nrf52833 standby current reaches 60uA

[justin0407]

First, NRF52833 work environment at 100 degrees celsius, standby current reaches 60uA.

Second, When the NRF52833 working in 100 degrees Celsius, what kind of processing software or hardware needs to be done.

Third, When the NRF52833 dormant, how to shielding the red signal in the frame? As is shown.

[Sigurd Hellesvik]

Hi Justin

First, NRF52833 work environment at 100 degrees celsius, standby current reaches 60uA.

See Electrical Specifications on Sleep for expected current consumption based on temperature.
What is your standby current in room temperature?

Second, When the NRF52833 working in 100 degrees Celsius, what kind of processing software or hardware needs to be done.

As far as I know, no special hardware or software is needed for operating the nRF52840 in 100 degrees.

Third, When the NRF52833 dormant, how to shielding the red signal in the frame?

I do not think I understand what you mean by "shielding". Can you say question in another way?

Regards,
Sigurd Hellesvik

[justin0407]

First, about 2uA standby current in room temperature.

Second, I'm using the nRF52833, can you tell me is the difference between nRF52833 and nRF52840 in 100 degrees Celsius environment?

Third, Change the way, why there will be a current pulse in this picture when standby?About 16 hz frequency and frequency varies with the change of environmental temperature changes.

[hmolesworth]

The refresh frequency is directly related to the reservoir/smoothing capacitance fitted to the regulator output. Assuming ceramic capacitors, the capacitance already should be derated 50% from the required value, so a 3.3 volt output requires a 6.3 volt capacitor.

With increase in temperature, the capacitance further decreases (by as much as 30% depending on capacitor type). A 30% decrease in DC-DC capacitance would result in an increase in frequency by a similar amount as the refresh tries to maintain a fixed mean voltage with a faster decaying capacitor voltage, which likely explains the figures you are seeing. There are capacitors which would have better temperature performance, or alternatively choose a larger capacitance value. Tantalum capacitors are entirely unsuitable in this application due to very high leakage. Also the internal leakage of the nRF52833 rises with temperature, and there is no control over this but it can be measured by testing an isolated part, crudely put swiftly heat the package of the nRF52833 alone to 100 degrees while not heating the rest of the board and components.

[justin0407]

Hi hmolesworth,

Refresh the frequency with which a capacitor?Should use what kind of material of capacitor?As shown in figure.

Regards,

Justin

[hmolesworth]

The 4.7uF or higher regulator capacitors are the key, these should be temperature stable and ceramic. Here are 3 example charts, select the expected variation (these are worst-case).

X5R

X7R

X7S

A sample data sheet is shown here avx. Typical manufacturers are Murata and Kyocera. Remember ceramic capacitance should be de-rated 50% from the required value, so a 3.3 volt output requires a 6.3 volt or higher capacitor.

[justin0407]

I try to increase the capacitance value, but has no effect on refresh rate and.

Refresh rate should only be affected by the MCU internal capacitance?

Under the environment of 100 degrees Celsius, standby power consumption is short of 15 ua manual, actual it is around 60 ua.

[hmolesworth]

Did you increase C6? C6 should affect the refresh rate.

.. and did you test this:

"the internal leakage of the nRF52833 rises with temperature, and there is no control over this but it can be measured by testing an isolated part, crudely put swiftly heat the package of the nRF52833 alone to 100 degrees while not heating the rest of the board and components"

[justin0407]

Sorry, I don't have such test equipment, can you help provide this data?

I modified the C5 and C6 respectively, but the refresh rate remains unchanged.

[justin0407]

Differences between the MCU version produced?I'm using the MCU of screen printing is N52833-QDAAA0-2117FX.

[hmolesworth]

If changing C6 doesn't affect the frequency then we must assume the nRF52833 uses a hysteretic controller with a fixed frequency and periodic charge time fixed by Vnom+-Vhyst. This fixed frequency is presumably some RC-oscillator, the details of which we do not know. The likely cause is then the RC components which are used to determine that frequency, however the very wide range reported seems improbable.

It is not clear what other components are on the board, or what power supply is used. Does the battery voltage vary with temperature? Testing a bare nRF52833 with no other components other than the required regulator C and R is probably the best step; preferably on a board where all other components are removed. Nordic could probably do this test but maybe you could test with your board. I only help out from time to time, if I get time I will try a local heat test but can't give a timescale.

Also internal leakage affects the 1.3 volt internal rail, whereas external components load Vdd; we don't know what effect that has on the hysteretic frequency.