QSPI clock waveform

[Poplar]

Hi

we configured a QSPI application on nRF52840.

its clock was assigned on p0.04/J1 in hardware design, now we notice that its waveform is close to a square wave as 30MHz maximum, but changes to similar as a sine wave if it is set to 32MHz.

does this mean p0.04 can't be run as such high frequency usage? or only the recommended usage is allowed for QSPI application?

Thanks

[Håkon Alseth]

Hi,

 

You should use the dedicated QSPI pins for QSPI operation:

https://infocenter.nordicsemi.com/topic/ps_nrf52840/pin.html?cp=5_0_0_6_0

Please note that you must also configure H0H1 (high drive) on QSPI pins:

https://infocenter.nordicsemi.com/topic/ps_nrf52840/gpio.html?cp=5_0_0_5_8_1_9#register.PIN_CNF-0-31

 

Kind regards,

Håkon

[Poplar]

Hi Håkon

the schematic is fixed now, we can't change to other pins, high drive mode is set already.

if we switch the p0.04 to 32MHz, not under QSPI but a simple IO, will the square wave still distort as a sine wave?

we didn't see p0.04 marked as some limitation as digital IO if take the reference from datasheet page 159. is there other information we shall take into consideration too? Or what is the difference between p0.04 and p0.19?

Thanks.

[Håkon Alseth]

Hi,

Q1: Can you share how the signal looks?

Q2: How large is the trace from nRF to QSPI-flash?

P0.04 isn't limited in terms of switching frequency, but QSPI has dedicated recommended pins:

 

The overall capacitance, when you probe the pin, will also change. I would recommend that you use a low-capacitance probe to ensure less further loading of the signal. 

 

Kind regards,

Håkon

[Poplar]

Hi Håkon

SPI_CLK, p0.04

SPI_CS, p0.06

SPI_D0, p0.26

SPI_D1, p0.08

SPI_D2, p0.22

SPI_D3, p0.20

overall trace length is 8.8mm

if we use the IOs which are not dedicated QSPI ports, is the easyDMA/DMA function also impacted?

thanks

[Håkon Alseth]

Hi,

 

Poplar said:

if we use the IOs which are not dedicated QSPI ports, is the easyDMA/DMA function also impacted?

The module will work as intended on other pins. See here: https://devzone.nordicsemi.com/f/nordic-q-a/47151/qspi-pins-in-52840/186409#186409

Poplar said:

overall trace length is 8.8mm

This should be short enough.

 

Can you share how the signal looks wrt. rising/falling edge?

 

Kind regards,

Håkon