We’ve got the opportunity few months ago to test a Jouloscope on small session (let’s say, less than 10 minutes of recording), this is an amazing tools, and we’d like to buy some of them to complete our lab equipment.
But we need to use them in a specific way.
We’re designing ultra low power devices, which could stay as long as several days in a standby/idle mode, consuming few µA. Our devices wake up 20 to 30 seconds to perform their activity (where they can sink up to 1A peak current) before going back to sleep.
Thus, we need to monitor the current consumption of our devices powered on small primary battery during days.
Is the joulescope able to monitor such long session?
Is their any tradeoff, like reducing sampling rate?
Do we need to setup a specific desktop configuration to make this kind of measurement operating?
Yes, Joulescopes are made for long, uninterrupted measurements. You have a few options.
You can buy large, inexpensive drives. The JS220 produces about 1 TB per day at the full 1 Msps rate. Large drives are inexpensive these days, but 2 weeks is about the max. We designed our proprietary but open source JLS file format to support nearly instantaneous read, pan and zoom of these huge amounts of data.
You can specify a lower sample rate. For long captures, many customers are happy with 10 kHz, which reduces the data rate and file size by 100. We recommend using the command line rather than the UI for even greater reliability. For example:
python -m pyjoulescope_driver record --frequency 10000 --signals i,v out.jls
Some customers are happy with the statistics data (2 Hz by default) for long captures. We provide the downsample_logging which also supports recovery on power loss or host computer reset. We have had customers capture full product lifecycle over many months with this script.
While we work hard to make the Joulescope UI reliable, the Python scripts remove lots of code found in the UI including PySide6, Qt6, and Nuitka. Less code means fewer latent issues for even higher reliability.
As you suggest, 10kHz sampling rate is a good target, and much more than the sapling rate we use today (most of the time nearly, 1kHz with our current monitoring tool).