Author Topic: My switching power supply (mc34063) makes audible noise  (Read 15936 times)

0 Members and 1 Guest are viewing this topic.

Offline sarfataTopic starter

  • Contributor
  • Posts: 28
My switching power supply (mc34063) makes audible noise
« on: August 11, 2012, 02:30:57 am »
Hi,

I have built a Arduino clone that uses a switching power supply because the input voltage might be as high as 20V and also to learn ;). It kind of works but when I power the board, I hear a very annoying high pitch noise.

The power supply is based around the MC34064 that Dave described in a video a while ago. It's being used in step-down configuration to get 3.3V. I have a Vin of 18V (if I lower Vin to 8/10V, the noise is significantly reduced but I can still hear it).



The values I have used are:
L: 33uH (should have been 36 but I could not find one)
R1: 2k
R2: 3.2k
Ct: 85pF
Rsc : 0.240 Ohm
Co: 10uF (I did not use an electrolytic here but a ceramic, could that be a problem?)
Diode is a B220A.

Where could this noise come from: the self? the diode? a capacitor? What can I do to diagnose that and reduce that noise?

I have attached two captures from the board. In red is the voltage at the output of the MC34063, getting into the self and the diode. In yellow is the "regulated" voltage measured on the board.

You can see that the output voltage goes from 3.04V to 4.24V that's a 1.2V "ripple" ... And also the switching frequency seems to be around 5kHz but is not very stable.

I should also say that there is not much power consumption from the board. Just an Atmega 34U4, a few pull-ups and a some leds right now.

Thanks for your suggestions!
thomas
 

Offline Rerouter

  • Super Contributor
  • ***
  • Posts: 4705
  • Country: au
  • Question Everything... Except This Statement
Re: My switching power supply (mc34063) makes audible noise
« Reply #1 on: August 11, 2012, 02:58:24 am »
at 5Khz there is a good chance your inductor is working as a speaker,
 

Offline T4P

  • Super Contributor
  • ***
  • Posts: 3697
  • Country: sg
    • T4P
Re: My switching power supply (mc34063) makes audible noise
« Reply #2 on: August 11, 2012, 03:42:47 am »
You need more than just 30KHz, our ears can also hear attentuated 30KHz (See human ear rolloff chart)
And the inductor CAN ring at any frequency because of magnetostriction
 

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5170
  • Country: ro
  • .
Re: My switching power supply (mc34063) makes audible noise
« Reply #3 on: August 11, 2012, 03:57:32 am »
Yeah, you need to increase the frequency to probably about 40-50 kHz.

The 10uF ceramic on the output may be an issue. Some regulators, especially low drop linear regulators, need a capacitor with a bit of ESR, usually above 0.1 ohm. Ceramic caps have lower esr. See this document: http://www.eetasia.com/ARTICLES/1999APR/1999APR08_ICD_AMD_TAC1.PDF

I don't understand why you'd use 2 100nF on the output .... I think the input could use a 0.1uF or a 0.47uF in addition to that 100uF cap.

 

Offline amyk

  • Super Contributor
  • ***
  • Posts: 8526
Re: My switching power supply (mc34063) makes audible noise
« Reply #4 on: August 11, 2012, 08:54:48 am »
Agree with the others above, 5kHz is way too low for SMPS. From the datasheet, the recommended frequency is 33kHz.

10uF might also not be enough bulk capacitance, but the low frequency can also exacerbate this problem. With a low frequency (= longer period), the cap spends more time discharging into the load.

Please note that a  MC34064 is a very different IC than a '34063 even though it also comes in an 8-pin package :D
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 28429
  • Country: nl
    • NCT Developments
Re: My switching power supply (mc34063) makes audible noise
« Reply #5 on: August 11, 2012, 01:58:11 pm »
If that 10uf is a normal electrolytic cap then there is your problem. Try using a ceramic capacitor. Read the datasheet on how much capacitance is required. It can also be a layout problem causing instability in the output regulation.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline JackOfVA

  • Supporter
  • ****
  • Posts: 350
  • Country: us
Re: My switching power supply (mc34063) makes audible noise
« Reply #6 on: August 11, 2012, 03:59:57 pm »
Not only the inductor can be suspect as a speaker, so can the 10uF ceramic capacitor.

Jack
 

Offline jahonen

  • Super Contributor
  • ***
  • Posts: 1055
  • Country: fi
Re: My switching power supply (mc34063) makes audible noise
« Reply #7 on: August 11, 2012, 04:05:47 pm »
The lack of error amplifier (it has just a comparator) and PWM modulator in MC34063 might generate audible subharmonic noise due to erratic switching pattern. Change it to something like LM2675 and your problems will be probably gone :) You can also get a fixed 3.3 volt version of LM2675 to remove the voltage setting resistors.

Regards,
Janne
 

Offline Berni

  • Super Contributor
  • ***
  • Posts: 5050
  • Country: si
Re: My switching power supply (mc34063) makes audible noise
« Reply #8 on: August 11, 2012, 07:55:14 pm »
Very likely that the 10uF ceramic is making the noise. These high capacitance ceramics have a very prominent piezo effect in them. I once put a 10uF 1206 ceramic across a few KHz square wave to test it out and it was so loud you could hear it clearly in to the next room!

The trick i found on finding where the noise is from is i take a scalpel knife and press the tip of it sideways on to the component, this takes up the vibrations from the component and makes them louder. When pressing on to a component makes the noise a lot louder you know you found the source.

Also yes 5KHz is a very low frequency and will defenetly be audible just because of the inductor
 

Offline sarfataTopic starter

  • Contributor
  • Posts: 28
Re: My switching power supply (mc34063) makes audible noise
« Reply #9 on: August 12, 2012, 12:27:54 am »
Thanks everyone for your replies!

In my initial post, I said the switching frequency seems to be 5kHz but that was incorrect. I was measuring the frequency of the noise on the output and not the switching frequency which are obviously not the same (I realize that now...).
The switching frequency for the 82pF I use is 127kHz (measured) and matches the datasheet.

Many of you guessed that the capacitor might be the problem. I replaced it with a 100uF electrolytic cap and the noise completely went away. The output signal is also much cleaner. Thanks a lot Berni for the "scalpel knife trick", it works also very well with my tweezers ;)

So I have re-read the datasheet and AN920 (very interesting) at http://www.onsemi.com/pub_link/Collateral/AN920-D.PDF. I have taken some more capture with my DSO and there are still a few things that dont seem right.

1 - Even with a 100uF capacitor, I still get some pretty large spikes (680mV) every 450uS



2 - With a 10 uf Capacitor (which is the value I had calculated and double-checked online with MC34063 calculators), I get bigger spikes and at a higher frequency (every  100/200 uS) (see attached file output-10uf-electrolytic.png)



3 - In my circuit, the timing pin goes from -0.46V to +1.76V but in AN920 I read that the timing pin should go from 0.75 to 1.25V - surprising... (see attached file timing-capacitor.png)



So changing the type of capacitor did help make the noise go away but there still seems to be something weird in there. It seems to me that it should work much better than that, correct?

thanks again,
thomas
« Last Edit: August 12, 2012, 12:30:21 am by sarfata »
 

Offline Psi

  • Super Contributor
  • ***
  • Posts: 10385
  • Country: nz
Re: My switching power supply (mc34063) makes audible noise
« Reply #10 on: August 12, 2012, 12:36:00 am »
Can you post a scope pic of the VCC pin, just want to check if your input is stable.
Greek letter 'Psi' (not Pounds per Square Inch)
 

Offline sarfataTopic starter

  • Contributor
  • Posts: 28
Re: My switching power supply (mc34063) makes audible noise
« Reply #11 on: August 12, 2012, 01:03:40 am »
Great idea, thanks. I am using a linear lab power supply (mastech hy3003d) by the way.

This is the result:



There is indeed 1v spikes in the input but this seems to be related to the load (my board and its power supply) because if I just put a resistor (taking much more current) on the output of my power supply I dont see any ripple that big.
 

Offline amyk

  • Super Contributor
  • ***
  • Posts: 8526
Re: My switching power supply (mc34063) makes audible noise
« Reply #12 on: August 12, 2012, 11:56:28 am »
You mention you were powering an ATMega with your SMPS, so is there anything that occurs at ~2.2KHz intervals on it? Have you tried a plain resistive load to see if the spikes go away? Also consider adding another choke at the output --- a few tens of uH should be sufficient. You can learn a bit from looking at PC PSU schematics.
 

Offline T4P

  • Super Contributor
  • ***
  • Posts: 3697
  • Country: sg
    • T4P
Re: My switching power supply (mc34063) makes audible noise
« Reply #13 on: August 12, 2012, 12:36:18 pm »
A resistive load makes the spikes go away, he mentioned
 

Offline dan.corneanu

  • Newbie
  • Posts: 1
  • Country: nz
Re: My switching power supply (mc34063) makes audible noise
« Reply #14 on: November 16, 2019, 10:34:45 am »
I know this is a very old topic, but was there any resolution on this issue?
I am also seeing the same level of noise on the input and it is not from the powersource bacause with a resistive load it is clean.

Regards,
Dan
 

Online iMo

  • Super Contributor
  • ***
  • Posts: 5570
  • Country: va
Re: My switching power supply (mc34063) makes audible noise
« Reply #15 on: November 16, 2019, 10:48:49 am »
You will always see some ripple caused by the switcher. You may minimize it with careful layout, construction and filtering (LC filters usually work best).
Readers discretion is advised..
 

Offline SiliconWizard

  • Super Contributor
  • ***
  • Posts: 15797
  • Country: fr
 

Offline sarfataTopic starter

  • Contributor
  • Posts: 28
Re: My switching power supply (mc34063) makes audible noise
« Reply #17 on: November 19, 2019, 05:23:58 am »
Yes I solved the issue. As far as I remember, just adding more capacitance on the output saved the day.

I agree with @SiliconWizard and have stopped using the mc34063. The TI based webtools to generate power supply designs are quite amazing. You dont need to do the math, they give you tons of options and yes the chip is more expensive but for 3.3V they have highly integrated circuits that will require a lot less additional parts. Less parts => less PCB space, less time spent on digikey searching for a specific capacitor, etc. You will save time and money by using them.
 

Offline OwO

  • Super Contributor
  • ***
  • Posts: 1250
  • Country: cn
  • RF Engineer.
Re: My switching power supply (mc34063) makes audible noise
« Reply #18 on: November 19, 2019, 05:43:03 am »
There are a LOT of cheap SOT23 all-in-one buck converter ICs, some with compatible pinouts and up to 3A output current. For example for low voltage (5V) there is FP6373, JW5211, etc with the same pinout. For higher voltage there is JW5033, RT8258, etc (easily found on LCSC). These also usually switch at 1MHz or higher which means you can use small inductors (there are size 1008 ones that can go up to 2A, still very cheap because they are used in cellphones).
Email: OwOwOwOwO123@outlook.com
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf