1
\$\begingroup\$

I am designing a 10 kW rectifier (single phase, 230 V, 50 Hz, 44 Arms, 62 App) with power factor correction (a bridgeless boost DC-DC converter working in continuous conduction mode). All works perfectly except the enormous inrush current need to charge the capacitor (1 mF) in the first quarter cycle (>>100 A). I also want to limit capacitor charging current in the case of few cycles power lost, so I need an active solution instead of a passive one, ie. thermistor. What are the most common strategies to deal with this problem?

I am using this topology:

enter image description here

\$\endgroup\$
  • \$\begingroup\$ Show the circuit you have including PF correction \$\endgroup\$ – Andy aka Aug 31 '15 at 12:19
  • \$\begingroup\$ Can't you set the PF correction circuit to do soft start? \$\endgroup\$ – Andy aka Aug 31 '15 at 13:13
  • \$\begingroup\$ I can't see how this configuration allow me to do soft start. At power on, capacitor acts as a short circuit, so a high current pulse will pass through diodes D3 and D4. If I turn on Q1 and Q2 it will be worse... I need to be able to control this current path. My idea is to place a MOSFET in series with Rs, buy I'm not an expert in electronics, so I need some literature to do it successfully. \$\endgroup\$ – user3368561 Aug 31 '15 at 14:20
  • \$\begingroup\$ How big is Rs. At powerup you shouldn't need to fire Q1 and Q2. The "inrush current" will be limited via Rs. Once you have reached a suitable link potential then enable control \$\endgroup\$ – JonRB Aug 31 '15 at 16:11
  • \$\begingroup\$ @JonRB Rs is a shunt resistor to measure current, not a current limiting resistor. At these power levels (remember, 10 kW, 230 V, >60 A) a fixed current limiting resistor will be a hot spot. \$\endgroup\$ – user3368561 Aug 31 '15 at 16:16
2
\$\begingroup\$

One of the most common solutions is to use a precharge resistor that is shorted out during normal operation.

You are essentially running a boost converter to facilitate power factor correction.

At 10kW you will be looking at using a SCR or an IGBT as the power device

schematic

simulate this circuit – Schematic created using CircuitLab

You would monitor the voltage across your capacitor until it reaches a level that the inrush is manageable (say ... 200V) & then gate your power device. Equally you would stop firing the power device if the voltage fell below say.. 185V.

The downside however is you have a PFC circuitry and that sort of relies on a very tight connection to the rectifier. IF it turns out this typical circuit is not appropriate the 2nd method would rely on pre-charge in the AC lines using back to back SCR. Whether you would need this 2nd method is downto specifics of your implementation. I had to use AC-side precharge with a 250kVA active rectifier.

|improve this answer|||||
\$\endgroup\$
  • \$\begingroup\$ This solution is OK at startup time, because load can be switched off, but in the case of few cycles power lost you never will reach target voltage with a fixed resistor because load current will create a voltage drop, capacitor voltage never will reach target voltage, and inrush current problem will remains. \$\endgroup\$ – user3368561 Aug 31 '15 at 17:01
  • \$\begingroup\$ do you want to maintain output load during such a blackout? if so some more thought. If the output load can be disabled (due to undervoltage) then what I stated is still valid as you would use the 200V:185V hysteresis as an enabler into your load circuitry as well - why have your load drawing power when there is an outage? \$\endgroup\$ – JonRB Aug 31 '15 at 17:05
  • \$\begingroup\$ In my region, micro (few cycles) voltage drops are common so if I am able to run the load with stored energy from the capacitor and recharge it without disconnecting the load, it will be a double victory. \$\endgroup\$ – user3368561 Aug 31 '15 at 17:18

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.