I currently have a chipwhisperer lite and from my experiments I can generate a glitch with a maximum width of 170mhz which corresponds to 5.8823529411765 ns using the CLKGEN output. However, this is probably not the maximum speed available for the spartan-6 gpio. Therefore, I wanted to ask how it looks in the case of the chipwhisperer husky in the individual descriptions I found some information > Clock Generation Range
10-350MHz - is this its absolute maximum that it can throw on the fpga gpio port as a glitch?
Regards
The scope API on our readthedocs page explains how the glitch is shaped. Make sure you look at the Husky-specific glitch section (which is different from the CW-lite/pro one).
The Husky Glitching demo notebook walks you through the glitch settings.
The glitch shape resolution is much higher than 350 MHz.
If I want to test my glitch sensor, I set the glitch width to 1, clock source to PLL with 1200MHz, then I will get a less than 1ns voltage spike?
Is there requirement on cables used for ns level glitch output (what does it mean- depends on cabling used for routing glitch output)? @jpthibault
It depends on your target clock; the step size is the target clock period divided by scope.glitch.phase_shift_steps, which is in turn controlled by scope.clock.fpga_vco_freq.
This will be the resolution of the digital signal that controls the glitch MOSFET’s gate. The shape of the actual voltage glitch that you get will depends on things like the glitch cable’s properties and the target itself. You can observe the glitch with an oscilloscope.