0

I'm working on an IoT project based on BLE (2.4 GHz) and SoC: nRf52832, and at this point, I need to do some power consumption measurements.

After going through some previously built similar projects, most of them use expensive oscilloscopes, which I can't afford as a hobbyist. I'm looking for something under $200-250.

The maximum and minimum current that could go on the PCB is from a few mA to a few uA. As for bandwidth and sampling rate, I got some ideas after doing some search on Google.

I'm still not sure how it's related to my circuit. For example, I'm using a crystal oscillator of 32.8 MHz, so do I need bandwidth more than that? Like 40 MHz+. How is the frequency relevant to the BLE antenna circuit on the PCB?

enter image description here

My other concern is the current measurement as low as uA and the waveform on the oscilloscope over the time, to measure how long the battery will last.

I also found I need different probes for the current measurement to measure the same (mentioned above) under the same price, if possible. Please suggest to me some good oscilloscope/HHD.

P.S.: I have gone through all the below-mentioned question links and a video. Most of them are 5-6 years old and even more.

  1. Opinions on an oscilloscope suitable for hobby electronics on a budget
  2. Oscilloscope Bandwidth, what's it all about?
  3. EEVblog #451 - Rigol DS1052E vs DS2072 Oscilloscope
  4. What to look for in an oscilloscope?

Oscilloscope:

  1. 100MHz Digital Storage Oscilloscope - ADS1102C : US $439.99
  2. DSO Nano V2 - Pocket-Sized Digital Oscilloscope : $99.95
  3. Rigol DS1052E : $314.99

Edited : I need this kind of waveform with measured currrent value over the sleep cycle.

Use a Timer as a wake-up source This is how to wake up the ESP32 that consumes the least power. Only the RTC Timer is on. It is a counter that triggers an alarm (like an alarm clock) whose period is configurable. enter image description here Consumption of 11.5µA in timer mode powered by 4.2V

Just doin Gods work
  • 361
  • 1
  • 8
  • 4
    An oscilloscope is not the right tool for power measurements. A bench multimeter would be better suited for that. – Lars Hankeln Dec 26 '22 at 08:29
  • @LarsHankeln We don't need to see the current waveform over time to measure power?If some multimeter can do that, please tell me which one is best suited for my requirement. – Just doin Gods work Dec 26 '22 at 08:35
  • <<< current measurement as low as uA >>> An INA219 should do it, with the change of the measurement resistor. – Antonio51 Dec 26 '22 at 08:37
  • @Antonio51 How do I measure it over a duration of time? As the MCU sleep and wake cycle is 60 sec. – Just doin Gods work Dec 26 '22 at 09:23
  • I used it to measure the current every 20 ms (minimum time sampling data for INA219). Then I made a "summation" to get average results every second (V, I, P, Energy, estimated Time of use of a given battery, and output all data through WiFi for saving on disk. – Antonio51 Dec 26 '22 at 10:48
  • 1
    Take a look at a dynamic measurement power supply, such as the HP 66312A, which is designed for this job, and can measure from amps to microamps - as others have said, you may not need an oscilloscope. – jayben Dec 26 '22 at 10:54
  • While a scope is very useful in general, for characterizing MCU power consumption a simple multimeter is cheaper and usually more sensitive. The main thing a scope does is let you see microsecond by microsecond transition power consumption, which can be interesting but is usually less useful for optimization. – user1850479 Dec 26 '22 at 17:20
  • Multimeters are generally not suitable for measuring tiny amounts of current. An oscilloscope also isn’t. I bought a Joulescope for this purpose, cheaper alternatives exist. Watch this clip: https://youtu.be/LUB8RWzzLWc – RemyHx Dec 27 '22 at 06:25
  • If all you need to measure is the energy consumed, then a current sense amplifier followed by an integrator is all you may need. Readout on a DMM. – Kuba hasn't forgotten Monica Dec 27 '22 at 07:48

3 Answers3

4

Nordic Semiconductor makes a "Power Profiling Kit", part number "nrf-PPK2" that probably does what you are looking for. And, it's actually available (!) at Digi-key, Mouser etc for USD92.

Daniel Chisholm
  • 209
  • 1
  • 3
  • 9
2

The current consumption waveform you show as an example doesn't need a high bandwidth or a high sampling rate.

Several hundred samples per second would be fast enough for that kind of plot. At those speeds, a datalogger would be enough.

Pretty much any digital oscilloscope would be fast enough for that task

More important is that the scope needs to be able to hold a lot of measurements so that you can be sure of capturing the entire power cycle. Again, any modern digital oscilloscope should be up for it.

  1. Use an appropriate adapter to convert current to voltage. This may be a purpose bought device or a device you build out of a current detector IC. A current transformer (CT) won't do because you are dealing with DC rather than AC. Transformers work on AC.
  2. Export the captured data to CSV or Excel or whatever format the scope's software offers that you can read into a spreadsheet (or other program) to do some math on it.
  3. Convert the voltage readings from the scope to current readings - the math will depend on how your adapter works.
  4. Sum the current over several complete cycles and calculate an average current per second (or hour or day - whatever unit suits your needs.)

Which scope is best for this depends more on the software than the hardware.

Which scope you buy depends on what other tasks you may use it on in the future.

The other questions you've linked to give advice on picking a scope based on the frequency ranges you expect to encounter.

JRE
  • 67,678
  • 8
  • 104
  • 179
0

I think you want to know only a long-term average of power consumption for evaluating battery life. A true RMS DMM with 2-3 measurement/sec is totally fine. Measuring low voltage (in 100mV range) with an oscilloscope on a shunt will be very noisy anyways.

novakt62
  • 51
  • 2