3phase PFC

You can get a wide range and good efficiency but you end up with a mechanically large inductor and switch device.

Each stage should be limited to about a 4:1 voltage range.

Jamie, normal 3 phase full wave rectified AC/DC conversion yields 93% PF if done properly. Why are you adding all this hog ware to get a few more %PF? Cheers, Harry

Cheers Terry

Hi Terry,

I think this is called the "Minnesota rectifier" right? Can it be used without the centertap neutral of the 3phase source? I think there is no current loop unless the 3phase neutral is available, but for my generator we don't have the neutral..

cheers, Jamie

The existence of the generator's neutral terminal is something that might be profitable to investigate further. Perhaps you could consult the manual, or take a closer look at the actual hardware wiring. It's your generator; you can do what you want with it.

You might want to think simple:

The single-switch 3ph pfc circuit is probably best applied off-line, where filtering requirement of the critical conduction current is set by known hardware, and not paper standards.

With a generator as a source, there is no actual inrush problem, as you are never going to be hot plugging the converter into a running generator.

RL

Great paper! Thanks, Harry

....but by no means error-free.

With Mr M., it seems better to supply single in-depth refs about single issues, rather than a host of references that just get him wandering again.

A single switch requiring no complex or super-high-frequency modulation is something that can be modeled fairly accurately.

You can also throw big uncomplicated components at it, though split three ways, none of the other parts, besides the switch and PFC rectifier need to swot the whole load.

RL

Just a little warning about this and other single-switch 3ph pfc topologies.

It may not be obvious, but when the pfc chokes are in the line like this, the entire downstream cicuitry has boost switch noise imposed upon it, including bulk caps and downstream converters. This places obvious cmrr problems (with respect to the rest of the universe) on any control circuitry directly connected to it, and introduces serious emc compliance issues.

Luckily, any information that needs processing or communication is relatively simple. Functional troubleshooting can be tricky - don't expect to be able to just stick you grounded scope probe anywhere you please.

RL

Hi,

A bit off topic, but I also have a split phase inverter that has a similar technique of injection, but its with the 120VAC common:

"

The red bar is a +440VDC rail, and then there are two half bridges to give 120VAC/60Hz 180degrees out of phase, and they are centered around ~220VDC, which is the 120VAC common, held by the middle half bridge.

Will this setup work ok, and does it have the same CMRR problems due to the series L in the common return? The L is quite small, probably 20uH or less for the common.

cheers, Jamie

..and only nine components. Amazing!

There's no high frequency interference if the 440VDC is clean. The inverter outputs should only inter-tie through a working load.

You seem not only to be jumping between topologies, but are now going backwards, ie generating three-phase from some unknown source.

Problems are a lot easier to deal with if there is a fixed goal.

RL

never heard it called that.

Ive used this with a neutral connection, havent thought about no neutral connection.

Wait a minute.

Are you saying that you intend to connect one of your inverter outputs to hydro neutral (ie ground) ?

Might I ask, why?

Unless there is some galvanic isolation somewhere back in the power train, this will look like a short to somewhere.

If there IS isolation, you'll be driving the entire system with the inverter 'output' voltage, where driving interphase loads normally will only induce hf and modulation balancing errors on the system hardware.

RL

Actually, it is more likely than not that it is wound delta rather than wye. There are reasons for doing this.

Not true. Many Motor-Generator sets feed into automatic bus transfer switches.

If that were the case, then inrush would have to be addressed.

So far, that spanner hasn't been thown into the works, but I'm sure the OP will eventually get round to it.

RL

"

Yes, the halfbridge output that is connected to hydro neutral (ground) is there as the current return for the two 120VAC outputs, it just holds a constant DC voltage, which the two 120VAC phases are centered around, since 120VAC loads run from phase to neutral.

cheers, Jamie