Open jps2000 opened 5 years ago
awesome, thanks!
ad 1. The vdiv with higher values is indeed present in the 52832 board. however such high impedance voltage dividers need capacitor to gnd for proper adc operation ( done in the 52832 board) --> suggest to add capacitor 1..10nF in next 52840 layout and change voltage divider to 2M + 2M. ad 2. 2M2 also works fine. The penalty is that inadvertently touching the pins with fingers can lead to a short shut down. So this may be a problem with modules lying around on the table. 220k or 470k may be a compromise. ad 3. I see no need for a 0.2V voltage drop as the diode just supplies the 3V3 low drop voltage regulator. So I actually use just a 1N4148 instead of the Schottky. However current consumption is then limited to 150mA. It is likely less robust to shorts. There are also some more capable SOD123 Si diodes but they are somehow rare.
With above changes I achieved <10uA (~7uA) off current (with EN = LOW). Maybe helpful for some users.
true it does not need to be a schottky since, but we need to supply up to 500mA peak to the regulator, can you suggest a diode that can do that? (in the same package)
There are these SL1 A... M types from Diotec for example they claim <1uA@ 25° leakage may be considered.....
seems to be actual https://forums.adafruit.com/posting.php?mode=reply&f=57&t=146360
(2): I made a current measurement on my Feather nrf52840 with an external power supply of 3.6 V on the BAT pin. I found 100 µA and I don't understand (although not critical for my project). 100 kOhm on the ENable pin: 36 µA. Battery voltage divider, 200 kOhm: 18 µA.
(4): There is a neopixel LED, ref.: WS2812B. The Quiescent Current is of 0,5 mA when the LED is off.
https://lowpowerlab.com/forum/general-topics/ws2812b-rgb-led-measurements/.
This should explain why I get a current of 600 µA at rest with a loopless sketch.
One should note that on the ItsyBitsy, there is DotStar LED with ref. APA102-2020. It is also having a gargantuan Quiescent Current of 0,8 mA.
https://memcpy.io/apa102-led-current-usage.html
The latter may be easier to unsolder, considering the SWD connector position on the Feather one. But, overall, it seems to be a big drawback to low power management of booth boards. Again, the Feather nfr52832 would be more suitable.
FWIW, I can confirm that...
...brings the total current draw down to 20µA. 18µA are still wasted for the VDIV (no need to verify that and I'm not particularly keen on replacing the 4x100k array), the remainder is likely 1.5µA for the Raytac module and 0.5µA reverse leakage from the STTH1R02ZF.
true it does not need to be a schottky since, but we need to supply up to 500mA peak to the regulator, can you suggest a diode that can do that? (in the same package)
The STTH1R02ZF can handle up to 1A with a max. 0.5µA reverse leakage. The AP7354D seems to be fully compatible with the AP2112.
@mhx thank you for being so helpful :) we'll check those parts out!
@mhx STTH1R02ZF has very high Vf, 0.75 - can you try MBR054? https://www.digikey.com/product-detail/en/on-semiconductor/MBR0540T1G/MBR0540T1GOSTR-ND/918576
WRT the voltage divider, there's a full article from Nordic Semi on how to design a high resistance divider for the nRF52. It's really straightforward and the only additional component needed is a capacitor as already mentioned by @jps2000.
@mhx STTH1R02ZF has very high Vf, 0.75 - can you try MBR054? https://www.digikey.com/product-detail/en/on-semiconductor/MBR0540T1G/MBR0540T1GOSTR-ND/918576
I don't have one of those lying around, but I can put a handful on my next digikey order and give it a try. Why are you concerned about the higher voltage drop? There should be plenty of headroom going from 5 to 3.3V.
USB can droop down to 4.25V when current spikes - at 1A Vf 0.9V and we also need some headroom for the LDO at spikes
The AP7354D seems to be fully compatible with the AP2112.
Actually, it's not — just had another look at the datasheet — it'll only output up to 150mA. :/
I see this thread is over a year old. Any chance an updated rev of the nrf52840 will have the lower current mods?
@ladyada I tried the mods with the AP7354D and MBR0501T1G and got down to 29 uA. 150mA is plenty for my application. Would really be nice if the production board didn't need hardware mods to be a 'low energy' BLE peripheral.
Just looked over the schematic of rev D To achieve lowest Vbat and OFF current pls consider following changes:
1) Make the VDIV divider same as in the nrf52832 board (high ohmic + add capacitor!) It takes now 20uA
2)EN Pullup use 1M or even more to allow low power shutdown (EN pulldown is 3M , >1.5V is required It takes now 40 uA
3) Replace D1 MBR120 by a Si diode which has less leakage current It takes now 25uA to 200uA Thanks