sparkfun txb0104 level sifter and nrf52

[Matts]

Hi.

We are trying to get nrf52 to work with 1.8v spi master. We are using level sifter with it. However when we connect the nrf52 to level sifter to spi output, the whole thing pretty much explodes logic vise. The sparkfun says this level sifter cannot be used with other than push-pull-type signaling. Is your spi not this type?

EDIT: with default SPIS lines the MISO is still crap:

#define SPIS0_CONFIG_SCK_PIN         2
#define SPIS0_CONFIG_MOSI_PIN        3
#define SPIS0_CONFIG_MISO_PIN        4

There the blue is clock and yellow is the MISO line. The picture is from nrf side of the level sifter. MOSI is OK. In below blue is again clock and yellow is MOSI.

[bjorn-spockeli]

SPI peripheral on the nRF52 is activly driven, i.e. Push-Pull, so this should not be an issue. Are you able to provide a logic trace/capture of the SPI lines? If so could you provide a trace of the SPI input and output to/from the level shifter.

I also found this in the comment section on the Sparkfun product page(here is the link) :

One problem I ran into–and it is a problem with the IC, not with SparkFun’s breakout board–is that the IC will oscillate if the load on any output is too heavy. If you find that is the case, add a buffer IC downstream of the output.

-Bjørn

[Matts]

I think the problem is the master or level sifter here. I measured the loose connector again (so just nrf side signal), and I saw nice looking square wave response on miso. However with the level sifter attached the signal clearly doesn't pull it low enough. The signal was OK, but it didn't go low enough. The master line is miso, so I don't think that's the issue, but could it be that the nrf doesn't have enough pull to actually drive the chip correctly? I mean this is level sifter is meant for SPI.

TI spec says +-2mA is required for driving the cirquit.