Re: Wait for mux to settle for PCI-DAS 1002

From: Calin A. Culianu <calin_at_ajvar.org> · Date: Tue, 11 Mar 2003 12:27:57 -0500 (EST)

Not that I am complaining.. I am all in favor of the driver doing the
right thing with respect to mux settling times... but I always thought
Dave's philosophy was to allow the application programmer to choose
whether he wants to wait or not (in order to avoid performance degradation
due to the usleep() calls)....?

-Calin

On Tue, 11 Mar 2003, Ivan Martinez wrote:

> 	Hello Lorenzo,
> 	Thank you for the bug report. I guess adding the delay won't harm anybody, so
> we better add it and see if there any complains. Let me make the patch and
> send it to David, so I update my e-mail address too.
> 	Greetings,
> 	Ivan Martinez
>
> On Tuesday 11 March 2003 12:47 pm, Lorenzo Dozio wrote:
> > Hi,
> >
> > I'm working on an industrial project using RTAI-Lab with Comedi support
> > to manage analog I/O of a ComputerBoards PCI-DAS 1002. I noticed that
> > the driver (cb_pcidas) misses a waiting time for multiplexer to settle.
> > This lack involves wrong readings with multiple consecutive
> > comedi_data_read calls on different AI channels. You can test the
> > outcome simply by adding other comedi_data_read to the "inp" demo under
> > Comedilib.
> > The right behaviour is retrieved by adding a delay of 5 us after the
> > writing to the ADC Channel MUX and Control Register.
> >
> > --- cb_pcidas.c.orig    Tue Mar 11 12:40:52 2003
> > +++ cb_pcidas.c Tue Mar 11 12:41:07 2003
> > _at__at_ -815,6 +815,8 _at__at_
> >                  bits |= SE;
> >          outw(bits, devpriv->control_status + ADCMUX_CONT);
> >
> > +       udelay(5);
> > +
> >          /* clear fifo */
> >          outw(0, devpriv->adc_fifo + ADCFIFOCLR);
> >
> > I don't know if this modification affects other boards supported by the
> > same driver. Anyone can test it?
> >
> > Ciao, Lorenzo.
> >
> >
> > _______________________________________________
> > comedi mailing list
> > comedi_at_comedi.org
> > https://cvs.comedi.org/cgi-bin/mailman/listinfo/comedi
>
>
> _______________________________________________
> comedi mailing list
> comedi_at_comedi.org
> https://cvs.comedi.org/cgi-bin/mailman/listinfo/comedi
>