PPK2 test peak current inaccuracy

[sunmnengnan]

   When testing the power consumption of the product with main IC 52832, we found that under the condition of 0dBm 1000ms, the peak current of beacon with PPK2 is about 15mA, while the data manual is about 5mA, we asked XUNTON (Nordic supplier found that they use the power analyser to test the TX peak in the vicinity of 5mA and said that the peak current is not permitted when the sample rate of PPK2 is adjusted to the maximum, and the peak current is permitted when the sample rate is adjusted to the second higher gear). maximum, the measured peak current is not accurate, when the sampling rate is adjusted to the second highest gear, the peak current is accurate). I wonder know is that right？What sample rate should we use to test the average current and which sample rate should we use to measure the peak current? Why is the difference in measurement results between ppk2 for a higher sampling rate and a power analyser greater ?

①average current

②One pulse amplification

[Simonr]

Hi

This very much looks like the phantom spikes you can see in current consumptions where the current draw changes rapidly. This is because of the level shifters in the PPK2 aren't able to reduce the voltage spikes as well as expensive oscillators are, so very brief high spikes can be seen. They are only caught at the 100,000 samples per second sampling because they are so brief, so lower sampling rates aren't able to catch them. They should be so brief that they don't affect the total average current, but they are indeed present. You can open the advances settings on the Power Profiler app with ctrl+shift+alt+A to try and tweak the voltage spikes parameter to mitigate this. The simpler fix will be to just use the lower sampling rate.

Best regards,

Simon

[sunmnengnan]

Hi

Does it mean that the sampling rate is adjusted to 10,000 times per second and the peaks captured are incorrect? Should I just go with what the power analyser picks up?

Thank you for your reply.

[Simonr]

Hi

They won't be incorrect, but with less samples it will of course catch less of the super short peaks and valleys of a measurement. The average will still be approximately the same, but you will catch fewer of these voltage shifted spikes that you also see the power analyzer mitigates in HW with stronger components.

Best regards,

Simon

[sunmnengnan]

I see，thank you！

[sunmnengnan]

Is the peak current picked up at 10K sampling rate closer to the actual value?

[PaKa]

The sampling rate inside the PPK is always the same. When you change the timebase in the user interface the data is averaged before sending over. This is what "removes" the spikes as they will be averaged out. When using the spike filter, the average will still be the same, it just removes the spikes from the display to make it easier to view.

TLDR; the peak current with the 10K will be OK as long as the pulses are long enough, if they are really short then the peak may be too low due to averaging. The average will always be OK independent of the sampling range. If you need better peak current detect then you may try to use an oscilloscope.

[sunmnengnan]

Thank you for your answer.

So should the peak transmit current of the chip at 0dBm be 15mA or 6mA?

[PaKa]

Depends on a lot of parameters. If you are running 3.0V it should be about 7mA with the CPU in Idle and about 9mA with the CPU active. This is with the DCDC enabled. If you are not using the DCDC then the current will be almost double. If you reduce the operating voltage then the current with DCDC will go up while the current with LDO will stay more or less the same.

infocenter.nordicsemi.com/.../pmu.html

I recommend checking out the online power profiler  Online Power Profiler for Bluetooth LE as that can give you some ideas on what to expect for average current. The peaks are averaged out here as well so may be lower than what you can see with an oscilloscope