1
\$\begingroup\$

I have two Recom R24-150B dc-dc converters (24V in, adjustable up to 220V out) and I want to have +/-220V output rails. The converters need a resistor between the - (COM) output and their adjust port in order to set the output voltage. This acts to regulate the output.

According to the data sheet it is ok to tie the outputs of these converters together to obtain higher voltages because they're isolated, as long as some capacitors are attached across the outputs of each unit to prevent beats appearing between the switching frequencies of the two units.

I tied two together and attached the filtering capacitors. I grounded the middle to create the bipolar output rails, and this produces the desired dc voltages on each rail, +220V and -220V. However I am seeing quite a lot of noise with my oscilloscope around 1 MHz on the load attached between the two HV rails, around 200-300 mV peak-to-peak. The switching frequencies should be around 200 kHz so I guess this is some higher harmonic content. This noise appears to be on the ground rail between the two converters because the noise on each of the HV rails is smaller and in phase. The individual output ripple from each converter is less than 100 mV peak-to-peak so clearly something is ringing up and increasing the output noise.

Is there some reason why bridging two converters together like this to form differential outputs is a bad idea? Or is there some additional filtering required somewhere?

\$\endgroup\$
11
  • 1
    \$\begingroup\$ "200-300 mV peak-to-peak... The individual output ripple from each converter is less than 100 mV peak-to-peak" - does not compute. The total should be the sum of the two individual outputs. Anyway 200-300mV on 440V doesn't sound excessive to me. What is your application that requires lower ripple? \$\endgroup\$
    – Bruce Abbott
    Commented Sep 6, 2016 at 20:01
  • \$\begingroup\$ @BruceAbbott, sorry I was not clear. The ripple on each converter on its own is less than 100 mV peak-to-peak if it's the only one in the circuit. If I connect two together via a load then they oscillate. The problem I have is that this noise on the ground rail enters a power amplifier I attach to these rails and shows up at the output with much higher gain. It's not so much that the ripple is too high, it's that I can't seem to get rid of it with the suggested filtering on the data sheet. That led me to think that maybe there's a reason not to chain converters together to make bipolar rails. \$\endgroup\$
    – Sean
    Commented Sep 6, 2016 at 20:22
  • \$\begingroup\$ How do you measure the "total" ripple? Where do you put GND of the scope? I once had a similar problem which turned out to be a measurement problem... \$\endgroup\$
    – Junius
    Commented Sep 6, 2016 at 20:33
  • \$\begingroup\$ @Junius, that's interesting. What was your problem? I was also suspicious about my measurement: I attached the scope ground to the (virtual) ground between the converters, and the probe on the output of my power amplifier (which is connected to the HV rails and has a signal input). Hmm, although there aren't ground loops external to the board (the 24V input is floating), there are small ground loops on the board I've built. Could this be the problem? \$\endgroup\$
    – Sean
    Commented Sep 6, 2016 at 20:39
  • \$\begingroup\$ To be honest, I have to recap what I have measured and think about your circuit before giving possibly wrong hints... I will check tomorrow \$\endgroup\$
    – Junius
    Commented Sep 6, 2016 at 20:46

1 Answer 1

Reset to default
0
\$\begingroup\$

It turns out the problem was due to interaction between the two converters via the link joining them (positive output on one to the negative output on the other). When I tried filtering the output previously, I only added components to the +/-HV rails and not the shared ground. Adding ferrite beads between the converters and ground got rid of most of the noise. Thanks everyone for their suggestions, they helped steer me towards the solution.

\$\endgroup\$
1
  • \$\begingroup\$ Nice! This might support the ground bounce theory. If your ground connection had appreciable impedance when one converter dumps a bunch of current to its ground, the resulting voltage glitch would be seen by the other converter. If you ensure good electrical connection as well as a nice decoupling cap close to the converter pin it should also help \$\endgroup\$
    – Daniel V
    Commented Sep 10, 2016 at 0:08

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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