Issue #3367 Changed delay calls to avoid conflicts with qspi_disable

[DavePutz]

Two calls to common_hal_delay_ms in common_hal_displayio_fourwire_reset() were changed to use common_hal_mcu_delay_us instead. This avoids a problem where qspi_disable was being called in the middle of the reset, which ended up disabling the SPI screen display on the Clue. This also fixes SPI TFT displays being used on other nrf52840 boards as well. It does not fix a similar issue with e-paper displays, which will need to be addressed separately.

[tannewt]

Sorry, I understand this fixes the issue but I'd really like to understand the root cause.

The only way I imagine the reset failing is if the pulse is shorter than it should be. We should be able to verify that it is too short with a logic analyzer or oscilloscope. I'm worried the root cause is causing other issues.

[DavePutz]

Some info regarding power consumption with and without this PR. Data was gathered using an Adafruit INA260 power sensor running the peripherals script from @xiongyihui
6.0.0-alpha-3 (SPI not working)
1.026 mA 1.275 mW
6.0.0-alpha3 with nrf errata disabled (SPI working)
2.174 mA 6.619 mW
6.0.0-alpha-3 with this PR (SPI working)
1.075 mA 1.333 mW

[tannewt]

Thanks @DavePutz. The root cause search ended in an errata so this is the best fix. More info on the issue: #3367 (comment)