Re: [PATCH v4 1/6] dt-bindings: adc: ad7173: add support for ad411x

[Jonathan Cameron <jic23@kernel.org>]

On Wed, 5 Jun 2024 09:54:31 +0300
"Ceclan, Dumitru" <mitrutzceclan@gmail.com> wrote:

> On 03/06/2024 23:00, Jonathan Cameron wrote:
> > On Mon, 3 Jun 2024 12:46:10 +0300
> > "Ceclan, Dumitru" <mitrutzceclan@gmail.com> wrote:
> >   
> >> On 01/06/2024 21:35, Jonathan Cameron wrote:  
> >>> On Fri, 31 May 2024 22:42:27 +0300
> >>> Dumitru Ceclan via B4 Relay <devnull+dumitru.ceclan.analog.com@kernel.org> wrote:
> >>>     
> >>>> From: Dumitru Ceclan <dumitru.ceclan@analog.com>  
> 
> ...
> 
> >>>> +          Supported only by AD7172-2, AD7172-4, AD7175-2, AD7175-8, AD7177-2:
> >>>> +            19: ((AVDD1 − AVSS)/5)+
> >>>> +            20: ((AVDD1 − AVSS)/5)−    
> >>>
> >>> That's what it says on the datasheet (so fine to copy that here) but I'm curious, what does
> >>> that mean in practice?  How can we have negative and postive signals of the difference
> >>> between two power supply voltages where I'm fairly sure AVDD1 always greater than AVSS.
> >>>    
> >>
> >> I have not tested that as I do not have a model that supports this wired up.
> >> If I had to guess they are the same signal but one should be connected to the
> >> positive input, one to the negative input...but I could be wrong.  
> > 
> > If they are, then as far as I we are concerned is this one channel with two
> > representations depending on whether it is 1st or 2nd in the list?
> > Can we use one number and hide that detail in the driver?
> > 
> > Seems odd though if that is the case.
> > 
> > I guess if we find out later this is the case we can tighten the binding to
> > enforce the right one instead of squashing them to one value, but that
> > is a bit ugly.  Any chance of digging out the info?  If not we can go ahead
> > but ideally answering things like this make a our life easier in the long run.
> > 
> > Jonathan
> >   
> 
> "(Avdd1/Avss)/5+ as positive input and (Avdd/Avss)/5- as negative
>   this is used for monitoring power supplies, the inputs must be selected in pair"
> Perhaps it's an internal voltage divider...? I dunno
> 
> So it seems like this cannot be used as a common mode voltage input.
> I'll restrict the driver to only allow these inputs paired together
> and rename the define for these selections.
Most mysterious :)  I'd be interested to know what value it reads
back if you ever get the part.

Ah well, great to have gotten that extra detail even if it leaves
more questions!

Jonathan

> 
> 
>