Probing medium speed signals

[Dubbie]

Hi,
I'm working on a design at the moment using an Artix-7 FPGA. It's not a particularly  fast design, but I've been a little surprised at how crappy the square waves look on the scope. What I want to know is; is what I'm seeing normal? is it the best I can expect?

With regards to probing technique, the signal is coming out an 0.1" header pin adjacent to a GND pin. I am using the standard 500Mhz keysight passive probe with the spring ground thingy. (no long crocodile lead or witches hat grabber)

Let me know your thoughts!

Thanks,
Rhys.

[Alex Eisenhut]

Did you compensate the probe before?

[Dubbie]

Sorry, forgot to mention, yes, the probe is properly compensated.

[Alex Eisenhut]

Data gets to where it's needed with no problem? Then it looks fine to me. There might be some sort of pre-emphasis or "strong drive" option on your drivers. Maybe you can turn it off and it'll settle down the edges a bit, you'll reduce current consumption and heat dissipation too.

[Dubbie]

Yes, everything seems to be functioning properly. It's just because I haven't done much before above a few tens of Mhz, I just wanted to check that I wasn't doing anything boneheaded! Thanks for the suggestion regarding drive. I'll play around with the drive settings and see what I see. This is only my second FPGA project, so I am very green.

[joeqsmith]

It's all about the edge rates.   

I bought one of those ARTY boards to play with and I still have not done anything with it.

[tggzzz]

The edge rate will depend on the type of driver inside the Artix; there are many, and they can be chosen inside the design tool (presumably Vivado).

The standard probe compensation is irrelevant w.r.t. the overshoot. The standard probe compensation is done with ~1kHz signals. If it is incorrect then at these frequencies all you would see is that the entire signal has a higher than expected amplitude, but the shape is unchanged. Some probes, especially older ones, have extra high frequency compensation "hidden" inside the case that plugs into the scope. Those, if any, would significantly affect that waveform's shape.

Don't forget that at 500MHz the "high" impedance probe's input impedance is <50ohms

Personally I would use an active probe, or a commercial "low" impedance Z0 probe, or make one myself. FFI, see https://entertaininghacks.wordpress.com/library-2/scope-probe-reference-material/

[Dubbie]

Thanks for the link tggzzz, some good reading in there.


Sent from my iPhone using Tapatalk