Experimental Physics and
Industrial Control System

"Dalesio, Leo" <[email protected]> · Tue, 2 Dec 2008 19:01:23 -0500

ESLO and EOFF were meant to characterize the transducer.
ASLO and AOFF are adjustment slope and offset that are meant to correct for calibration errors in the ADC board.
As I recall......

-----Original Message-----
From: [email protected] on behalf of Andrew Johnson
Sent: Tue 12/2/2008 4:46 PM
To: [email protected]
Subject: Re: Device Driver Initialization routine

On Tuesday 02 December 2008 13:45:59 David Dudley wrote:
> According to the 3.14 WIKI on the ai record, the ESLO and EOFF fields
> aren't configurable through the DCT.  I take it that that's wrong.  Should
> that be changed?

Yup, that's wrong; it changed when we added the SLOPE conversion mechanism, 
which was needed because some generic device support layers can't know the 
range of their raw values so can't implement special_linconv().  I just 
changed the ai and ao pages, but be aware that there are *lots* of other 
things still wrong in the wiki Record Reference Manual.

> I assumed I couldn't do anything with it, so I didn't try, and used ASLO
> and AOFF instead, however looks like I need to correct some things.

I believe those were originally intended to permit adjustment of individual 
I/O records to correct for hardware issues, where the actual conversion 
results differ from the design parameters.  You can of course use them 
however you wish though.

> Also, if you are using LINEAR, do you have to have a special_linconv()
> routine in the driver in order to get values scaled correctly?

Not necessarily, as long as ESLO was properly set in init_record() or the .db 
file.  If a special_linconv() routine is provided it will be called whenever 
LINR, EGUL or EGUF are changed and LINR is LINEAR; its sole purpose is to 
compute ESLO and set EOFF if its value is not EGUL.  If you never change EGUL 
or EGUF (or you don't set LINR to LINEAR) there's never a need to recompute 
ESLO, so not providing a special_linconv() routine would have no effect on 
the value scaling.

- Andrew
-- 
Talk is cheap. Show me the code. -- Linus Torvalds

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System