5 days later

thanks bela_robert for the BOM, will place an order .
(also the instructions look excellent, thanks for putting so much effort into this!)

glad your enjoying the series ... hoping next one will be with Trill
going to order some more grove to grove cables at same time as ordering above BOM

odd thing...
Ive noticed (rev 1 pepper) has a DC offset e.g. the sine example, is about 3v peak to peak (as expected) but its centred around 1.5v (approx)
is this correct? i don't remember noticing it before?

Im going to test my other pepper which is in a different rack, but I need to hook up my scope to do this.

as I say odd... as Id have thought I would have noticed this before...

EDIT:
ok, so my other pepper is the same, and given its in a completely different rack, and Im measuring it with a different scope - I'll assume this is 'normal' - or I did the same thing wrong in both builds 😉

I think the reason Ive only just noticed is I was using a VCA to amplify, and the DC offset makes this distort and cut out when pushed too far... I think previously i was using an input module, which possibly removed the DC offset without me noticing (that'd mean I also would see on scope!)

I assume the above amp remove this DC offset? (its aim being to -5v to +5v centred on 0v)

is R49/R55 the biasing? i.e. if it wasn't exactly centred replacing with a trim pot might allow adjustment.
(but perhaps its not that relevant, a small DC offset is ok)

      thetechnobear Ive noticed (rev 1 pepper) has a DC offset e.g. the sine example, is about 3v peak to peak (as expected) but its centred around 1.5v (approx)
      is this correct? i don't remember noticing it before?

      Yes that's right: the line out is DC-coupled, and we thought it would be a feature 🙂

          giuliomoro DC coupled is indeed useful 🙂

          (I just can't believe i hadn't noticed previously ?!)

          the question is how does it map to the API? ,

          i guess we expect -1 to 1, to represent -ve, +ve , so 0 =0v

          imho , in this context - DC coupling, just means I can continually send 1.0f and get 1.8v (or whatever peak is) continuously output... if it was not dc coupled, this would not be possible, since it would fall to zero. regardless of my code 😉

          Im assuming this is not a code change, rather how the line out is biased?

          as above, I guess different in rev 2/one the above amp circuit is in place...
          (and I assume salt is -5/+5)

              The range you can achieve is 0.4V (sending out -1) to 2.2V (sending out +1).

              PS: this was somehow documented here,

              thetechnobear as above, I guess different in rev 2/one the above amp circuit is in place...
              (and I assume salt is -5/+5)

              Audio out on Pepper rev2, Pepper rev1 + amp board and Salt is the same: -5V sending -1 , +5V sending +1.

                thetechnobear Im assuming this is not a code change, rather how the line out is biased?

                Yes this is it's just how the line out comes out of the codec (plus a series resistor for protection)

                  In fact, that DC-coupled line out is available on both the Bela cape (balanced, on J9) and BelaMini (unbalanced, on the 3x1 socket header).

                  20 days later

                  Is anyone experiencing a similar problem with cv outs? I get lots of switching noise when hooking up to ms20.

                  You could try to add a capacitor in parallel with the jack output. The series resistance from the Bela's DAC and the output pin is 970ohm (220 on the Bela cape and 750 on R36-R43. A 100nF capacitor between the jack's tip and GND would give you a cutoff of 1/(2*pi*970*0.0000001), that is 1.64kHz. This should reduce noise drastically (clearly at the expense of bandwidth). Here's an example drawing of how to install the capacitor (in red) at the jack socket's terminals (with my extra-poor drawing skillz). Capacitor is most easily mounted on the back of the PCB.

                  alt text

                  9 months later