0
\$\begingroup\$

I want to know how to dimension a bridge rectifier.

As an example,in a SMPS, the bridge rectifier has repetitive peak current, up to 10A at rate of 100kHz, but in a small period of time, like a small duty cycle. But the average current is 1A. How do I choose the component (bridge)? Based on its mean current value, on its peak current, or calculating the I²t to mach the datasheet specifications?

I read something about dimensioning it using I²t specification from datasheet, but on GBJ3510 (for example) there's a disclaimer about this value being for non repetitive peaks.

The situation expressed above and on the previous edits were only examples to serve as a tool for the main question of how to dimension the bridge based on current aspects.

\$\endgroup\$
8
  • 1
    \$\begingroup\$ How large is your bulk capacitor? Do you have a PFC? Could you maybe add a rough block diagram of your SMPS? \$\endgroup\$ Commented Apr 25 at 16:04
  • \$\begingroup\$ What is the "duty cycle" here? Do you have a PFC stage? \$\endgroup\$ Commented Apr 25 at 16:14
  • 2
    \$\begingroup\$ It's so easy to simulate this and get a far better (and more comprehensive) answer than by using archaic formulas designed for people using slide rulers. \$\endgroup\$
    – Andy aka
    Commented Apr 25 at 16:18
  • \$\begingroup\$ I just need to select the right component based on the current ratings I supplied. Everything else is already settled down. I did some research about how to select the bridge but didn't find any answer specific for this situation or found and didn't find it reliable or understood it. \$\endgroup\$
    – IMorimoto
    Commented Apr 25 at 17:46
  • 1
    \$\begingroup\$ (mach match? meet?) \$\endgroup\$
    – greybeard
    Commented Apr 26 at 15:44

1 Answer 1

1
\$\begingroup\$

Your PFC-less design is fundamentally flawed. That 14.4V 200A output alone can provide almost 3kW. At these power levels, a PFC is actually mandatory.

If you plug this thing in as you've drawn it in your schematic, you're going to detonate the bridge rectifier and trip all breakers within line of sight of that contraption. The capacitors act as a dead short as they get charged up by mains.

Additionally, even if you managed to somehow get this circuit to survive the inrush current, it'll consume mains power in horrible spikes approaching 200A of instantaneous current (I simulated it). That's again pretty much guaranteed to (magnetically) trip most breakers rated for 32A or less. In other words: Your circuit is drawing power in spikes of almost 40kW, 120 times per second. This is really bad.

AC rectifier simulation

If you still want to spec a proper rectifier, you'll need one that can handle 200A at 10% duty cycle, as can be seen in the simulation.

Please, though, build a PFC, even if just a passive one.

An active PFC (power factor correction) circuit usually consists of a boost converter that tries to match its input current draw to the sinusoidal waveform of the incoming mains voltage. This makes it essentially present itself as a resistive load to mains. The output of such a PFC provides a more-or-less smoothed ~400V DC, which is exactly what you need to power your subsequent isolated converter stage. If you don't want to build an active PFC, a passive PFC consisting of suitable inductor in series with the line input might help a bit.

\$\endgroup\$
8
  • \$\begingroup\$ In fact, nowadays, it works with a 30A rectifier bridge for days without problem. In certain conditions, like too hot ambient and surge voltages from power line, it damages, but using some varistors, at line input, seemed to solve most of the cases. That's why I want to change it. I agree PFC is important, but I'm not able to insert it at the moment. The surge current from discharged capacitors is a problem that I already fixed. \$\endgroup\$
    – IMorimoto
    Commented Apr 25 at 19:05
  • \$\begingroup\$ Well in that case, you know the answer: It's drawing 200A pulses of current. You know the average current (around 20A) and pulse current (200A), so from that you know that the duty cycle is approximately 0.1 (10%). Select a rectifier that can handle 200A pulses at 10% duty cycle, or ideally a bit more. And don't be surprised if the utility company turns off your power due to the stress (or even damage) you're causing to their equipment. \$\endgroup\$ Commented Apr 25 at 19:14
  • \$\begingroup\$ Sorry, 200A is the output current at low voltage. The maximum peak current that passes though the line input is 73A in 13% duty cycle, resulting in 23A (average). There are other chargers like this for sale, without PFC (I have two of them. One, uses GBJ5010). As I said, PFC is important but I can't add it, only modify what I have. \$\endgroup\$
    – IMorimoto
    Commented Apr 25 at 19:28
  • 2
    \$\begingroup\$ No, 200A is the peak mains input current I determined in a simulation with LTSpice. The fact that the output current is also 200A is just coincidental. The peak SMPS transformer current is unrelated to the peak mains current. I've added a screenshot of the simulation. \$\endgroup\$ Commented Apr 25 at 19:30
  • \$\begingroup\$ I indeed intend to add a passive PFC and EMI filter, but only after this stage of changing the bridge. However, you can abstract this situation, because, the main doubt is not because of this equipment. But on how to dimension the bridge (and other diodes) based on the current aspects (mean value and high frequency peaks), generally speaking for any other situation. I used this charger as an example to serve as a tool for the answer, but it could be a small power SMPS as well, or even other devices. \$\endgroup\$
    – IMorimoto
    Commented Apr 26 at 15:33

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.