On Tue, 15 Jun 1999 09:36:07 +0100 Peter Chadwick
<Peter_Chadwick@mitel.com> writes:
>Marv says:
>
>> In this implementation, energy storage, to filter the AC line, is
>>provided by a 4000 uF / 450 V capacitor at the output of the PFC.
This
>>provides the "power" and all that the small output cap is required to
do
>>is pacify the edges of the output converter which is already very close
>>to being a "DC" signal.
>
>Marv, am I correct in thinking that the feedback loop for regulation
comes
>from the final output? If so, I think I'm correct in saying that the
loop
>bandwidth has to be wide enough to keep the impedance low up up to
>several kHz so that the DC-DC converter regulation is cancelled out. And
>doesn't the 4000mFd/450volt capacitor (do you really mean 4000mFd, not
>400mFd?)have to handle a pretty hefty ripple current? What sort of order
of loop
>bandwidths do these devices usually use, anyway?
There are actually two loops.
The first loop is around the PFC and that has a rather low frequency
response... on the order of 10 Hz. This provides approximately 5%
regulation of the input stage at 375 volts. There is a small 120Hz
component on this power. The low frequency ripple component in the
filter cap(s) is less than in a conventional supply since the power
input/output is spread over the full line cycle rather than being
concentrated in short peaks. Note: The maximum output current of this
stage is only 8 amps. The high frequency ripple components must
necessarily be filtered with small value HF capable parts.
The second loop controls the phase modulated ZVS H-bridge DC-DC
converter and provides fine regulation with high bandwidth.
Before I continue... I do hope your mFd means microfarad and not
millifarad. (I have never gotten used to the British billion!)
I really do mean 4000uF. Unitrode suggests a design rule of thumb (for
400 volt PFC's) that the filter cap in uF be equal to 2X the power output
in watts. Some other designers have cut the requirement by a factor of 5
or so depending on the nature of the load and the allowable transient
response. Since this unit is designed to power an SSB linear, the high
end of the C range is indicated for stability. Note: 4000 uF @ 400 V
has the same energy storage as 40 uF @ 4000V. So, even though I am
using modern witchcraft to regulate the voltage, I still can't defy the
laws of Physics! :-)
>Sounds a 'challenging project'........professionally, that means that it
>probably can't be done anyway, but certainly not in the time scale and
with
>the resource that the boss is going to allocate!
Most of the hard work has been done! Mature IC's are available for
the PFC controller, the Phase controller and FET drive plus there are now
great FET's available at reasonable prices.
The gotcha's are in unexpected sectors.
The geometry of the PFC power components is rather sensitive though,
a new APT ap note (APT9901, if I recall correctly) does an excellent job
of detailing the caveats.
Some kind of soft start circuit (relay, SCR's, or big thermistor) is
a must to keep the input capacitor from popping the mains breakers at
initial turn on. This kind of thing is almost never shown in the trade
magazines!
And, as I mentioned previously, the Litz winding techniques required
for the output transformer: Take 27 strands of #26 wire and wind 48
turns for the primary of the toroid! Perhaps, the boss would want to
help wind that one?
>The filters on the input aren't likely to be negligible, either....
The topologies used in this design are inherently a lot cleaner than
the typical flyback switchers commonly used in consumer level products
with their attendant poor reputation for creating both line disturbances
and RFI.
If good RF design techniques are employed (toroid inductive
components, short fat wires, etc.) which are likewise necessary to avoid
overstressing the semiconductors then the result should be fairly radio
friendly.
Incidently, commercial standalone PFC units (1/2 of this power supply)
in this power range have been available for over 5 years. They cost in
the neighborhood of $500. Perhaps it is about time for some of them to
show up on the surplus market!?
73,
Marv WC6W
*
___________________________________________________________________
Get the Internet just the way you want it.
Free software, free e-mail, and free Internet access for a month!
Try Juno Web: http://dl.www.juno.com/dynoget/tagj.
--
FAQ on WWW: http://www.contesting.com/ampfaq.html
Submissions: amps@contesting.com
Administrative requests: amps-REQUEST@contesting.com
Problems: owner-amps@contesting.com
Search: http://www.contesting.com/km9p/search.htm
|