This simple function will remove any non-ASCII character. Feel free to fork and extend!

convert_ascii.php

<?php
/**
 * Remove any non-ASCII characters and convert known non-ASCII characters 
 * to their ASCII equivalents, if possible.
 *
 * @param string $string 
 * @return string $string
 * @author Jay Williams <myd3.com>
 * @license MIT License
 * @link http://gist.github.com/119517
 */
function convert_ascii($string) 
{ 
  // Replace Single Curly Quotes
  $search[]  = chr(226).chr(128).chr(152);
  $replace[] = "'";
  $search[]  = chr(226).chr(128).chr(153);
  $replace[] = "'";

  // Replace Smart Double Curly Quotes
  $search[]  = chr(226).chr(128).chr(156);
  $replace[] = '"';
  $search[]  = chr(226).chr(128).chr(157);
  $replace[] = '"';

  // Replace En Dash
  $search[]  = chr(226).chr(128).chr(147);
  $replace[] = '--';

  // Replace Em Dash
  $search[]  = chr(226).chr(128).chr(148);
  $replace[] = '---';

  // Replace Bullet
  $search[]  = chr(226).chr(128).chr(162);
  $replace[] = '*';

  // Replace Middle Dot
  $search[]  = chr(194).chr(183);
  $replace[] = '*';

  // Replace Ellipsis with three consecutive dots
  $search[]  = chr(226).chr(128).chr(166);
  $replace[] = '...';

  // Apply Replacements
  $string = str_replace($search, $replace, $string);

  // Remove any non-ASCII Characters
  $string = preg_replace("/[^\x01-\x7F]/","", $string);

  return $string; 
}
?>

Jay,

Thanks for this. I have spent the last couple of days looking for a clean way to sanitize some data before sending it up to a web service that triggers an error every time it encounters a non-utf-8 character. I'll fork it if I need to add any extra characters.

-Matt

Thanx dude, it saved my day :)

YOU DA MAN!!!!

Jay you are the man, I have been struggling with content I have been pulling from a third party vendor and this sorted it out. THANK YOU!

Nice.

Thanks.

Very nice and thank you

that was helpful, thank you.

Thank you very much, so useful! Exactly what i needed :)

Finally found what I need to solve my problem, very nice, thank you so much

Thanks for your sharing.

:-D when you re-discover the only good solution the internet has to offer, ty

Another great solution that will cover every edge cases (é will be turned into e, etc)

iconv('UTF-8', 'ASCII//TRANSLIT', $string);

(from https://stackoverflow.com/a/3542748/1822742 )

@paulintrognon, that's correct, however if you have some invalid UTF-8 input, it will error out. So that only works with valid unicode.