firefreak11:

Are they extremely complex?

 Obviously, yes. Other calculators are simply based on a logic chip, and does not contain any adders. They only work with +, is a one digit calculator, and therefore can only count until 18.

Each "actual" calculator has full adders, which consist of 2 half adders.

The half adder is a XOR and AND, with each input directly connected to these two. How the half adder works is that if only 1 out of 2 inputs is on, The XOR lights up and the AND doesn't, but if the half adder has 2 inputs on, because of logic, XOR doesn't light up and the AND does.

Here is the logic circuit for a half adder:

The dots mean they emerge (do not merge with other wires) to create another wire, which means they are linked. This means if A is on, the link emerging from A is also on.

The half adder is completely useless for calculators, because it does not "carry" the outputs, and will only work with 2 inputs. It is only useful as a component for creating full adders.

The full adder is completely essential, and how much the calculator can add depends on how much there is in the calculator. Now this is where explanation becomes very complex. However, if you could judge how the picture works, then you could judge how the full adder works already:

   (darn wikipedia)

The problem is converting these results to decimal and vice versa for inputs. I do not know any methods for converting binary into decimal and viceversa in TPT as I do not have this much intelligence into logic gates yet. You can search in the Internet for showing any methods, and following the instructions for connecting the logic gates and components involved.

PTuniverse:

Other calculators are simply based on a logic chip, and does not contain any adders. They only work with +, is a one digit calculator, and therefore can only count until 18.

Also, there are different kinds of calculators. There are adder-based calculators, that store data and calculate the results with all adders in an instant. This kind is very fast, but (could be) massive. Another kind is queue-based calculators, that queue the data, and read them slowly to be processed by only a few adders. This kind is slow but compact. There's only a few calculators in TPT based on this kind. 

These are the fake calculators that are only based on logic chips:

This is one of the real calculators (I suppose this is queue-based) that is based on both addition/subtraction:

If you want to make a calculator I suggest you make a fake memory based one first. If you want to convert it to a real one later, used decimal to base 18 instead of decimal to binary.