The following topic areas are discussed on this page.  To jump to a specific topic, click on the link to that topic.  Alternatively, you can read the whole document in order.  The order in which I have chosen to describe the commands is not representative of the order in which they should be used, which will vary.  This is not necessarily a complete list of commands, but you should be able to achieve everything you want to using them.

Introduction
A brief introduction to the form PCL-3 commands take, and how to read my descriptions.
Page Setup
Paper size and orientation, print area and text wrapping settings.
Font Selection
Selection of character set, font, size, pitch and spacing.
Character Formatting
Setting of italic, boldface, underline, superscript and subscript.
Setting and Moving Print Position
Positioning and moving by row and column, dot position and decipoint.
Miscellaneous Commands
Printer reset and extended character printing.
Printing Raster Graphics
Setting resolution, compression and offset and printing graphics.

Most of the characters which make up the PCL-3 commands are standard printable characters.  There are some exceptions, most notably escape character 27 (1Bh) which all PCL-3 commands begin with.  I will represent this character with the § character.  To eliminate any ambiguity caused by different character sets, I will show all commands as a series of hexadecimal values within square brackets as well as printable characters.

PCL commands follow a standard structure starting with the escape character.  After the escape character comes the command type, the command value and a terminator identifying the specific command.  An example of this is found in the font selection commands, all of which begin with §(s or [1B][28][73].  Commands of the same type can often be strung together without the need to include the command type bytes by changing the terminating character from upper to lower case.  For example, to choose a 12pt 10cpi font:
§(s12V§(s10H or [1B][28][73][31][32][56][1B][28][73][31][30][49]
can be replaced with §(s12v10H or [1B][28][73][31][32][76][31][30][49].

In a lot of cases, I have only listed one example of a command.  This is when it is clear what the command value represents.  An example of this is font size selection.  I have only listed §(s12V for 12pt, but clearly §(s6V is equally valid for 6pt.  Consult your manual or hunt around for a full list of valid typeface and size combinations for your printer.

Typical Graphics Printing Procedure
To print raster graphics you must first set your graphics resolution.  The Deskjet 320 has a maximum resolution of 600x300dpi, but later models allow 600dpi in both dimensions.  Next you must set the graphics compression mode, which tells the printer in what form the graphical data will be sent.  When you have positioned the printhead where you want to start printing graphics, issue the Start Graphics command.  For every line of graphics you must have one new Transfer Graphics command, optionally preceded by a Horizontal Offset command.  The Horizontal Offset command specifies how far across the page to start the current line of graphics.  When you have finished transferring the graphics data, issue a Finish Graphics command.

Graphics Compression Modes
The HP Deskjet allows graphical data to be transferred in three formats.  All of these formats deal with dots which can either be black ink or paper left white.  These dots are dealt with in groups of eight, which occupy one byte of data.  The following paragraphs should explain each of these formats in detail sufficient for you to produce your own raster graphics.

Graphics Mode 0 (No Compression)
This mode transfers raw pixels in groups of eight.  Each arrangement of pixels is translated into one byte of information.  This is done by treating a group of eight pixels as binary 1s and 0s where black ink is represented by 1 and white paper by 0.  The lefthand pixel in the group of eight is the most significant bit, and the righthand one is the least significant bit.  To print [White][Black][Black][White][Black][Black][White][Black], convert to binary 01101101 = m or [6D].

Graphics Mode 1
In this mode, graphics data bytes occur in pairs.  The first byte of the pair represents the number of repetitions of the second byte, which is the data byte.  The data byte is in the same format as a data byte used in mode 0.  If you wanted to print 80 black pixels in mode 0, you would have to send ten bytes: [FF][FF][FF][FF][FF][FF][FF][FF][FF][FF].  The byte [FF] occurs 10 times, which is 9 repetitions, so you need only send: [09][FF].

Graphics Mode 2 (TIFF Group 4 Compression)
This mode is concerned with both repeating bytes and conventional non-repeating bytes.  The data bytes occur in groups of variable length.  The first byte of a group indicates whether a repeated byte or a series of unrepeated bytes follow.  If the first byte has a value from [00] to [7F] then it indicates a number of unrepeated bytes follow.  A value of [00] represents one unrepeated byte, and a value of [7F] represents 128 unrepeated bytes.  If the first byte has a value from [81] to [FF] then it indicates a repetition of the following byte.  A value of [FF] represents one repetition (two occurrences), and a value of [81] represents 128 repetitions.  If you wanted to print a pattern represented by the following thirteen data bytes in mode 0: [FF][FF][FF][FF][00][18][00][FF][FF][FF][FF] you could compress this data in mode 2 to just eight bytes: [FD][FF][02][00][18][00][FD][FF].