[Boatanchors] Resonant Choke Filters (was Please review & comment on my Heath Warrior HA-10 amplifier design changes)

[Rick Poole WA1RKT]

A followup on my Warrior HA-10 design revisions project of the last couple of months, along with a request for your opinions on some further changes contemplated...

The redesign and reconstruction was completed using most of the suggestions offered here (THANKS!!); in summary:

* removed 866's and added solid state rectifiers;

* added a 6.3VAC 20A filament transformer in place of the original-equipment filament/bias transformer which was apparently lost in shipment from the previous owner;

* directly grounded the 811A grids;

* raised the filament transformer centertap to nominally 4.7 volts above ground to provide bias;

* changed the metering circuitry including recalculating metering resistors and adding some protection;

* a few other minor changes.

The result works beautifully... up to about 700 watts PEP output when driven with a Heath Marauder, and the trapezoid pattern on a Kenwood SM-220 monitor scope looks almost perfectly linear, and excellent signal and audio reports with no distortion noted.  You can see the amp mounted in a rack above the 51J-4 and Marauder on my QRZ page.

As Gary Schafer noted below, I did have to retain the 60K 100W bleeder resistor to maintain a minimum load on the choke-input power supply.  That resistor generates A LOT of heat and consequently I leave the HV turned off unless I'm actually going to transmit.

Some work I needed to do on my Collins 516F-2 power supply (for the KWM-2A) over the last few days introduced me to the resonant-choke-input filter where a capacitor is added across the input choke to resonate at 120 Hz, which then led me to http://www.qsl.net/i0jx/supply.html which is a lengthy and detailed explanation of various filtering types including the resonant choke.  I'm thinking I could add a capacitor across the swinging choke to help keep the output voltage in range at lower minimum loads, so that I can increase the bleeder resistance and reduce the power (and heat) dissipated by that resistor.

Wondering if any of you have any experience with that.

I calculate that at the minimum current, which corresponds to the maximum inductance (50 Hy) of the swinging choke, I'll need a capacitance of a bit over 0.035 uf to resonate the choke at 120 Hz.  With the proper (higher) bleeder resistor value, which I haven't figured out yet, that should keep the HV at a reasonable value (well below 2 KV) under idling conditions.  Increased load current during operation reduces the swinging choke inductance, which takes the tuned circuit off of resonance and tends to increase the output voltage, resulting in better output voltage regulation over the load range.  Result is output voltage regulation at least as good as it is now and hopefully somewhat better, and MUCH less heat dissipated under the chassis by that monster bleeder resistor.

The Warrior design puts the choke in the plate transformer center tap to ground, rather than between the rectifier and the filter cap, something else I hadn't run into until I started working on this.  Seems like the resonant-choke concept should still work, can't think of a good reason why not, but I'm not really sure.

The way I have the amp mounted means a very large two-person job to dismount it and get it onto the workbench, so I'm not planning on doing this until the next time I have to work on the amp because something else broke, but I wanted to run all this past you folks and get your thoughts.

Thanks...

Rick WA1RKT
Londonderry, NH
Catch me on 14.300 MHz Intercon / MMSN most days


----- Original Message -----
From: "Gary Schafer" <garyschafer at largeriver.net>
To: "Rick Poole WA1RKT" <wa1rkt at comcast.net>, boatanchors at mailman.qth.net
Sent: Sunday, November 23, 2014 9:01:53 PM
Subject: RE: [Boatanchors] Please review & comment on my Heath Warrior HA-10 amplifier design changes


Another of the joy's of a choke input filter.

With the choke input filter you can't just arbitrarily change the bleeder
resistor size. You need a minimum load on the supply or the voltage will
soar very high with no load.

Right now the output voltage should be around .9 times RMS voltage of the
transformer. Without a substantial load it will start to look like a
capacitor input filter and the voltage can go as high as 1.4 times
transformer RMS voltage with no load.

<snip>