<@U016EM8L91B> any idea what might be causing this...
# mpw-2-siliconm
Matt Venn
03/02/2023, 2:42 PM@Tim Edwards any idea what might be causing this?
t
Tim Edwards
03/02/2023, 2:56 PMWhich circuit are you measuring, and what were you expecting the output to look like (I assume you are not knowingly doing anything that would cause the output to change over time)?
m
Matt Venn
03/02/2023, 3:03 PMno, it should just be a regular pulse
Matt Venn
03/02/2023, 3:03 PMI've seen this on a few designs now, they 'work' for a bit and then the signals fade away
Matt Venn
03/02/2023, 3:03 PMfirst time I've captured it
t
Tim Edwards
03/02/2023, 3:04 PM"regular pulse" means you're just generating a pulse with a loop in the C code that is driving a GPIO output?
m
Matt Venn
03/02/2023, 3:05 PMno, it's coming from a user design
Matt Venn
03/02/2023, 3:05 PMa hyperram controller
t
Tim Edwards
03/02/2023, 3:08 PMI would want to look at the user design, then, to see what might be going on with the power supplies. It could certainly be an issue where the power supply is dropping in the user area until the input to the GPIO is no longer full swing. I think separating out the vccd1 supply and measuring the current through it might be helpful. I don't think you're going to be able to tell much by measuring vccd1 off-chip, because noise and level can be quite different from what's happening locally inside the user area.
Tim Edwards
03/02/2023, 3:09 PMAlso check if there's anything going on on vccd2, vdda1, and vdda2. Might want to keep them grounded as a precaution, though I'm doubtful that that's going to make any difference in this case.
Tim Edwards
03/02/2023, 3:10 PMThere is a register that reads back "power good" from the user area, but it may not catch what you're looking for if it's something like the ground level rising instead of the power level dropping (even assuming that it is, in fact, a power supply issue).
m
Matt Venn
03/02/2023, 3:14 PMI'll need to check the circuit diagram to see if we can measure current or disconnect the psus
t
Tim Edwards
03/02/2023, 3:16 PMThe development board header J3 connects vccd1 to 1V8. It has an existing trace between the two points. if you cut between the two vias with a knife to cut the trace and then install a header, you can put an ammeter across the pins to get the current just through vccd1.
a
Andrew Wright
03/08/2023, 2:03 PMIdeally we wanty something like a 0.1Ohm resistor and an oscilloscope trace of the difference in voltage on both sides...So we can see the current and the current spikes in the design. Also what was the interface to the external hyperram and what was the frequency of operation?