# Garbled Characters

The layout is right, the text is in the right places, and the characters are wrong - accented letters
turned into symbols, Cyrillic into line-drawing characters.

This is a **code-page mismatch**, and it is the single most common surprise when moving from a driver
to raw ESC/POS.

## Why it happens

ESC/POS text is bytes, not Unicode. A single byte means whatever the currently selected code page says
it means. Your application encodes the text with some code page; the printer decodes it with whichever
one it was last told to select. When the two disagree, every non-ASCII character comes out wrong while
ASCII looks perfect - which is why the problem often hides until the first accented name.

## The fix

Select the code page explicitly at the start of every job, and encode your text with the same one:

- `ESC t n` selects the page, where `n` is the printer's identifier for it - `0` for CP437,
  `17` (`0x11`) for CP866, `46` (`0x2E`) for Windows-1251.
- Then encode your strings with that same page in your application, rather than sending UTF-8.

Do not rely on the printer's power-on default, and do not assume it survives between jobs.

## Confirming what was sent

Switch on **Raw Data** for the document. Each element is shown with its raw bytes next to the parsed
command, and text is decoded with the code page in effect at that point - enough to tell a wrong
`ESC t` from text that was encoded wrongly before it ever left your application.

If the document contains no `ESC t` at all, that is the answer: your application never selected a page.

## Still stuck?

Email [support@virtual-printer.online](mailto:support@virtual-printer.online) with the printer's
protocol and port, what your application sent, and what you expected to see.