Measuring the Raspberry Pi Pico W's Power Consumption - Workbench Wednesdays
ฝัง
- เผยแพร่เมื่อ 3 มิ.ย. 2024
- The Raspberry Pi Pico W microcontroller board is the latest addition to the Pi Pico family. This new MCU adds a Wi-Fi and Bluetooth module to the existing board’s RP2040 microcontroller. With those additions, James got curious about the power consumption of the board.
In this video, he walks through how to make the Pi Pico W’s LED blink, how much power it consumes, and then you'll see an IoT example using MQTT. Ask him questions and find resources on the element14 Community at this link! bit.ly/3QkT9OV
An intro to the Pi Pico W on the element14 Community: bit.ly/3QqeSop
Buy the Nordic Semiconductor kit James uses for these tests: bit.ly/3TQjQhw
James's tutorial on profiling battery usage for IoT devices: bit.ly/3mKHvAu
James's introduction to MQTT: bit.ly/3qgUuLU
Engage with the element14 presents team on the element14 Community - suggest builds, find project files and behind the scenes video: bit.ly/3tmdewv
Visit the element14 Community for more great activities and free hardware: bit.ly/3q6YMpu
Tech Spotlights: bit.ly/3qPrDhM
RoadTest and Reviews: bit.ly/3pV5Bux
Project14: bit.ly/31wbnJY
#0:00 Welcome to Workbench Wednesdays
#1:02 Blink An LED
#3:20 MQTT Example
#5:22 Sleep Modes
#6:40 Sleep + MQTT
#8:22 Give your Feedback
#raspberrypi #raspberrypipico #rp2040 #raspberrypiprojects #micropython #mqtt #iot #internetofthings - วิทยาศาสตร์และเทคโนโลยี
Excellent overview, and I'm also happy to see sleep modes can make battery-powered applications a lot more reasonable. I would love to see if the C SDK is any better on sleep modes, and/or if you could run the same/similar code on an ESP32 device and get a comparison.
I would say Micropython is quite good actually. Pico data sheet (page 11) shows that the lowest power consumption that pico can achieve is around 0.8mA at 5v with its dormant mode. And the sleep mode is around 1.3mA.
Hey Jeff. I would like to share my experience with ESP32 and sleep modes. In my tests, ESP32S2 mini power consumption was around 180uA in deep sleep. However, there were spikes up to 200mA when was connecting to WiFi. Still a better performance than Pi Pico W in terms of power consumption. I was able to keep ESP32S2 mini with an oled display and bme680, running for 3.5 days on an 18650 which had attached an always on LED (SMD). The ESP was waking up every 10 seconds, then sending the readings to blynk and went back to sleep. With 30 seconds or 1 minute in deep sleep, I guess running time without charging, can be extended up to 6 days. I understand that ESP32C3 is even better regarding the power consumption but I didn't had a chance to test it yet. Pico W lasted 1.9 days in the same conditions, so I don't find it great for battery power applications in my case.
Compared with both, ESP32S2 mini and Pico W, the Microbit V2 was capable to run for 5 days on 2 AAA batteries. That was without any sleep mode, displaying time, date, temperature every minute (except night time) and with BLE always on. Therefore, is probably the best for some battery projects as it's extremely capable and it has a few on-board sensors with a decent accuracy.
However, I resume to ESP32S2 mini for most projects due to wifi capabilities, power consumption and its size.
Pretty cool! I may have to get a couple of these to play with.
Very nice! Thanks for the video. Really help!
Interesting stuff, thanks for that. I wonder if batching transmission would be effective (like Android devices do). DO as much work as possible when the wifi is starting and active.
Amazing video! Wondering if it’s possible to power on Pico W using 18650 battery and how long will it last on deep sleep mode
I couldn't find a companion video for the non-W version. Is there one? I think that comparison could be useful as well.
(I hate IoT dependence although I'm not averse to Wi-Fi or cable connections. I do prefer the device to be fully self-contained when possible. At least this can then give an overall run-time consumption.)
I did some simpler measurements on the temperature monitoring project a few weeks back with a regular Pico. 22mA basic, with 24mA with the LED on. Got that down to 4mA by changing the clock speed down to 10MHz and a little lower by turning off the ADC and USB. Couldn't get the C SDK example to work with my code but that would have enabled the same deep sleep that James mentions. Have seen some other TH-camrs get down in the 1mA range with the SDK example code.
thanks for great video .. sad thing is that 1mA in deep sleep is still far above usable value for battery operation
In a sense I am happy about trying the original Pico first. I power it from the RPi 3B which CAN complain (a power bank solution).
Great Video, I bought one of these PPK II devices and can’t find much newbie into on how to connect it to a pico.
You briefly mention “it power pins” at 1:21 but it’s not clear what you actually did.
I got mine working in “Current Mode” where the PPK powers the pico but that does not use the USB cable so I can’t launch different programs during testing.
What connections to the pico should I use in “Ampere Mode”?
All the examples I find online use Nordic proprietary boards. I don’t want to fry something on my first go 🤠
Thanks
David
"Ampere Mode" is the same idea as a DMM in current mode. You break the circuit so it can measure the current.
To use the PPK with a PIco, you have to power the Pico FROM the Pico in Source Measurement Mode. If you're connected via USB then the Pico will get power from USB and bypass the PPK.
@@bald_engineer wow speedy reply! Perhaps you made a typo “power the Pico from the Pico in Source Measurement Mode”
Should the second “Pico” be PPK?
If that’s the case then source mode is my only option. So for each test I do I need to upload my program as main.py disconnect the USB and then power the Pico up from the PPK. I was confused how the PPK was going to measure anything when the pico was apparently connected to two power sources.
Have I got it right now?
(I’m an old dog learning new tricks) 🤠
Amazing savings
Do you think reducing supply voltage to 3v (instead of 5v) will reduce/increase power consumption?
Slightly reduce.
It seems no matter what value I enter into the deepsleep function- it just restarts immediately. Any ideas?
What if you have a factory full of PLC's, think about the savings by the "simple" sleepfunction....
In the real world, vs just running all the time, how much of a time extension do you get, on, say a 10,000 mAh battery pack? Hours? Days? Weeks? Interested in something like that for a mail box monitor...
In the last example graph going by the average current draw of 40mA a 10Ah pack works out to 250 hours. Assuming 80% efficency from a pack that size I would expect around a week of usage before charging is needed. If you change the sleep time to around 10 mins for a mailbox then I would imagine you could get a good month out of it assuming you have healthy cells in your powerbank that don't discharge quickly. You could probably use a small 12v solar panel with a small bank of supercaps and a buck converter and run the device indefinitely! Sounds like a fun project
In your mailbox idea the issue I see would be keeping the CPU active to register the "mailbox open" event. I guess you could keep the CPU active during times you would expect mail and only powering the WiFi module when the event is registered. If the CPU consumes 18mA for 8 hours a day then we can assume an average of ~6mA. 10Ah @ 80% efficiency = 8Ah. So 8000mAh/6mA = 1333 hours. So maybe 6 weeks is possible? I think I need to play with this idea (edited as I calculated for 8mA instead of the actual 18mA the CPU consumes)
@@-Graham that is definitely the way to go. You can configure the Pico to wake on interrupt and then put it into deep sleep.
@@-Graham @Workshopshed Interesting. The interrupt idea sounds intriguing, along with supercaps and a small solar charger, as it gets cold here (friggin' cold, like -20F/-30C) and hot (hellishly, like 105F/41C) regularly, and I think that has killed a battery or two for me. The interrupt might be a physical switch or light sensor indicating the door to the mailbox has been opened, telling the PicoW to 'fire up, send the message, and go back to sleep. Thank you both for the suggestions.
have a look into the "esp8266 01s" .... add a physical latch mechanism that fires up the device when the mailbox is open along with having the device in deep sleep mode when the code is complete and you just made yourself a device that can in theory run for a year off a x4 pack of AAs .... GPIO is very limited though.
wow, quite small sip for wifi.
iirc esp32 eat hundreds of mA during wifi tx.
This is the best video i've ever watched.
And the first source of data for power consumption on the Pico W.
Thanks!
CYW43439 TX active mode 300 mA
3.6 V
Deep sleep use 1-2 mA? That's a lot. I though it would be someting in the order of 1 microamps. Why so high? Whats wrong with the board? Next time round numbers correctly - I expected better from an engineer!
There is a full schematic for the Pi Pico in the hardware-design-with-rp2040 data sheet. There is not much to it. A crystal, regulator circuit the RP2040 and a memory chip.
The datasheet suggests 0.18mA just for the RP2040 chip in deep sleep but then suggests that it could be as high as 4.2mA depending on the design of your board. So not a system that has been designed for ultra low power.
My guess is that MicroPython isn’t putting the RP2040 into its dormant mode. And I don’t know what the flash memory draws. The WiFi module probably draws a bit. Maybe one the regulators is not shutting down. Plenty of reasons.
@@bald_engineer flash memory should be uA in standby if I have read the datasheet correctly. Regulators can make a lot of difference
mqtt is not light weight. And if it’s so popular, why isn’t it used more?