0
\$\begingroup\$

I have an AC to DC converter that outputs 5V. I then run this into 2 regulators on my board. The 4.2V regulator goes into a cellular module, and the 3.3V regulator goes into a microcontroller. I noticed the micro resetting in the debugger and scoped the 4.2V and 3.3V on single mode. I saw this transient. It happens about 3 times a day. The frequency of the transient is very close to 100Mhz. My micro runs at 64Mhz, not sure where this transient could come from. The magnitude of the transient on the 3.3V has gone to 16V pk-pk. This would seem like it would cause severe damage to the electronics. How can I find the source of this transient? I emailed the maker of the AC DC converter. He said it is unlikely coming in on the AC line, and is more likely coupled downstream from the regulators.

enter image description here enter image description here

\$\endgroup\$
4
  • \$\begingroup\$ install a 1uH inductor in series with power into the switcher; follow that with 10uF to ground, with voltage rating to handle the transient. \$\endgroup\$
    – analogsystemsrf
    Commented Nov 14, 2019 at 2:11
  • 1
    \$\begingroup\$ Exactly how are you probing with the scope? At 100 MHz, your 'scope could be seeing a bogus transient (that your circuit does not see). \$\endgroup\$
    – glen_geek
    Commented Nov 14, 2019 at 2:47
  • \$\begingroup\$ "It happens about 3 times a day." at any particular times? What happens if you disconnect the cellular module? Please show schematic and layout of the power distribution. \$\endgroup\$
    – Bruce Abbott
    Commented Nov 14, 2019 at 5:41
  • \$\begingroup\$ The ac dc power supply says switching frequency of 100kHz. The regulators are linear. \$\endgroup\$
    – Bill
    Commented Nov 14, 2019 at 19:08

2 Answers 2

Reset to default
1
\$\begingroup\$

The transient comes from the control loop of your AC DC converter. Unfortunately the only way to eliminate this is to tighten up your control loop. But Why? The transient only lasts for some microseconds. This would not cause any damage normally. A decoupling capacitor across the rails would provide the protection. A varistor can also be used instead. Also most likely there is a regulator in your cellular module that will provide enough protection.

\$\endgroup\$
2
  • \$\begingroup\$ I thought I would add a TVS diode, but they seem to be made for data lines. The datasheets I looked at said not for power lines. There are bypass caps going into and out of each regulator, I thought these would prevent ac dc converter noise. I can look into a varistor. I thought the regulators would reject the noise with their high PSRR. \$\endgroup\$
    – Bill
    Commented Nov 14, 2019 at 18:40
  • \$\begingroup\$ The last one seems a viable option. \$\endgroup\$
    – User
    Commented Nov 15, 2019 at 1:03
1
\$\begingroup\$

BSEE, I struggled for some time with a similar issue that turned out to just be poor probing technique with a single ended oscilloscope. I wrote up the resources/techniques to clean up the measurement in this answer: Massive pk-pk ripple on output using isolated DC-DC converter module

You may find as glen_geek said that the issue is not as severe as it seems from your first samples.

\$\endgroup\$
1
  • \$\begingroup\$ I had the probe and ground clip connected 1 inch each from the output and ground of the regulator. I can solder wires on the output cap for a better measurement. My scope is 100Mhz. I wonder if I near the limit of what it can do. \$\endgroup\$
    – Bill
    Commented Nov 14, 2019 at 18:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.