Surge current limiting on large capacitors and power factor correction (PFC)

[P E Schoen]

I've posted previously about failures in my push-pull high power DC-DC
converter, which have most likely been caused by start-up surge currents
when first connected to the DC source (batteries), and when the converter
starts charging the output capacitors through a bridge rectifier with a
square wave. I think I have found several options such as a resistive
precharge resistor or a current-limiting switching circuit either as a
front-end add-on, or as part of the push-pull PWM.

Now I am looking at two designs used in electric vehicles and described in
threads:
http://www.diyelectriccar.com/forums/showthread.php/10kw-60a-diy-charger-open-source-59210.html
and
http://www.diyelectriccar.com/forums/showthread.php/evnetics-developing-dc-dc-converter-72862.html

The charger uses AC mains of 120 or 240 VAC to charge the battery pack,
which may be 72 VDC to as much as 700 VDC, and generally about 100 A-Hr,
which is 86 to 840 kW-hr. So an 8 hour charge may require 10 kW to as much
as 100 kW. One charger design is:
http://www.enginuitysystems.com/pix/VMcharger-V13.3.pdf

There have been some problems with getting maximum power because of the
input surge currents and high PF due to the capacitors. I offered some
improvement using a precharge resistor and an inductor as follows:
http://www.enginuitysystems.com/pix/240Sine-320VDC_FWB-s.png
http://www.enginuitysystems.com/pix/240Sine-320DC_FWB_Inductor-s.png
http://www.enginuitysystems.com/pix/240Sine-320DC_FWB_Inductor_Precharge-s.png

The LTSpice ASC files are in the same folder and you can read the directory:
http://www.enginuitysystems.com/pix/

To be really effective at line frequency, the inductor must be rather large
and expensive. And the power factor was not really improved all that much
with the smaller ones that are of reasonable size and cost.

So, I plan to try some switching techniques to reduce the initial inrush and
also to improve power factor. I found an interesting resource on PFC at:
http://cktse.eie.polyu.edu.hk/Tse-IEEElecture2.pdf
But I didn't understand all of it, and I wanted to propose an idea that may
or may not be one of the topologies discussed there.

If you know of any better resources or actual circuits, please let me know.
I did some searching but it seems that PFC circuits may be trade secrets.

Basically, I plan to just rectify the incoming power through a FWB. Then I
would like to design a switching circuit consisting of an inductor, diode,
and series MOSFET as a forward buck regulator, so that it would allow the
incoming voltage to apply current through the inductor until it reaches like
1.5 times the maximum expected output current, and then disconnect until the
current drops to about 0.75 times the output current. Thus it would run in
continuous mode and the input current would remain within reasonable limits
of the available source current.

As the input voltage increases, the duty cycle would decrease, although it
would be more a function of the difference between the input and output
voltage (on the main storage caps as they are charging). This process would
continue until the capacitors are fully charged and hopefully the PWM would
continue to draw a relatively constant current from the supply line. It may
even be better to operate this switcher as a buck/boost so that it will
provide a higher output voltage near the zero crossings of the input supply,
and a lower voltage when the supply peaks exceed the intended fully charged
output voltage.

This is still a conceptual design, so I wanted to get some feedback about
its chance of working and any pitfalls to avoid. This will be a fairly
high-end piece of equipment so it is no problem to use microcontrollers
and/or dedicated ICs. I'll discuss the other item (DC-DC converter) in a
separate thread.

Thanks!

Paul

 

[Joerg]

I've posted previously about failures in my push-pull high power DC-DC
converter, which have most likely been caused by start-up surge currents
when first connected to the DC source (batteries), and when the
converter starts charging the output capacitors through a bridge
rectifier with a square wave. I think I have found several options such
as a resistive precharge resistor or a current-limiting switching
circuit either as a front-end add-on, or as part of the push-pull PWM.

Now I am looking at two designs used in electric vehicles and described
in threads:
http://www.diyelectriccar.com/forums/showthread.php/10kw-60a-diy-charger-open-source-59210.html

and
http://www.diyelectriccar.com/forums/showthread.php/evnetics-developing-dc-dc-converter-72862.html


The charger uses AC mains of 120 or 240 VAC to charge the battery pack,
which may be 72 VDC to as much as 700 VDC, and generally about 100 A-Hr,
which is 86 to 840 kW-hr. So an 8 hour charge may require 10 kW to as
much as 100 kW. One charger design is:
http://www.enginuitysystems.com/pix/VMcharger-V13.3.pdf

There have been some problems with getting maximum power because of the
input surge currents and high PF due to the capacitors. I offered some
improvement using a precharge resistor and an inductor as follows:
http://www.enginuitysystems.com/pix/240Sine-320VDC_FWB-s.png
http://www.enginuitysystems.com/pix/240Sine-320DC_FWB_Inductor-s.png
http://www.enginuitysystems.com/pix/240Sine-320DC_FWB_Inductor_Precharge-s.png


The LTSpice ASC files are in the same folder and you can read the
directory:
http://www.enginuitysystems.com/pix/

To be really effective at line frequency, the inductor must be rather
large and expensive. And the power factor was not really improved all
that much with the smaller ones that are of reasonable size and cost.

So, I plan to try some switching techniques to reduce the initial inrush
and also to improve power factor. I found an interesting resource on PFC
at:
http://cktse.eie.polyu.edu.hk/Tse-IEEElecture2.pdf
But I didn't understand all of it, and I wanted to propose an idea that
may or may not be one of the topologies discussed there.

If you know of any better resources or actual circuits, please let me
know. I did some searching but it seems that PFC circuits may be trade
secrets.

It's no trade secret. This is how real PFC is done and they have a
circuit ready-to-go to get you started on LTSpice:

http://cds.linear.com/docs/Datasheet/3798f.pdf

There are also controllers that are essentially working as glorified
boost converters. Not much to it, but usually you end up having to wind
your own magnetics. Especially at the power levels you are talking about.

Basically, I plan to just rectify the incoming power through a FWB. Then
I would like to design a switching circuit consisting of an inductor,
diode, and series MOSFET as a forward buck regulator, so that it would
allow the incoming voltage to apply current through the inductor until
it reaches like 1.5 times the maximum expected output current, and then
disconnect until the current drops to about 0.75 times the output
current. Thus it would run in continuous mode and the input current
would remain within reasonable limits of the available source current.

Disconnect wouldn't be real PFC. If you are this ambitious about it I'd
do it right so the distortion in the current is really minimal.

[...]