Confusing traces of CW305

Embedded Security ChipWhisperer Hardware
1

Hi,

I am a newer for CW305, and I’m using it for some single-trace attack experiments.

However, I found, in CW305 , the same data and operation leads very different power samples. So, I tried to observe how it works when the targeted FPGA is unprogramed. A trace I observed is shown as the following image:

图片
图片2505×1207 253 KB

the blue trace is sampled from CW305.

In my experience， the power trace should be close to flat at this point rather than fluctuating like this. I don’t understand what causes this.

So, HELP!

Thanks.

2

Hmm, a few questions:
1- How are you powering the CW305?
2- How are you measuring the traces?

Do you have any ChipWhisperer capture hardware? If so, it would be much easier to diagnose that way, since it would eliminate a few variables.

3

Thank you for your help.

  1. I’m powering the CW305 via the USB-B connector.
  2. The trace in the figure above was acquired with an Pico-5000 oscilloscope connected to the X4 SMA connector.
  3. Yes, I do have a ChipWhisperer capture board, however, it has similar problems (below)
    图片
    图片1231×877 77.1 KB

    图片
    图片1231×890 108 KB

    The spikes at the very beginning of the waveform are the AES example projects followed by an empty run phase where you can see that the waveforms from the two acquisitions have a non-overlapping undulation. If the leakage of my attack target is weak, this undulation will seriously affect the utilization of the leakage of the side channel.
4

It looks like fluctuations are caused by induction from power lines (50/60Hz). Probably you have different power supplies for your target board, PC, CW… You have to solve GND connection problems. When GND for different boards connected at different points, you have a small AC current flow between those connections