AD2: Switching Digital Inputs Induce Noise on Analog Inputs

[AlexanderGDean]

What circuit methods do you recommend for keeping digital input signal transitions from showing up on the analog inputs? 

I'm using the AD2 as a mixed-signal scope to view low-amplitude differential analog signals and twiddle/debug GPIO bits from an MCU (48 MHz). I'm using the flywire harness (connected to a custom PCB mounted on an NXP FRDM-KL25Z MCU eval. board). The analog signals are picking up the transitions on the digital input signals and ringing at ~40 MHz. Adding a series resistor (e.g. 1k) reduces the noise about 5x, and capacitor (e.g. 10 nF) does by about another 2x, though I haven't looked at the impact on signal timing perceived by the AD2. 

Before I spend too much time reading up on snubber networks and RC filters I figured I'd ask if your other customers have run into this and what the common solutions are.    

Thanks for any help!

Alex

[attila]

Hi @AlexanderGDean

This crosstalk could between the wires which act as antennas.
This could be highly reduced by shielded cables but twisted wires also help a lot.
Twist the positive and negative scope inputs together and for single ended usage connect the negative to the ground near the test point; twist each of the used digital, trigger, wavegen and supply wire with ground.

Edited May 18, 2022 by attila

[AlexanderGDean]

Thanks, that helps quite a bit!

• Baseline (+/- analog inputs twisted together): 300 mV p-p noise
• Add 1k resistor in series with digital output: 60 mV
• Twist ground wire (grounded at both ends) around digital signal lead: 25 mV  

And now to figure out the most feasible way to procure 154 such wiring harness modifications/adaptors for 154 students by August 22 🤔. 

Edited May 18, 2022 by AlexanderGDean

[JColvin]

Hi @AlexanderGDean,

I get that this isn't the most economic solution, but Digilent does have twisted wire pairs with embedded 100 ohm resistor on one end of the wire that we offer to go with the Digital Discovery but are also compatible with the Analog Discovery (though the adapter is not as it's one too many pin pairs wide), https://digilent.com/shop/digital-discovery-high-speed-adapter-and-logic-probes/.

Thanks,
JColvin

[AlexanderGDean]

Thanks, @JColvin. It is good to see additional options, though it's too bad that the adapter isn't compatible without some hacking. A custom adapter PCB would be easy enough to make, but the connectors could blow the budget.    (...Once again I am stricken by the Curse of the Engineer: the more ways you know how to solve a problem, the easier it is to get distracted or lost in the details.)