I2C timing spec

[peanutbutter]

Tsu_dat (300ns min) and Thd_dat(500ns min) are more stringent than  NXP I2C spec.(Fast mode)

Since most IC vendors design their I2C bus with NXP I2C spec, they are not meeting the nRF spec.  Does nRF work reliably with NXP I2C spec Fast Mode spec? Thanks

[Håkon Alseth]

Hello,

 

Tsu_dat (300ns min) and Thd_dat(500ns min) are more stringent than  NXP I2C spec.(Fast mode)

 Isn't this the other way around? We guarantee that the setup time is extended to allow slow I2C slaves to get "extra time" to prepare themselves for a command. If we had a minimum setup time (t_su;dat) that was lower than 100 ns, then we could have created a compatibility problem with slaves that require more than 100 ns to prepare its bus interface.

 

Since most IC vendors design their I2C bus with NXP I2C spec, they are not meeting the nRF spec.  Does nRF work reliably with NXP I2C spec Fast Mode spec?

 We have not gotten feedback, that I know of, that we do not work with certain devices.

We have a theoretical place where we can potentially break the I2C spec from NXP, where we can go under the minimum "tlow" defined parameter for fast mode:

https://infocenter.nordicsemi.com/topic/errata_nRF52840_Rev2/ERR/nRF52840/Rev2/latest/anomaly_840_219.html#anomaly_840_219

 

Kind regards,

Håkon

[peanutbutter]

I'm confused. The Tsu_dat is 300ns min so anything smaller 300ns is a failure. eg. That indicates that a Tsu_dat of 150ns will fail the NRF spec but 150ns is greater 100ns of NXP I2C spec.