Before we get into actual disassembly of machines for the purpose of repair, a certain idea or concept must be covered in order that the person attempting repair fully understand the job to be done.  We'll do that now.

From the very beginning, typewriters have been what we might call "Precision Manufactured Machines."  A great degree of skill in manufacture was required even for the first successful machines, and as time went on and machines became more complex, this skill in manufacture was matched by precision manufacturing of the parts and assemblies contained in the machines.
Because of this requirement for precision in parts and in assembly, many early typewriter builders were also successful in other areas requiring equal competence -- areas such as sewing machine or firearms manufacture.  Indeed, both the Remington and the Smith Premier firms first built firearms before typewriters; the design of the Smith Premier No. 1 was wholly developed based upon the idea that the levers used in the first Sholes & Glidden (later Remington) would eventually work out of alignment -- and so the Smith Premier used rotary or crank motivation for its type bars.  Such things were, then, recognized even very early -- that precise control of the dimensions of parts, tolerances between parts or of machined or drilled openings, and assembly and particularly alignment were not only desirable but vital for a machine which had to write many thousands of letters which, if all things were proper, could stretch thousands of miles straight with not one being up or down, too far right or left, or too dark or too faint when compared with all the other typed letters.
Most of the early machines were built using assembly line methods in which the whole cast solid frame of the machine was moved along, at each station having a workman add one or several parts and making up connections between operative functions as needed.  Eventually, when finished, the machine would reach a bank of very highly skilled technicians who actually had to test and completely align the machines, one by one, before they were crated up for shipping.  The methods of assembly were just not sufficient to prevent the machine from being built imperfectly.  Eventually, machines such as the Harris Visible and the Woodstock appeared which were built on the premise that precision-built SUB-assemblies were mounted into the machine, aligned before such installation; in such machines, any misalignment of one section was not supposed to be able to transfer (by linkages) to another.  Mountings for the subassemblies were highly precise, so that very high-quality machines requiring little attention at the packing end of the factory were built.
That premise continued through newer and newer typewriters, through the adoption of stamped frame members and plates, and through to the employment of dowel-mounted key levers mounted in blocks or plates -- right to the end.  In every case, even with the most modern and most simple machines, the idea of precise alignment was paramount.  Any machine that was offered HAD to operate well when new -- letters in line, carriage moving and returning smoothly, tabulator releasing and stopping the carriage positively, and so on.

WE MUST REMEMBER THIS as we begin to consider disassembling ANY part of ANY typewriter today.  What are the important points, regarding precision and fit and alignment regarding this part, or that part -- this assembly or that?  Every time a part or assembly is removed it must be understood that various adjustments and alignments were likely made on a machine when built and later if or when rebuilt or repaired.  Disassembly does two things -- it allows access to parts, and thus removal and repair of parts, but it also disturbs alignment and fit.  We must always be sure that a part or assembly can and WILL go back in exactly the orientation, fit and alignment it is supposed to.  We must also remember that disassembling and then reassembling a machine in top shape prior will very often result in a misaligned machine, just simply by way of our own imperfections in work, if proper steps are not taken during the reassembly.  And we must know what those steps are -- BEFORE we disassemble anything.
Terminology is extremely important in any mechanical pursuit; it is the terminology that allows everyone to understand what anyone says or writes.  Let's take a look at some mechanical diagrams of various typewriters, and explain the terminology and actions of some various mechanisms.  We will start with that part of the typewriter most familiar to typists; that part which operates as a direct function of the keys to print characters on the paper.  This entire mechanism can be described as the TYPE BAR MECHANISM, or sometimes as the KEY LEVER AND TYPE BAR MECHANISM.

KEY -- the actual button depressed by the fingers to operate functions.  The term "key" is in some places considered inaccurate, in that the actual part is normally called a KEYTOP.  The Keytop is that part attached to the KEY LEVER, and may be one solid part with a letter or number painted on or cast in, or else may be made up of parts.  In the COMPOUND KEYTOP there is usually a RING surrounding a glass cover and a LEGEND (usually paper) on which are printed the characters.  The key ring holds the whole assembly to the key lever by way of tabs bent around below; the end of the key lever is formed into a round table of sorts, sometimes cast in one piece but also often fabricated by welding.  The KEY LEVER is that first lever which is operated by the key's depression, and there can be many other parts as well.  Let's examine a machine designed by Halcolm Ellis, which became the French-made M.A.P. machine.
On the left, we see the keys and the key lever indicated by (1) on the diagram.  When a character key is pressed, the key lever (1) rotates downward in an arc, hinging at its farthest end at point (2) with motion restricted by spring (3) which will help to return the mechanism to rest when the key is released.  Lever (1) moving down causes slot (7) naturally to angle downward to the front, and riding in this slot is pin (8) which is on the intermediate lever (9).  The pin (8) moves down or clockwise as seen here, moving intermediate lever (9) with it hinging around point (10).  Intermediate lever (9) moving to the rear imparts motion at point (11) to the intermediate link (13).   Intermediate link (13) is connected directly to a type bar at point (14) and since the link is moving backwards, the type bar, hinged at point (23) then rotates bodily until the type slug strikes the ribbon, paper and platen.  Gravity acts on the type bar assisted by spring (3) to return the type bar to rest horizontally and the key to its fully raised position.
The machine above has what we might call FULL LENGTH KEY LEVERS; they extend all the way through the machine, essentially underneath all the working parts.  This is a very common design feature.  The machine at left, a variant of the Victor typewriter, has what we might call HALF LENGTH key levers.  The key at right, being depressed, causes key lever (11) to hinge at point (12) seen as a dark circle here.  The opposite end of the lever thus moves up as the key end moves down, and this causes intermediate lever (9) to rotate backwards in the machine (or counter clockwise as seen here.)  This imparts motion to intermediate link (8) which pulls the type bar to the print point.  Note location of springs (25.)
In both of the designs above, the type bars are pulled to the print point from the rear.  In the Stearns typewriter, seen at right, the type bars are pushed to the print point from below and to the rear.  Key lever (2) moves down, hinging at its rear; note location of springs near the rear of the full-length key levers.  Mounted in a row above the key levers are the intermediate levers (12); a pin and slot connection is again used, which in this case pulls the intermediate levers forward which is well shown in phantom form in this drawing.  This imparts motion to a set of intermediate links, which are connected at their opposite ends to the type bars. 

In all three designs, note how the "action" is coming from below and 'around behind' the type bars.
Some typewriters do not employ intermediate links in their type bar actions; at left is the Burroughs typewriter.  The action is easy to follow, with a phantom view of the key lever at left (at rest) and a type bar at rest.  Note the hinge for the intermediate lever at point (10) but note that the intermediate lever connects directly to the type bar with another form of pin and slot connection at (11).  Motion of the intermediate lever forward, then, directly imparts motion to the type bar itself causing it to be raised to the print point.
Of course, it would also follow that there are typewriters which do not employ intermediate levers.  At right we see the National Portable typewriter, which we have colored for easy viewing.  The key levers, in red, are hinged at just forward of their half-way points and act to raise the opposite end when the key is pressed.  Motion is imparted vertically to an intermediate link, shown in light blue, which is connected at its other end to the type bar (dark blue) shoving it to the print point.  Thus, in this design only two hinging actions occur.  Typewriters of this design tend not only to be sloppy in their actions but also are most likely to transmit misalignment of the key lever to the type bar if the actual mounting of the type bars allows any side play at all, through excessive tolerance at manufacture or through wear and tear.
At left we see a 1932 design for an Olympia portable typewriter.  We can easily see the full length key levers, and note that an intermediate link is connected to the type bar near its forward end.  This is the 'first intermediate link' and it is connected to a BELL CRANK (3.)  This bell crank rotates forward when pulled down at its forward end by the first intermediate link, and thus pulls the second intermediate link which pulls the type bar to the print point.  Note that two different sizes and lengths of bell crank are shown; these correspond to actual positions in the machine and are longest at the outsides, shortest in the middle.
All of the machines seen above include as a design feature key levers that run some way underneath the machine, whether all the way under and through or just part way.  Many small typewriters did not use long key levers such as these, and if there is no key lever under the machine the action is quite different.  I coined the term "front mounted type bar action" some years ago to describe machines of this variety; let's look at one.
Here we see the design of the Smith-Corona SKYRITER portable typewriter.  It is immediately obvious that it is small and flat and has almost nothing below the type bars.  Depression of keytop (40) causes the key lever to rotate at point (12) which pushes backward on the top of a bell crank, seen in red and numbered (13).  The lower end of (13) pulls on the intermediate link (18) which, being directly connected to the type bar (green) pulls it until it reaches the print point.  Note that the shortness of the key lever causes the keytop to move in a much sharper or shorter arc than in any of the machines seen above, which is a common source of complaint with small portables such as this one.  Some typewriters on the other hand have no arc to their keytops at all when depressed; the keytop remains flat or horizontal throughout its travel.  This is known as "parallel key action" and is seen below in the late Erika portable typewriter.
We see here that each keytop is mounted on a slightly complicated looking linkage; a phantom view of a mechanism in operation is provided.  The use of two parallel links (2 and 3) forces the key to travel straight down, or nearly so; motion for the type bars is derived from the lower parallel link (3) which is extended to connect to intermediate link (9) with springs (8) also attached.  The intermediate link then pulls the type bar (10) to the print point.

Every single key lever / type bar mechanism employed in a model line of any manufacturer was usually either patented or produced under license, and it is impossible to illustrate them all since we would need one illustration for every model line of every manufacturer.  However, these illustrations should give a good indication of the wide variations in design that the repairman might confront -- and the more complicated or crowded they are, the harder they are to repair.