Pictured above is an Anker brand LiPo battery pack for use in charging USB devices. Although Anker is an Amazon phenomenon, their products stand out in that they aren’t garbage from Shenzhen. They’re actually very well engineered pieces of hardware. Qualcomm Quick Charge enabled, and with 2x DC outputs for prototyping stuff (as I do), this thing rocks. 2000mAh battery capacity too. Pretty beef for being small. I use it to do power supply experiments with development boards.
This is a Portapow brand (Amazon again) USB power monitor. They work very well, although I’ve received a few broken. I think that’s Amazon’s fault, though, not the company itself. Their charge cables are beefy and pretty well made. Their products just seem to work. Also it gives you it’s own specifications for it’s power drain so you can account for that in your experiment. Super slick. It’ll monitor up to 20v DC and 4A of current, although its more suited to 2A under constant load.
Then you’ll need a dev board to test, like the Raspberry Pi seen above. It’s got it’s IQaudio DAC onboard, but not enabled in software as we will be booting from scratch to take a measurement of how many mAh the unit takes to run an update process after first boot, with a DAC, of course. Why? Just for the hell of it.
I’ve cleared the memory on the USB power supply monitor, and written a fresh boot image to SD card for the Pi, which has serial terminal enabled in the /boot/config.txt file. That way we can see exactly how much energy the process of booting, running updates, and configuring the device to use the DAC takes us from the first miliamp of power used. This is very useful information: if you plan to go set up a Shairport-sync client at a friend’s house, and all you have is a USB battery, this test will tell you how big a battery will be suitable to the purpose. If you write python or develop ARM software from scratch, this same procedure can be used to see how much energy a newly minted product would be using to boot and get an Over The Air (OTA) update.
With the Pi plugged into the USB monitor and the monitor plugged into the battery, we’re ready to power up. I’m using an FTDI programmer and a cable I’ve prepared to use UART 1 for a serial comms terminal. I’ll be using Putty on my PC to talk to the Pi and monitor it’s progress during updates and I’ll also be configuring the DAC the same way.
As I run the apt-get process and update the firmware (after configuring wifi and country info, of course) I watch the amount of current bounce around in pulses as the microprocessor does it’s calculations and operations. Microprocessors tend to demand large pulses of current versus a smaller comparable amount of current over time at idle.
Usually the idle current is roughly .5A, but during updates for the software and firmware, the device pulls a peak of roughly .98A. Couldn’t get a photo cuz I was too slow on the draw. After 14 minutes of time (Gawd-Dayum!) the Pi is configured with the Fake KMS to run the DAC card, had all of it’s firmware updated, run all of it’s software updates, and is ready to rock out. This is considerably less time than the original Pi series took to run updates without 5G wireless and without the upclocked ICs (chips.) Less time to run updates translates to less power used overall (I have to test that! Cuz I don’t actually know that.) 148 mAh was used in total during all of the updates and reboot cycles. That includes the power needed to supply the DAC board.
I hope this mundane test cycle has helped you understand a little bit about one way to put a device under test for power supply requirements. This is not the most useful example. A better example would be for me to create a Linux image with my custom configurations and software on it, and to test that package for boot and update requirements. This can be a critical part of product design, wether it’s a product for you and your friends in a recording studio or a product Designed For Manufacture.