High Current Draw on SNVS_IN (part 2)

[mjbcswitzerland]

Hi All

I would like to make a part 2 to this thread: https://community.nxp.com/t5/i-MX-RT/High-Current-Draw-on-SNVS-IN-with-imxrt-1062/m-p/1370981

I am rather worried about the current consumption on battery backed up mode, which is specified to be typically 20uA.
My EVKs measured about 60uA and prototypes between 40uA and 60uA.
Today I did the following with a couple of new i.MX RT 1062s:
1. They were originally not soldered into a board so I carefully applied 3.0V between the BGA balls VDD_SNVS_IN and a GND pin. The current was about 50uA
2. I then had two i.M RT 1062s soldered into their boards with NO other components (this ensures all GND pins are connected together). The currents was about 40uA on one board and about 50uA on the other (the same as in 1, when that chip not soldered in).

Therefore from my experience with EVKs and chips on and off boards (and 1051, 1052, 1062 based designs on custom boards with chips from different batches) the current (and thus the batter backup current - with and without RTC running) has always been between 40uA and 60uA.

This shows an average value of around 50uA with deviation of +/-20%.

This is of course not in accordance with the data sheet's typical 20uA (with no variation details).

What does this mean? Is the spec too low or is there something missing in the measurement or circuit details? Is there some way to reduce it and achieve the claimed value by ensuring other pins are set to a certain state?

Thanks in advance

Regards

Mark

 

[lucas_cao]

Hi,

For EVK, you may need to remove some resister to get 20ua.

You can check the following AN, chapter 4.1.2 Hardware rework for the EVK

https://www.nxp.com.cn/docs/en/application-note/AN12245.pdf

[mjbcswitzerland]

Hi

I am interested in battery backed u RTC and some memory and so the VDD_SNVS_IN current in battery backed up state if of concern. Even if resistors are removed on the EVK I don' get much below 60uA (3x the typical value according to the data sheet), which corresponds with other measurements with chips that have absolutely no other connections (and thus no 32kHz oscillation either) - with the spread between 40uA and 60uA with the batches I have.

Since I need 5 years battery life I have concluded that it is not practical and now foreseen an external RTC with memory so that I can reduce the battery backed up current to below 1uA and easily achieve 5..10 years of battery backup with a much smaller battery. Another advantage being that the RTC in the i.MX RT 1062 is not so accurate since its calibration/compensation is of limited use since it is not fine enough.

Basically all investigations have concluded that the data sheet typical value is too optimistic and the spread rather large, so a 2'000mAh battery would be the only way to ensure to a good degree of confidence achieving 5 years, but these are large, heavy and expensive.

If anyone can actually find a chip that has 20uA current when only its VDD_SNVS_IN  and GND are connected to 3V I would be interested to hear about it. In my case I also used multiple current meters to ensure accuracy (all gave the same reading to better that uA).

Regards

Mark