0
\$\begingroup\$

I would like to purchase or build a simple AC-DC power supply, with 400V/32A 3-phase input, and DC output adjustable (by a microcontroller which I would add) between 300-400VDC and upto 100A (not exceeding 22kW total). The purpose would be to charge an electric vehicle via the fast charge port (CCS or ChaDeMo), which provides direct DC access to the vehicle's battery (usually around 350VDC and 15-70kWh depending on car type). Such power supplies are really expensive (I haven't found any under 5000$) and I have a hard time understanding exactly why they are so expensive. It doesn't need to be isolated, since all parts are extremely well protected from touching. The output voltage is pretty close to the input voltage, and never higher, so shouldn't a rather simple PWM regulation be enough, no transformers, no big voltage differences? (I know such a power supply is not enough to charge an EV, the microcontroller part needs to communicate with the car via CAN (ChaDeMo) or Powerline (CCS), and multiple safety circuits must be implemented, but I already know how to do that, what's missing is the power supply part).

\$\endgroup\$
  • \$\begingroup\$ So you want it to be simple and therefore circumvent the legally required power factor correction circuitry for high power power supplies too? These power factor correction circuits might nearly triple the BoM part cost for an electronic circuit. \$\endgroup\$ – Andy aka May 26 at 17:26
  • 2
    \$\begingroup\$ If it goes wrong, it will explode the battery and burn down the car and the house it is parked in. Then your insurance will argue that your DIY charger was not certified. Also, output power can't exceed input power unless energy is stored in a stationary battery. \$\endgroup\$ – peufeu May 26 at 17:26
  • 8
    \$\begingroup\$ I love the oxymoron, "simple AC-DC power supply" followed by "400V/32A 3-phase input, and DC output adjustable (by a microcontroller which I would add) between 300-400VDC and upto 100A (not exceeding 22kW total)" -- it just doesn't get any better. \$\endgroup\$ – jonk May 26 at 17:30
  • 2
    \$\begingroup\$ I’m voting to close this question because questions seeking recommendations for specific products or places to purchase them are off-topic and requests for complete designs are also off topic. \$\endgroup\$ – Charles Cowie May 26 at 17:52
  • 1
    \$\begingroup\$ @CharlesCowie I'm not asking for a product recommendation (although I probably would happily buy one for under 2000$ if I knew of one) and I'm also not asking for someone to provide me a complete design. I'm sincerely asking for general knowledge (which I haven't found in days of searching the web) so I can better appreciate the value of the products in question. The "PFC" keyword was, so far, the most helpful fragment of an answer. \$\endgroup\$ – Ethan Arnold May 26 at 17:58
5
\$\begingroup\$

Power factor correction is the hidden biggy with any powerful 3 phase AC to DC converter. It’s simply not the case that the rather elegant 3 phase rectifier circuit can be piggy backed with a boosting power factor correction circuit. This will not work. This is because the elegant 3 phase rectifier (used for many tens of years if not a hundred years) just doesn’t allow it to happen. See this for an explanation.

This won't work: -

enter image description here

Because the current in each phase falls to zero while its respective voltage is still very high: -

enter image description here

The only viable alternative is to regard the 3 phases as 3 individual supplies and have power factor correction replicated for each. Then there is the requirement to maintain load balance so, there has to be significant cross checking between each phase and finally, each of the 3 boost converter's outputs has to be managed so that it can be adequately merged with the other two boosted DC voltages: -

enter image description here

There is a lot of overhead here and I reckon power factor correction alone is about 50% of the BoM cost for the circuits.

I’m not saying that $5k is a justifiable price to pay of course but given the market and trends and hidden complexity of PFC, it doesn't surprise me at all.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Fantastic answer, the linked reference explained exactly what I wanted to know. Thank you! Especially eye-opening is the fact that 3-phase is apparently no advantage, 1-phase is actually much simpler if your electric lines are dimensioned big enough. I suppose using 3 separate AC-DC converters, one on each phase, might actually be easier (outputs centrally controlled)? \$\endgroup\$ – Ethan Arnold May 26 at 19:57
  • 1
    \$\begingroup\$ Suprisingly, it’s probably easier to merge the three boosted PFC voltages then have a single brute of a dc to dc converter. Regulations require that phase Load balancing is a must and this means that the common point occurs before the brute force dc converter. \$\endgroup\$ – Andy aka May 26 at 20:08
  • 1
    \$\begingroup\$ How about a Vienna Rectifier (en.wikipedia.org/wiki/Vienna_rectifier)? \$\endgroup\$ – Ethan Arnold May 26 at 20:32
  • \$\begingroup\$ I think that might work from memory having looked at articles on it some time ago. Not an expert on them so I can’t speculate. \$\endgroup\$ – Andy aka May 26 at 21:16
  • 1
    \$\begingroup\$ The Vienna rectifier would work - my employer engaged with Dr. Kolar a few years back and we released a design based around it. It requires a lot of DSP-based control - I wouldn't go so far as qualifying it as "simple" per OP's requirements. \$\endgroup\$ – Adam Lawrence May 27 at 13:11

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.