Choke Input Power Supply

[n4wc]

What effect does placing a capacitor across the reactor?  I assume the idea is to tune this to a parallel circuit at 120 cycles.  What benefit is it, if any?

Bill Cook

[WD5JKO]

Bill,

   Several things happen when you resonate that choke. Just make sure the choke is a smoothing choke and not a swinging choke. I will cover a few and leave others to fill this in:

1.) The ripple reduction of a single L-C choke input filter will improve dramatically.
2.) The switching transients of silicon diodes in the rectifier will go straight through a standard choke input filter, but with a resonant choke these are virtually eliminated. Switching to diodes instead of tube rectifiers results in these transients because for a portion of the power cycle both diodes (FW CT) or all 4 diodes are on (FW Bridge) for anywhere from a 1us to a 1ms depending on the parts used. When the slow turn off diode finally shuts off, a back EMF spike comes from the transformer.
3.) min load critical inductance changes .....
4.) Voltage regulation changes ....

The capacitor must be tough, and preferably oil filled with a voltage rating at least 2X the output DC voltage. I tried this once in a Central Electronics 20A, and used a .47uf 600V black beauty cap across a 4 H inductor. Things were wonderful until about 1 year later the capacitor shorted out.

73,
Jim
JKO

[WD5JKO]

Some good stuff about resonant choke input filters:

http://www.mail-archive.com/amradio@mailman.qth.net/msg17525.html

http://lists.contesting.com/_amps/2004-08/msg00030.html

[John K5PRO]

There is one misconception about using resonant choke topology for power supply filters. It doesn't do a thing for higher frequency harmonics of power line, i.e., 240, 360, 480, and higher. That's why I had to add a second section of L and C (albeit smaller components than a one section LC filter) after my resonant choke and shunt C section.

I did this work at Broadcast Electronics for the model FM1.5A, and model FM3.5A single phase powered FM transmitters in 1982. I summarized some of the work in that first link, some years ago on the Amps@contesting.com list.

[Steve - WB3HUZ]

In general, it's probably more trouble than it's worth. If you have a choke with sufficient inductance, you can always put enough capacitance behind it to make a good filter using the standard LC topology.

[WQ9E]

Jim makes a great point about needing a robust capacitor if you use one to resonate the choke; it will see significant voltage across the choke.  Bill Orr makes a similar point in advising caution when putting a filter choke in series with the normally grounded center-tap of a transformer secondary since this point is now raised well above ground potential.  Although you can generally get away with this it is possible to have a transformer failure due to voltage breakdown at a point that the transformer manufacturer expected to be at ground potential.

Rodger WQ9E

[KE6DF]

Bill Orr makes a similar point in advising caution when putting a filter choke in series with the normally grounded center-tap of a transformer secondary since this point is now raised well above ground potential. 

Rodger WQ9E

This is an excellent point. I have a simulation program for designing power supplies.

With a choke input supply, the voltage across the filter choke reaches a peak value approximately equal to the output voltage of the power supply.

Suppose the output of your transformer is 1150 - 0 - 1150 and you use a full wave center tapped circuit with choke input filtering. You would get about 1000 volts DC from the supply.

If you attach one side of the choke to the transformer center tap, and the other side to ground you would end up with a peak voltage on the center tap of about 1000V. The waveform is complex AC, but that would be the peak.

Some transformers like many of the old S series UTC units were designed to have the center tap grounded and you would have trouble using negative lead filtering.

Another reason people put the choke in the negative lead is to enable them to use a choke with a lower voltage spec. But this may lead to trouble also because the transient voltages in a choke input filter are quite high.

You would risk a breakdown in the choke unless it has a voltage spec near the output voltage of the supply.

[Opcom]

Which program? Is it PSUD by Duncan?

[KE6DF]

That's it. It doesn't allow you simulate resonant choke systems, unfortunately, and once in a while there is a glitch, but the price is right.

Free.

[WA1GFZ]

Linear Tech, will give you a free copy of SWCad all you need to do is go to their site and download it.