kelvin) has very different values. Should it not be the same? 

BTW, I am using 'imstat' to calculate the K/(Jy/beam) conversion factor to convert Jy/beam to Tmb (K) scale. Then I multiply the cube with this factor to obtain the cube in Tmb (K). However, the unit in the header of the cube (in Kelvin) still says 'Jy/Beam', and 'scale' task looks for this unit to calculate the scale factor correctly. In other words, 'scale' still sees 'Jy/beam' unit in the cube which actually have the values in Kelvin. Thats the thing makes 'convol' task confused I think. What do you think? Sorry in advance if the question is so trivial. I just want to get the convolved cube in Kelvin correctly.

Many thanks!