Anelok: first batch of changes from board 0 to 1

Werner Almesberger werner at almesberger.net
Mon May 12 06:34:48 EDT 2014


I made a first set of changes in the schematics. The most visible ones
are going from wheel to capacitive slider and switching from AT86RF232
(IEEE 802.15.4) to CC2543 (hopefully BTLE).

Most of the work is however related to power distibution. Note that I
removed the rfkill switch for now, but I'd like to hear some more
opinions on that.

Board 0 schematics:
http://downloads.qi-hardware.com/people/werner/anelok/tmp/anelok-20131005.pdf

Board 1 schematics (work still in progress):
http://downloads.qi-hardware.com/people/werner/anelok/tmp/anelok-20140512-wip.pdf


A bit more details:

- "upgrade" MCU from KL25 to KL26. They're almost identical as far
  as Anelok is concerned, but:
  - the KL26 is newer and may thus also have better availability,
  - some sources claim the KL25 isn't RoHS-compliant,
  - more pins are interrupt-capable. The downside of using these
    pins for interrupts is that one couldn't just switch to KL25
    in case of supply problems with KL26.

- replace wheel with capacitive "slider" sensor,

- replace AT86RF232 with CC2543 (and related circuit), as already
  tested in Y-Box,

- get rid of the "current throttle" design,

- remove dodgy reverse current protection, merge Vbat and Vsys, add
  a FET to diconnect the battery when USB power is available, and
  also add transient voltage protection across the battery,

- add 33 Ohm series resistors to D+/D-, as recommended by Freescale,

- power switches:

  - memory card FET probably wouldn't be able to cut off the card
    if battery voltage is low. Resolve this by adding a second FET
    to drive the other.

  - add such a two-FET power switch also for OLED.

- remove the test points "for future extensions", i.e., TP4 for a
  light sensor, and TP5-TP8 for a navigation switch,

- DISP_SCLK was on a pin that doesn't have access to an SPI clock.
  Swapped places with CARD_DAT2.

- add footprints for boost converter options AAT1217-3.3, AAT1217-1.2,
  and TLV61220

- document memory card holder options (and use of R4, role of CARD_SW)


MCU pins will have to be reassigned, but I can start this only after
rfkill and MCU package (cf. "puppet masters") are settled. I also
have to go over the test points and see if we need any more or can
remove some.

I'm also struggling with the decision of whether I should use FETs
in SOT-523 (1.8 x 2.2 = 4 mm^2) or in SOT-323 (2.4 x 3.2 = 7.7 mm^2)
packages. They're small but now have quite a lot of them (6, up from
2), so ...

The drawback of 523 is that Digi-Key lists only one manufacturer for
p-FETs of that size and only one not so great-looking alternative for
n-FETs.

Another option would be to use complementary pairs in SOT-563. Saves
even more space (Anelok uses them in pairs anyway), lots of choices,
even a little bit cheaper, but I don't have any at home, and with the
customs situation in Argentina, that's a bit of a problem ...

anelok.cmp (component to footprint mapping) isn't up to date yet and 
neither is the layout. Last but not least, I'll renumber all the
parts once the (draft) schematics are finished, to make things easier
to find.

- Werner



More information about the discussion mailing list


interactive