sandeepmistry / arduino-nRF5

Arduino Core for Nordic Semiconductor nRF5 based boards
Other
892 stars 276 forks source link

Taida_Century_nRF52_minidev - pin mapping #200

Open micooke opened 6 years ago

micooke commented 6 years ago

First up - the README reference no longer has a valid link. Im planning on hosting the schematics i was provided, so i can provide somewhere to point to if needed. For reference, it seems to also be called 'Gold Core' at this point in time on aliexpress.

Regarding the variant.h pin mapping by @carlosperate and @d00616, was the pin mapping for tx, rx, i2c and spi arbitrarily chosen?

I ask because the 12 pin header exposes tx and rx, which are connected to either;

I currently have RGZ soldered, but either option creates a conflict with the SDA, SCL pin mapping.

Is there a relationship between the pin allocation of TDI/TDO and MOSI/MISO on the nRF52? If so then one of RGZ or RSM may be preferable as the remaining two jumpers on the reverse side map TDI/TDO.

dlabun commented 6 years ago

We probably shouldn't have been linking to AliExpress in the first place, but then again I can't find a manufacturer page for these modules.

carlosperate commented 6 years ago

Yeah, I don't see a better alternative either than the online shop product pages. It'd be great if somebody could archive the data somewhere a bit more dependable, so that would be quite helpful.

About the pin out, not sure where it originated from, @d00616 would be better position to know that answer. It's possible that similar modules might have slightly different pin outs, @micooke are you certain your module was the same @micooke had when https://github.com/sandeepmistry/arduino-nRF5/pull/47 was introduced?

micooke commented 6 years ago

@dlabun - I agree, and no i haven't found a manufacturer either. This is why i might just put up some info and schematic in my github.io repo for archival purposes.

The schematic i have claims its a 'CC2640RSM/RGZ Module Test Board V1.0' from 'rf-smart.taobao.com', which looks like another store.

micooke commented 6 years ago

Hi @carlosperate - im not 100% sure, but if i google 'taida century nrf52832' the first result is an aliexpress link (that is no longer available) which looks identical to the board i have, down to the silk screen.

If you search aliexpress for 'nrf52832 gold core' (2 results only) that's what i have. It is a nrf52832 module soldered to a 'CC2640 test module' (aliexpress - 1 result), which matches a schematic i have (and will host and post tomorrow). I have seen modules for the CC2640, nrf52832 and nrf51822 that match this test module/breakout board.

The 2xbutton and 2xled bindings are identical to the Taida pin mapping.

micooke commented 6 years ago

See https://micooke.github.io/nRF52832_TaidaCentury_GoldCore Files are located here : https://github.com/micooke/micooke.github.io/tree/master/nRF52832_TaidaCentury_GoldCore

d00616 commented 6 years ago

The reason for the layout is, that if haven't realized the existence of the jumpers. My documentation is at https://github.com/RIOT-OS/RIOT/wiki/Board%3A-nRF52-minidev

micooke commented 6 years ago

@dlabun - I was thinking that i can accomodate for the original pin mapping, plus the pin mapping enabled by soldering the RGZ or RSM configuration by using board_variant.

The changes i have made locally are a bit of a hack. They set compile-time defines using {build.lfclk_flags} (as there is only one static option for this), based off the board_variant, which are used in variant.h to change the pin mapping.

I couldn't use extra_flags as this is set by the softdevice chosen, and there are two options.

ec1oud commented 6 years ago

From the RIOT wiki, we have

MCU pin Board pin Function
P0.11 D18 RXD (software defined)
P0.12 D19 TXD (software defined)

and that works for me, but on my board as received from AliExpress, none of those RGZ/RSM jumpers are soldered, so the signals don't get through to the pins on the end connector. And @micooke you're saying there's no way to jumper them such that there isn't a conflict with other signals, so maybe that's why they left them open? @d00616 is yours that way too?

micooke commented 6 years ago

@ec1oud - this pin mapping is the same as the original. The solderable link bring out some different pins to the 2x6 as listed on the silk screen on the bottom of the board. The original configuration has 2,3 for i2c pins which conflict with the serial pins for rgz, rsm so i deconflicted by moving these pins to 7,8 when the rgz or rsm configuration is used. No more conflict :)