0
\$\begingroup\$

Currently I am designing a Constant Current charger for a capacitor which charges it from 0 to 350 VDC using a peak current mode controller.

As we expected from simulation, I see there are sub-harmonic oscillations in reality at a duty cycle of 50% and higher.

But as I am charging the capacitor from 0V to the input voltage of 350V (D=100%) means the sub-harmonics occurs only for a short time and when it's charged I set the duty cycle to 100%.

Can I ignore/accept the sub-harmonic oscillations for a short period of time, or should I make a slope compensation to fix it?

Thank you for your time.

The oscillation looks like this: enter image description here And is similar as reality.

\$\endgroup\$
  • \$\begingroup\$ Welcome to EE.SE! Please show simulation/oscillogram and schematic/output stage and feedback mechanism. \$\endgroup\$ – winny May 14 at 18:40
  • \$\begingroup\$ Thanks for your answer, editted my question. \$\endgroup\$ – Klaas Pommes May 15 at 20:10
  • \$\begingroup\$ I thought you had to compensate for it or else the regulator is unstable at those duty cycles? It's not just like a nasty noise or ripple that your converter will continue to operate through. The converter goes haywire doesn't it? \$\endgroup\$ – DKNguyen May 15 at 20:17
  • \$\begingroup\$ Oh, I just noticed you have a caveat in the way you are using the SMPS. I guess it's might be okay...but I'd be careful if you are changing your loads around. \$\endgroup\$ – DKNguyen May 15 at 20:21
  • \$\begingroup\$ Well it does work, I see those oscillations in reality too! \$\endgroup\$ – Klaas Pommes May 15 at 21:51
1
\$\begingroup\$

I'd consider potential negative effects of oscillation:

  1. EMI. Do you need to pass EMC tests? Will there be undesirable audible noise due to magentostriction? Is the EMI likely to disrupt nearby circuits?
  2. Ageing. If you were charging a battery I'd be concerned about localised heating and non-ideal chemistry effects. With a capacitor these issues are negligible.
  3. Instability. Are the oscillations bounded? That is, what is the frequency response of the control loop and what's your margin for stability? As long as you have a healthy margin from where the oscillations become self-perpetuating, there's no stability issue.

As long as these issues are manageable, I can't see from what you've posted, that there'd be any issues. A PWM (bang-bang) charger, for example, would have much more significant oscillations in current, and they can be made acceptable.

\$\endgroup\$
  • \$\begingroup\$ Thanks for your answer. As I am not Charging a battery but a capacitor the ageing is not a big issue. I can clearly hear the inductor switching when it's oscillating. As I only trip at 6A, I don't know what the frequency response is of my loop, how can I determine? And at last, will there be any safety issues when I keep Charging the way like this? \$\endgroup\$ – Klaas Pommes May 15 at 21:51
  • \$\begingroup\$ You can try to model the frequency response with system analysis, or you can do it empirically - try to insert current oscillations of different frequencies to see how their amplitude is effected. Safety issues? That would need a lot more information. From what you've posted I can't see any, but 350V and 6A is inherently dangerous! You need to take all due precautions. \$\endgroup\$ – Heath Raftery May 15 at 22:55
  • \$\begingroup\$ About the safety I mean, can the oscillation be dangereous? \$\endgroup\$ – Klaas Pommes May 16 at 18:43
  • \$\begingroup\$ Apart from the secondary issues already raised (noise, interference, heating/vibration due to magnetostriction, heat/gasification due to chemistry) there are no inherent safety issues due to current oscillation. As mentioned, PWM controllers often exhibit significantly more dramatic oscillation with no cause for alarm. \$\endgroup\$ – Heath Raftery May 16 at 21:03

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.