11uA to 30 uA low power pulses at 4ms intervals.

RE: 11uA to 30 uA low power pulses at 4ms intervals.

PeterInStillwater — Tue, 16 Jun 2020 14:06:33 GMT

Yes - That's how I'm triggering the 1 seconds pulse. Im sure there's room for improvement there as well.

This project is a port from the 51 to 52. In the 51 code there are multiple references to being able to directly control the power - for example:

    NRF_UART0->POWER  = 1; // Universal Asynchronous Receiver / Transmitter 0
    NRF_SPI0->POWER   = 1; // Serial Peripheral Interface (Master) 0
    NRF_TWI0->POWER   = 1; // Two Wire Interface (Master) 0 Note: will not work without enabling this
    NRF_SPI1->POWER   = 0; // Serial Peripheral Interface (Slave) 1
    NRF_TWI1->POWER   = 0; // Two Wire Interface (Slave) 1
    NRF_SPIS1->POWER  = 0; // Serial Peripheral Interface 0
    NRF_GPIOTE->POWER = 1; // GPIO Tasks and Events
    NRF_ADC->POWER    = 0; // Analog-to-Digital Controller
    NRF_TIMER0->POWER = 1; // Timer / Counter 0
    NRF_TIMER1->POWER = 0; // Timer / Counter 1
    NRF_TIMER2->POWER = 0; // Timer / Counter 2
    NRF_RTC0->POWER   = 1; // Real Time Counter 0
    NRF_TEMP->POWER   = 1; // Temperature Sensor
    NRF_RNG->POWER    = 1; // Random Number Generator
    NRF_ECB->POWER    = 1; // Crypto ECB
    NRF_AAR->POWER    = 1; // Accelerated Address Resolver
    NRF_CCM->POWER    = 1; // AES CCM Mode Encryption
    NRF_WDT->POWER    = 0; // Watchdog Timer
    NRF_RTC1->POWER   = 1; // Real Time Counter 1
    NRF_QDEC->POWER   = 0; // Rotary Quadrature Decoder
    NRF_LPCOMP->POWER = 0; // Low Power Comparator

I've not been able to find reference to equivalent functionality in SDK16.  In 
the NRFX structures several/most of the "POWER" member.  

Peter

RE: 11uA to 30 uA low power pulses at 4ms intervals.

ketiljo — Tue, 16 Jun 2020 12:23:49 GMT

There's something that wakes up the CPU every 100 ms. Do you have the RTC running?

RE: 11uA to 30 uA low power pulses at 4ms intervals.

PeterInStillwater — Tue, 16 Jun 2020 11:56:33 GMT

Here are some screen shots of what I'm seeing. I mentioned the 4ms pulse above - I'm also interested in trying to understand the 100ms pulses.

Here are two one Second pulses where I wake up the timer: Note 10-100ms pulses in between (not sure what these are either).

Below you will see 100 ms pulses with 24-four MS pulses in between.

And here is the four MS pulses.

For Reference,

Thanks again - for your support,

Peter

RE: 11uA to 30 uA low power pulses at 4ms intervals.

ketiljo — Tue, 16 Jun 2020 06:53:20 GMT

Refresh mode is not the same as DCDC. The refresh mode is only in use when the device is on it's lowest power state and barely "kept alive". The way it works is that you have a capacitor that is charged quickly with a short pulse. The regulator senses the voltage on the cap and gives another pulse when the voltage drops below the threshold. This gives quite a lot of ripple on the regulator output but that doesn't matter as none of the power critical parts are on anyway. So you sacrifice ripple and load regulation for better efficiency.

RE: 11uA to 30 uA low power pulses at 4ms intervals.

hmolesworth — Mon, 15 Jun 2020 21:51:05 GMT

This is unexpected behaviour; I thought the automatic refresh mode was conditional on external components and therefore could be disabled (or rather not enabled before enabling the softdevice). The external components are optional:

Automatic switching between LDO and DC/DC regulator based on load to maximize efficiency
Note: Two additional external passive components are required to use the DC/DC regulator

Breaking execution with and without DCDCEN enabled shows that the softdevice does not enable the DC-DC converter; does this imply a pulsed LDO? How does that reduce power consumption?

RE: 11uA to 30 uA low power pulses at 4ms intervals.

ketiljo — Mon, 15 Jun 2020 17:56:14 GMT

The regulator is in refresh mode independent of LDO or DCDC mode

RE: 11uA to 30 uA low power pulses at 4ms intervals.

hmolesworth — Mon, 15 Jun 2020 17:29:26 GMT

Try turning the DC-DC converter off during the period when you want to avoid the current pulses, using instead the LDO (Low Drop-Out regulator) which doesn't have current pulses. The LDO is less efficient when running at full blast, of course.

  nrf_power_dcdcen_set(false);

RE: 11uA to 30 uA low power pulses at 4ms intervals.

ketiljo — Mon, 15 Jun 2020 16:49:58 GMT

No, that's not possible, it's part of how the regulator works. When you start to draw more current, it will switch mode to be more efficient for larger currents. The internal regulators are optimized for the lowest possible current draw depending on the state of the chip, how many peripherals that are enabled etc.

RE: 11uA to 30 uA low power pulses at 4ms intervals.

PeterInStillwater — Mon, 15 Jun 2020 14:06:06 GMT

Thank you Ketiljo - is there a way to disable the capacitor charge - put the regulator out of refresh mode? Or is this not a desirable condition? I would like to eliminate the pulses all together.

Thank you again!

RE: 11uA to 30 uA low power pulses at 4ms intervals.

ketiljo — Mon, 15 Jun 2020 12:24:46 GMT

What you see is the regulator being in refresh mode. When the regulator is in the lowest power mode, it will charge a capacitor in very narrow bursts at regular intervals. The width of the pules are just a few µs so it makes it a bit hard to capture and measure them correctly.

RE: 11uA to 30 uA low power pulses at 4ms intervals.

PeterInStillwater — Fri, 12 Jun 2020 20:53:50 GMT

P.S. I'm using the Otii to capture traces and make my measurements. I've been very pleased with the performance of this tool - I recommend checking it out ! https://www.qoitech.com/

Peter -
Stillwater Minnesota