Skip to content

Search is only available in production builds. Try building and previewing the site to test it out locally.

FIS character set

Which byte code shows which character on the cluster's centre display: close to ASCII, but with the lower-case letters a to o moved to 0x01-0x0F and accented letters, arrows and symbols from 0x80. Transcribed from one 2016 table, uncertain cells marked.

Applies to: B6 onlyCommunity write-upFrom an owner’s write-up or forum thread. Not checked against factory documentation.

The text messages for the cluster display (see Writing text to the cluster display) carry one byte per character. Which character each byte shows is set by the cluster’s own table, which is close to ASCII but not the same.

The post says the table Audi uses is “more limited than the official ASCII table”, and that “especially the little letters ‘a’ to ‘o’ are in a different position than regular”, so text has to be converted before it is sent to the display. The table shows:

  • 0x20 to 0x5E match ASCII: space, punctuation, digits, @, the upper-case letters and [ \ ] ^.
  • 0x5F is Ä, not the underscore.
  • a to o are at 0x01 to 0x0F, and p is at 0x10. Their ASCII places 0x60 to 0x70 hold other characters (Ö, Ü, punctuation) or are empty.
  • q to z stay at their ASCII codes 0x71 to 0x7A, and { | } at 0x7B to 0x7D.
  • 0x80 to 0xBF hold accented letters, arrows and symbols such as ß, £, µ, ½ and §.
  • Some characters appear at more than one code: the four arrows (0x18–0x1B and 0xAC–0xAF), ñ (0x8A, 0x9A), ç (0x8B, 0x9B), ş (0x8C, 0x9C), the space symbol (0x00, 0x17, 0x20, 0x65), the short upright stroke (0x27, 0x62), . (0x2E, 0x63), , (0x2C, 0x64), _ (0x12, 0x66), * (0x2A, 0x6D), > (0x3E, 0x6E), - (0x2D, 0x68), % (0x25, 0xA3), V (0x56, 0x67), and i (0x09) if 0x8E turns out to be an i rather than ¡. The table does not say whether they differ on the display.
  • The screenshot shows rows 0x0 to 0xB only. Codes 0xC0 to 0xFF are not in it; whether the original table had them is unknown.

So TEST is 54 45 53 54 as in ASCII, but Test is 54 05 73 74: the e moves to 0x05, while s and t stay. The worked example is on Writing text to the cluster display.

Row is the high hex digit, column the low one: the cell in row 4x, column x1 is code 0x41. Hover over a cell (or tap it) for the code and any note.

FIS character codes: row is the high hex digit, column the low one
x0x1x2x3x4x5x6x7x8x9xAxBxCxDxExF
0x␠abcdefghijklmno
1xp√_¤␠↑↓→←
2x␠!"#$%&'()*+,-./
3x0123456789:;<=>?
4x@ABCDEFGHIJKLMNO
5xPQRSTUVWXYZ[\]^Ä
6xÖÜ'.,␠_V-?*>
7xqrstuvwxyz{|}¯
8xáàéèíìóòúùñçşß¡IJ
9xâäêëîïôöûüñçşĝıij
Axªα©%π₠£←↑→↓
Bx¹²³±µ¿º¼½¾§
Cx
Dx
Ex
Fx
A
reading uncertain
?
cannot be read in the source
empty in the source (no character)
not in the data yet
␠
space

Transcribed from a screenshot of an English machine translation of the post 'Audi ASCII table' (janssuuh, 1 April 2016); the original post could not be retrieved, so the screenshot could not be checked against it. The screenshot shows rows 0x0 to 0xB only: codes 0xC0 to 0xFF are not in it. Several characters appear at more than one code (for example the arrows, ñ, ç, ş); which of them the cluster prefers is not known. Nothing here has been checked on a car.

Every cell was read from the screenshot, enlarged; cells were checked one by one, and empty cells were confirmed by measuring the ink in each cell. The marks mean:

  • Normal cell: the character is clear in the screenshot.
  • Uncertain (marked): there is a likely reading, but the glyph could be another character. Nine cells: 0x00, 0x17 and 0x65, which show the same open-box space symbol (⎵) as 0x20, but nothing says they act as a space; 0x27 and 0x62, two almost identical short strokes that could be an apostrophe, a prime or an accent; 0x8E, which is either i or ¡ (its dot is taller than on the i at 0x09 and its stem reaches one pixel below the baseline, which suggests ¡); 0x7E, a short bar at the top (a macron or overline, not ASCII’s ~); 0x9D, a g with an accent that looks like a circumflex (ĝ) but could be a breve (ğ); and 0xA9, which looks like ₠ and may stand for the euro sign.
  • Cannot be read (?): one cell, 0x6B. The table shows the Unicode replacement character (a black diamond with a question mark), so the character was lost when the table was put on the web.
  • Empty in the source: 23 cells that the table leaves blank, so no character is known for them.
  • Not in the data yet: 0xC0 to 0xFF, which the screenshot does not show.

In total: 192 codes, of which 159 are read with confidence, 9 are uncertain, 1 cannot be read and 23 are empty.

The data is in site/src/data/protocols/fis-charset.json, written by hand; docs/data-model.md describes the fields, and npm test checks the file.

  • Send each code to the display and photograph the result, starting with the uncertain cells, the empty cells and 0xC0 to 0xFF.
  • Check whether the duplicate codes listed above look different on the display.
  • Check whether 0x8E shows i or ¡, and what 0x27 and 0x62 show.
  • Check whether a B7 cluster, and a lowline cluster, use the same table.

Parts in the wiring data

CodeNameModelsLocation
J285Control module with indicator unit in instrument panel insertB6 and B7Behind display unit

Sources

  • Audi ASCII table (janssuuh, 1 April 2016) — janssuuh
    The whole table. Original not retrieved (Dutch blog janssuuh.nl, URL unknown); read from a screenshot of an English machine translation supplied by the owner, so the screenshot could not be checked against the original.
  • Known CANID's (janssuuh, 18 March 2016) — janssuuh
    The example frame 54 45 53 54 20 20 20 20 ("TEST") on 0x261 to 0x267.