ADP5091/92 Supercapacitor Selection

I am designing a very small 4 layer wireless device (34.25 mm x 20 mm) where the ADP509x is powering an STM MCU from a source like indoor solar. The small size means that space and weight are a big factor for component selection. Height of components is especially important as there are many practical advantages to having the device largely limited to 2 dimensions.

The storage element must be a supercapacitor and the MCU runs on 3.3V down to 1.8V (mandatory brownout reset at 1.7V falling edge). I have configured SETPG for 3.3V using 6.98/3.01 MΩ with the understanding that this should set SYS for a target of 3.3V to power the MCU.

My understanding is that during normal operation the BAT storage element will effectively match SYS when charged, i.e. VBAT = 3.3V.

As mentioned, I am trying to limit the physical size of components and naturally the supercapacitor is the biggest. I have come across the DSK-621 (0.33F) from ELNA which states "Max. Operating Voltage" of 3.3V (and lacks other information such as leakage, which is not good). This is by far the best available size for the capacitance however I am wary of the voltage rating.

Could the DSK-621 be used with the expectation that the input voltage to the supercapacitor can be safely maintained at 3.3V or lower at all times with appropriate selection of the TERM threshold?

If that supercapacitor cannot be used what would be the minimum voltage rating required for a 3.3V maximum?

The DSK-621 also states 200Ω internal resistance which seems high. Could this be an issue for performance in a very low power harvesting scenario?

Thanks in advance for your time.

  • 0
    •  Analog Employees 
    on Oct 30, 2020 3:12 AM 1 month ago

    The supercap should be okay during charging if you parallel with a 10uF capacitor at BAT pin but there ma be problem if the operation current of the system is large. 

    Regarding the charging protection, you can set TERM threshold about 3.3V*95% to keep margin. So you should set a lower SETPG threshold. 

    Btw, have you tried the entire system? If you connect the MCU at SYS pin with PGOOD as the enable pin. How about the shutdown current or standby current when PGOOD is low? if the current is large, it will take longer charging time. 

  • I have not tried the entire system as the MCU is not yet available for purchase (STM32WLExx series). This device is being designed for my honours thesis so I haven't had time to try things like the evaluation boards for the 5091/92 and instead need to rely on scouring the datasheet.

    If SETPG was lower could I still get 3.3V at SYS? I'd like to maintain the higher voltage as it directly relates to RF transmit power.

    I'm using REG_OUT for powering the analog subsystem of the MCU but at 200mA it would be enough to supply the entire device. If SYS cannot be set to 3.3V while using the 3.3V supercap, could the regulated output be kept at 3.3V instead?

    The MCU will be generating RF transmit signals with around 120mA input current (at 3.3V) but there's a bunch of internal subsystems and decoupling inbetween so I doubt the SYS pin would be experiencing large transients. These would be momentary and the typical power consumption would otherwise be very low. The MCU also has its own LDO and SMPS which feed the RF power amplifier.

    I'm not sure if the MCU can be enabled/disabled with PGOOD as it has battery backup capability for maintaining things like RTC. Currently the battery backup pin on the MCU is tied to VDD which is being supplied from SYS on the 5091.

    I have considered using PGOOD to trigger the "wakeup" pins on on the MCU to exit the backup power state but I feel like it may just be better to leave it in a powered shutdown state for very low power consumption to maintain internal memory and program operation.

  • 0
    •  Analog Employees 
    on Nov 3, 2020 2:09 AM 29 days ago in reply to Diozoh

    TERM is to set the terminal voltage of SYS and SETPG is just to set the threshold of PGOOD based on SYS voltage. To be noted, REG_OUT can only support 150mA and the voltage cannot be higher than SYS. for the RF transmission, you may need keep the BAT pin not lower than SETSD or else the supercap would be disconnected with SYS pin.  

  • To confirm, if using the DSK-621 there is no way to reliably go above 3.135V (i.e. 95%) to supply the MCU at 3.3V?

    Would it work to make SETSD 1.8V = minimum for MCU?

    I could route PGOOD to one of the MCU inputs and use it as a signal (when LOW) to enter deep sleep modes for reduced power consumption. What voltage level would you recommend for SETPG and the hysteresis range in this case?

  • 0
    •  Analog Employees 
    on Nov 3, 2020 9:39 PM 28 days ago in reply to Diozoh

    As the reference of the TERM threshold has variation, I suggest you keep 3% as a margin, 3.3V*97%.

    A lot MCU can work from 1.8V to 3.3V so you have a lot of energy can be used. And I would suggest set SETSD to 1.9V.

    Regarding PGOOD, it depends on your application. If your input power cannot supply your system always on, you need set a higher PGOOD with large hysteresis so that you can use more energy for the processing and radio transmission. For example, set PGOOD to 3.2V and hysteresis like 500mV or more so you can use the energy from 3.2V to 2.7V of the supercap.