Dual Supply (Bipolar) from an SMPS DC Adapter (Wall Wart)

Thread Starter

reynolds087

Joined Jul 22, 2020
38
It's funny because I've tried a ton of different configurations using electrolytic capacitors, ceramic of a few different values, inductors in the microhenry range, a couple of toroids wound into the tens of millihenries, and multiple stages with LC and RC filters, but nothing as performed better than just a single 10uF ceramic capacitor bypassing the negative and positive. The other circuits that include that capacitor and then add to it, are indistinguishable, so at least they don't reduce the performance, but it's still a little frustrating that I can't find any way to improve the results further. Lol

I tried stacking 3 of the 10uF capacitors in parallel, then a series inductor 25mH, and then another 3 ceramics in a pi filter configuration, but that looks exactly the same as the 1 ceramic capacitor. :oops:
 

DickCappels

Joined Aug 21, 2008
10,246
It looks like you are trying the right sorts of solutions.

When troubleshooting a problem like that it is very important to have your scope's ground clip in exactly the right place because to fast switching glitches ground paths can easily look like inductors.

1642093609186.png

If you don't already have one of these low inductance ground probes around you can make one with bare wire.

You might find changing the type or value of the ceramic capacitors works better than the one you are using now. In my experience type X7R ceramic works well at high frequencies. A low value is best, like .01 or .001 uf.
 

Ian0

Joined Aug 7, 2020
10,276
Also check the earth connection. Does the plug-in power supply have a connection to mains earth? Does the scope?

Start with a simple RC filter (made with a thick-film resistor, not a wire wound) and check that it is does its stuff. If not, it’s probably your measurement setup that is at fault.
Do your results change from day to day?

If you have an old scope probe that you don‘t need anymore, solder it to the circuit. A bit extreme, perhaps, but you know it‘s connected!
 

Thread Starter

reynolds087

Joined Jul 22, 2020
38
Also check the earth connection. Does the plug-in power supply have a connection to mains earth? Does the scope?

Start with a simple RC filter (made with a thick-film resistor, not a wire wound) and check that it is does its stuff. If not, it’s probably your measurement setup that is at fault.
Do your results change from day to day?

If you have an old scope probe that you don‘t need anymore, solder it to the circuit. A bit extreme, perhaps, but you know it‘s connected!
RC filters do not attenuate the noise at all. The only thing that works is the 10uF capacitor.

The power supply is one of those very cheap variable bench supplies from Amazon. It has a grounding plate that allows you to reference either the positive or negative side to ground, and I've tried it floating or ground referenced with no change in the amplitiude of the noise.
 

Thread Starter

reynolds087

Joined Jul 22, 2020
38
Even though the RC filters are not working for this power supply, I am still planning to use them. I have confirmed that when I filter the noisy supply, and then drive the base of a pass transistor with the filtered signal, the resulting waveform with a load connected to the emitter of the transistor is just as clean, even though the collector is drawing current from the unfiltered DC. So I'm still including the RC network in case it has a benefit that I am not seeing with this particular power supply. Here are some pictures of the schematic and the layout for the circuit I'm planning. I haven't tested this configuration exactly yet, but tonight I'm going to mill the PCB and assemble it so I can test its performance.
 

Attachments

Ian0

Joined Aug 7, 2020
10,276
I think that we can take it as fact that an RC filter does reduce high frequency noise, and that if it appears not to, then the problem must be with the measurement method.
 

Thread Starter

reynolds087

Joined Jul 22, 2020
38
I think that we can take it as fact that an RC filter does reduce high frequency noise, and that if it appears not to, then the problem must be with the measurement method.
I'm not saying RC filters don't attenuate high frequencies. I'm just saying that the one I implemented in the schematic, does not attenuate the noise in my test circuit. I don't think it's a problem with the measurements. It's only one power supply that creates this noise. If it was a measurement problem, why wouldn't the noise be present when using other DC power supplies? I'll check it again with the little spring clip, but I am certain the noise is coming from that supply. It actually puts noise into my mains too. My speakers which are not connected to it except through the AC lines in my home, hiss when I turn it on.
 
Top