1
\$\begingroup\$

I am using Lm2596-ADJ for input power regulation in my circuit. Below is the schematic.schematic

It powers an AVR and a sim800H gsm module. When the sim800H is about to power On or when making a call I hear a buzzing sound from the board as at these times the current requirements are higher. LM2596 should handle this easily, the Input current I am providing is 1.2A @12VDC which should be enough as well. I would like to know which component is responsible for this noise and How to eliminate it. Apparently there is no effect of this on the working of the board as neither AVR nor sim800H shows in problem.

enter image description here

Here is the sound recording ... https://soundcloud.com/dost-muhammad-shah-531566807/buzzing-sound

\$\endgroup\$
2
  • \$\begingroup\$ At first glance, I'd suspect the inductor. You can get tiny mechanical movement of the coil or nearby metalwork from its magnetic field. \$\endgroup\$
    – TonyM
    Jun 14 '17 at 10:55
  • \$\begingroup\$ Indeed, would sound like coil whine. It's the "only" thing that can normally rattle. You can try gluing the coil to fix it's windings, but beware of dissipation. At least you could do it to try and confirm source of noise. Of course the noise in the end is due to the regulator, there should be a straight correlation between PWM duty cycle and the noise you hear, if you can capture that \$\endgroup\$
    – Andrés
    Jun 14 '17 at 11:12
1
\$\begingroup\$

The sound is probably coming from the inductor.

However, that indicates that the power supply control is going meta-stable. The frequency of individual pulses should be way past what you can hear. What you hear are oscillations of the control system.

Usually this causes no real harm, just a little drop in efficiency. I see that this switcher chip doesn't have any external compensation, so there is nothing you can do to tweak the controller.

Make sure that instability isn't caused by bad or open C2. That should be a ceramic mounted physically close to the chip. If C2 is the electrolytic in the top left corner of your picture, then that's clearly bad. Replace it with a ceramic, or possibly two in parallel, then next time pay attention to the recommended practices in the datasheet.

\$\endgroup\$
6
  • \$\begingroup\$ If the loop is metastable, a phase lead cap across the 2.4k resistor might tame it. Somewhere around 10nF or so would be a decent start (quite large because the resistor is quite small). \$\endgroup\$ Jun 14 '17 at 11:59
  • \$\begingroup\$ c2 is ceramic here. 33uF. The regulator does get hot on higher input voltages although they are in range of the regulator. This PCB is a prototype so I can change the design according to your suggestions \$\endgroup\$
    – dmSherazi
    Jun 15 '17 at 8:21
  • \$\begingroup\$ @dmsh: No, C2 is clearly not ceramic. I just noticed that the cap in the top left corner of the picture is actually labeled "C2" and is of the right value. It is obviously electrolytic. \$\endgroup\$ Jun 15 '17 at 10:49
  • \$\begingroup\$ @OlinLathrop sorry I wanted to say is not ceramic but missed it while typing \$\endgroup\$
    – dmSherazi
    Jun 15 '17 at 11:39
  • \$\begingroup\$ @OlinLathrop can you please explain why c2 should be ceramic? Most of the lm2596 modules available have an electrolytic caps as can be seen here \$\endgroup\$
    – dmSherazi
    Jun 16 '17 at 14:29

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.