1
\$\begingroup\$

I've just put together my first busk converter with a LMR16006Y-Q1. http://www.ti.com/lit/ds/symlink/lmr16006y-q1.pdf

Here is what it looks like on the scope. The Pink signal is unloaded, the other is with a 90 ohm load. It seems noisy to me. I think I need to make some changes. But I'd like to get input from someone else.

enter image description here

I'm using a 6.8uH coil, 2.2uF cap on the input a 15uF cap on the output (ignore the second output cap in the schematic) The schematic is here: enter image description here

It's laid out on a PCB almost exactly like this, with the only difference being that the Feedback is tapped straight into the inductor output pad rather than stretching all the way to the output after the cap. I'm not sure if that will really mak a difference. :enter image description here

\$\endgroup\$
  • \$\begingroup\$ Which regulator IC are you using? What are the ripple and noise requirements for your load? \$\endgroup\$ – The Photon Oct 31 '18 at 18:10
  • \$\begingroup\$ Noise seems to be around 50mV. I'm guessing that you want to feed it to an IC or mC... it' s not terrible. But are you sure that both channels are on the same V/div \$\endgroup\$ – thece Oct 31 '18 at 18:25
  • \$\begingroup\$ @ThePhoton It's a LMR16006Y-Q1. The Datasheet is here: ti.com/lit/ds/symlink/lmr16006y-q1.pdf I'm waiting to hear back from a SOC manufacture on noise specs but being an automotive application I think my noise is going to go up once I feed this with 12v from an alternator. so I'd like to make improvements if possible. \$\endgroup\$ – Steven Lutz Oct 31 '18 at 18:32
  • \$\begingroup\$ @thece Yes, I need to feed a small microprocessor. Both signals are on the same div. What is a realistic expectation to be able to get this down to from 50mv? \$\endgroup\$ – Steven Lutz Oct 31 '18 at 18:32
  • \$\begingroup\$ That is mostly high-frequency noise, the switching frequency itself seems to be dampened okay. The rest will be handled by the decoupling capacitors next to the microprocessor. Connecting the feedback to the output capacitor or even further downstream will improve behaviour if your load impedance is low and fluctuates wildly. At a static ~30mA (90Ω), you won't see much of an effect here, but you should see a difference in ringing when you connect and disconnect the load. \$\endgroup\$ – Simon Richter Oct 31 '18 at 18:44
0
\$\begingroup\$

Looks good to me, there aren't any frequencies that really stick out, so it looks good. If your running a microprocessor, spurious frequencies can contribute to error by shifting the voltage levels that transistors switch at. If your worried about this board being an unintentional radiator (FCC testing), then it would be a good idea to build\buy an antenna for that oscilloscope and see if the circuit is radiating.

The microprocessor may have a limit for input ripple, check the datasheet. Then measure the ripple at the power filter cap. If 50mV is beyond the recommendations of the micro, then you'll need more filtering. Either an LC filter on the output or more power filter caps.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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