I wanted to know how much current our PSOC6 MCU is consuming.
Our supplier has sent us this document with locations where I can remove resistors / inductors that are in series and then I can solder my fly leads.
Yes
The down side is that it's a bit intrusive. Depending on how skilled you are, you might damage the pads, the components, or the board, unsoldering and re-soldering wires and components.
You will need to remember to connect your ammeter before you power up the board, some MCUs could break if you power some rails and not others.
Removing components and putting an in-line ammeter will give you the most accurate results.
An alternative is to remove the inductors, and replace them with low value resistors, and measure the voltage across those. This has the advantage that you're not breaking the supply with your ammeter. Also if you switch ammeter ranges while powered up, this could unintentionally break the supply to the MCU. It also has the advantage that you need only one voltmeter to measure all the rails sequentially.
Less accurate, but avoids using extra resistors, is to estimate or measure the residual resistance of the inductors, and measure the DC voltage across those. Many DMMs on the DC range will average the current properly, but if they are upset by any AC component, use an RC filter in front of them. It's no worse than the DC averaging that would need to happen properly with the in-line ammeter method. Note that the resistance will change a few percent with temperature. It's more than good enough to tell the difference between low and high consumption, for instance sleep, normal operation, and RF transmitter on.
Those beads are doing something for you: they're blocking digital noise from getting back into the rest of your system. If you can keep them in you're doing yourself a favor.
When faced with this kind of work I prefer to use a digital scope, as it will give more meaningful information than a DMM and make for less-intrusive probing. As a bonus, the scope can also do lots of nice math on the waveform.
I suggest a couple probing options with a scope:
To get better results, especially for low currents, I've sometimes used a lash-up of a good current sense amp (like the LT6105) in place of a diff probe to improve accuracy.
Admittedly this isn't the best, since the MCU standby draw can get really small. And good, accurate clamp-on DC current probes are heinously expensive. You can improve the sensitivity of current probes by using multiple turns through the jaws; however the added wire length begins to have its own influence.
The sense amp lash-up I suggested above is still better as it's less intrusive. You can get one for the LT6105 here, from Mouser.
Either way, your scope will give you dynamic information on what's happening with the current draw, with better resolution than a DMM. And you can correlate current activity with operation state by using other signal traces on the scope.
Could you use the ferrite's DC resistance in place of a sense resistor? Probably not. The currents you are measuring are small, and the actual DCR value isn't known. (Not to mention that its impedance varies with frequency, which is kind of the point.)
You can even try to measure the currents without replacing the coils (or beads) at all.
The coils/beads have a DC resistance which can be measured with a decent ohmmeter (could range from a few mOhms to a few ohms).
You could then measure the voltage drop across the coils/beads (which could range from µV to a few mV) and calculate the current from this.
This can give a reasonable indication of the average current consumption, but may not give a very high resolution of the current over time, as the coils/beads form a low pass filter for the supply, limiting the bandwidth of the measurement.
If an ADC is used to measure the voltage, it should have differential inputs, a high enough resolution, and allow a common mode voltage slightly larger than the largest supply to be assessed. An integrated PGA can be a very useful feature to increase the resolution of the measurement. Finally, of course, it's the sampling rate that determines how good the time resolution is. If you want to measure microsecond spikes in power consumption, you might want to use a fairly fast ADC.
