[alsa-devel] Wolfson 5102 on Beaglebone Black using ALSA

[Charles Keepax]

On Wed, Jun 15, 2016 at 09:14:39AM -0700, Clark Dunson wrote:
> Hello Charles, Peter, and Mark
> 
> Another progress report, some feedback, and a question.
>
> We have everything pretty much working now.
>
> The clocking configuration I identified in an earlier post,
> (ie. MCLK1 begin fed from BBB McASP0 24.576 MHz, and BCLK
> being derived from it) works, though we still have a patch to
> arizona-core.c to make it so.  Ready for any inputs on how we
> can make this change so we can upstream it.
>

This week is very busy for me let me put together some stuff and
get back to you next week on this one. I have some changes that
add the stuff required but it probably needs to be reshuffled so
we can get it upstream in a reasonable time frame.

> Selecting single-ended mode vs. differential mode being a
> pairwise choice (ie. one selection for IN1, not independently
> for L/R), caught us by surprise and required a harrowing
> ECO.  (This selection appears to also not be supported by the
> driver) In the end we had to abandon single-ended mode, as

The wlf,inmode device tree entry lets you select between
differencial and single ended.

> the PGA acted in a manner that defied our understanding.  To
> wit:  If we had a signal connected to IN2LP, (AC through 1uF
> capacitor), then what we did with IN2LN was pivotal.  If we
> connected I2LN to ground, the signal would have large positive
> offsets, eg. +0.5 on WAV scale of 1.0 to +1.0, with unity
> gain set on PGA, Digital Volume, and mixer.  Cranking the PGA
> gain would rail the signal in a positive direction before we
> had adequate gain.  If we left IN2LN floating, we got closer
> to zero-centered, but when we set PGA to +6dB, we got huge
> negative offsets (-0.6 on WAV scale) and still we were dead in
> the water.  Finally we connected a 1uF capacitor to 1N2LN, and
> set back to differential mode (default).  Life is better now.

Alas I am afraid you have exceeded my understanding of
electronics here, you would need to discuss this with one of our
hardware FAEs. I do know we have used both single ended and
differencial in the past, but I am at a loss to explain what you
are seeing there.

>
> In our status meeting yesterday, we concluded that our moving
> forward with thousands of units would be benefited both on
> our side and Wolfson’s by an FAE as a go between.  Is your
> Cupertino office still alive (we are in Bay area), or can we
> get a point of contact?  Is it Joao?

Yes Joao is probably a good place to start, and we do still have
the Cupertino office.

Thanks,
Charles