Hi all,

I am using a DCP020505 unregulated 5V DC/DC converter which is sourced from another 5V supply. The output of this IC is isolated from the supply and has its own isolated ground.

I am using the input supply to power a MCU and some op amps and wanted it to be isolated from the electronics powered from the DC/DC converter.

I am fine with some spiking on the output of the converter but its switching is causing brutal spiking to my input supply. I have tried placing some electrolytics across the input supply and that does a decent job at getting most of the noise down, but its still larger than I would like (~200 mVpp).

Wondering if anyone had any suggestions.

*EDIT I am testing this on a breadboard before I solder to a PCB so the noise might be worse off.

Thank you!

 

The data sheet recommends ceramic caps at the input due to their lower esr I think they mentioned 2.2 ufd.
They are a little pricy.
http://www.mouser.com/Passive-Components/Capacitors/Ceramic-Capacitors/_/N-5g8m?P=1z0wquuZ1z0z819

You might try a smaller value (Cheaper) in parallel with your electrolytic.

 

You might also try adding a small resistor, ferrite bead, or inductor in series with the power to the converter (before the capacitors). Those help the decoupling capacitor do its work.

Also make sure the leads to the decoupling capacitors are as short as possible. The inductance of even a short lead can significantly reduce the capacitor effectiveness.

 

The data sheet recommends ceramic caps at the input due to their lower esr I think they mentioned 2.2 ufd.
They are a little pricy.
http://www.mouser.com/Passive-Components/Capacitors/Ceramic-Capacitors/_/N-5g8m?P=1z0wquuZ1z0z819

You might try a smaller value (Cheaper) in parallel with your electrolytic.

Thanks for the reply.

I have a 2.2u polystyrene across the inputs and it cuts down on the major noise but still roughly around ~200mVpp.

I tried putting smaller ceramic caps in parallel but it didn't seem to help much. The electrolytic tended to make it worse unfortunately.

 

You might also try adding a small resistor, ferrite bead, or inductor in series with the power to the converter (before the capacitors). Those help the decoupling capacitor do its work.

Also make sure the leads to the decoupling capacitors are as short as possible. The inductance of even a short lead can significantly reduce the capacitor effectiveness.

Thanks for the reply.

The inductors cut down the noise very well!

One quick question regarding the inductors. I have two types of 100uH inductors, one that looks like a resistor (axial) and a one that is radial, much larger, and looks like bob.

Wondering what the difference between these two are and which might be a better choice, even though they seemed to do roughly the same.

 

Thanks for the reply.

The inductors cut down the noise very well!

One quick question regarding the inductors. I have two types of 100uH inductors, one that looks like a resistor (axial) and a one that is radial, much larger, and looks like bob.

Wondering what the difference between these two are and which might be a better choice, even though they seemed to do roughly the same.

I've seen quite a few inductors, but one that looks like bob is a new one on me.

Is it your contention that noise from the switching in the DC-DC converter is coming back through the isolation barrier and affecting the input to the circuitry powered from the 5v source? If so I'd concentrate on that before I started hanging caps and beads all over the place.

Noise comes in several flavors, you're talking about conducted noise in this case -- right?

 

I've seen quite a few inductors, but one that looks like bob is a new one on me.

haha probably not the right word for it. The bob one looks like the 330uH in this image.

 

So it looks like wire on a bobbin with some shrink tubing or other protective coating. What is important to you is core material and how much current it can handle without raising it's temperature, which affects the magnetic properties of the core. The ones that look like resistors are fabricated with very thin wire and are not likely to handle much current.

You probably meant "...looks like a blob."

 

So it looks like wire on a bobbin with some shrink tubing or other protective coating. What is important to you is core material and how much current it can handle without raising it's temperature, which affects the magnetic properties of the core. The ones that look like resistors are fabricated with very thin wire and are not likely to handle much current.

You probably meant "...looks like a blob."

haha thanks for the explanation

 

Hello,

I often have seen 6hole ferite beads to decouple the noise.
Here is a picture I have taken of a couple:

Bertus

 

................................
Is it your contention that noise from the switching in the DC-DC converter is coming back through the isolation barrier and affecting the input to the circuitry powered from the 5v source?
......................

What "isolation barrier"?

 

I've seen quite a few inductors, but one that looks like bob is a new one on me.

Is it your contention that noise from the switching in the DC-DC converter is coming back through the isolation barrier and affecting the input to the circuitry powered from the 5v source? If so I'd concentrate on that before I started hanging caps and beads all over the place.

Noise comes in several flavors, you're talking about conducted noise in this case -- right?

The input supply has very little ripple when the converter is not connected. Once connected the input supply contains high frequency (>500 kHz) noise from the DC/DC converter. I am pretty certain it is because of the switching current requirement of the DC/DC converter and the supply can't keep up without some sort of cap or inductor to smooth the load in between. So yep, talking about "conducted" noise.

I do not think this is due isolation being compromised from the output of the converter back to the supply. The isolation breakdown from input to output is much greater than 1 kV.