Power consumption features on oscilloscope
Hi all
I've been investigating the power consumption of the AudioMoth Dev board for my masters and have built a current-sense circuit to look with an oscilloscope. Here's what I'm getting at a 48kHz sample rate during recording. The SD card I'm using is the Samsung Evo Plus.
From what I understand, the large square waves (repeating every 340 ms) are the current drawn to write data to the SD card flash memory, but I can't figure out what the smaller, more frequent peaks are (see zoomed-in photo below). They seem to repeat every 2.7ms (~370Hz), which is too slow for clocking speed, sample rate or power switching and too fast for LED. I'm thinking SD card memory refresh, writing to SRAM , or some other component? Any advice on other features (such as which peaks correspond to the LED) or related information on power consumption would also be much appreciated.
Potentially different SD cards would move the whole line up as well as resulting in a higher square wave. Between the SD card writes the SD card is powered up and enters a standby mode (under its own control when no more data arrives). The consumption of this standby mode may vary from card to card. The different SD cards might then also have different write speed, and different write currents, resulting in peaks of different widths and heights. The SD cards are quite non-determininistic as they do lots of internal load balancing and delete and write pages autonomously. I think that's what shows up as the spikes in the square wave - the AudioMoth is doing the same thing continously over this period - sending data to the SD card in chunks of 512 bytes - but the SD card does a lot of internal work.