Experimental Physics and
Industrial Control System

In fact, setting EOFF did the trick, now the conversion is happening as expected.

Eric, I trusted the top Google result because the URL looks fine and suggests, at least to me, that it's the official documentation for the current version. I will take more care in the future :) I believe it would be more clear if a warning would be put on top of that old page, saying that the most up to date documentation can be found at the wiki.

Dirk, the 'type' variable was already defined in a private structure, so it agrees with your suggestion. I had created a custom function (not specialLinconv()) to setup those conversion parameters. Is there any special reason to use specialLinconv() instead of a custom function? I thought that EPICS would call specialLinconv() every time a conversion was needed (for, say, weird conversion rules), apparently that's not the case. 

special_linconv(precord,after)

This routine is called whenever any of the fields LINR, EGUF, EGUL or ROFF is modified. To support linear conversion, EOFF and ESLO must be set accordingly. The record support sets EOFF to EGUL before calling this routine, which is the very common case when RAWL is zero below.

A useful formula for calculating EOFF and ESLO is this one:

EOFF = (RAWF * EGUL - RAWL * EGUF) / (RAWF - RAWL)
ESLO = (EGUF - EGUL) / (RAWF - RAWL)

Here, RAWL and RAWF are the lowest resp. highest possible raw value. For instance, a 16 bit bipolar ADC might have RAWL=-0x7fff, RAWF=0x7fff.