[AMRadio] Smoothing Choke

Donald Chester k4kyv at hotmail.com
Thu May 4 17:53:21 EDT 2006

>    Sorry but I am going to throw a wrench into this discussion. Those
>swinging choke power supplies, although useful, have been out of favor for
>years, especially when the loading is transient such as for a CW rig. Sure
>if the critical inductance versus minimum load current are correct, the
>output voltage will not soar under a light or no load. But what happens 
>the load is applied suddenly such as with CW? The answer is arguable of
>course, but I think the answer is a sudden significant and momentary dip in
>B+ which all depends on the L-C-L-C values chosen.

This is called the dynamic regulation.  When keying a transmitter on cw, or 
running a class-B modulator or SSB linear, the plate (kilo)voltmeter may 
show little variation, indicating good regulation.  But if you set up an 
oscilloscope to display the instantaseous power supply voltage, you may be 
surprised how much the voltage kicks around with the transients.  The 
voltage not only dips down after the load is applied; when the load is 
removed the voltage will kick back upwards, and the momentum may swing it 
well above the nominal output voltage.  I have seen power supplies where the 
voltage dropped below 70% nominal volatge and back up to 140% nominal or 
higher - a 2 to 1 voltage variation while the sluggish mechanical movement 
in the analogue voltmeter just barely wiggled with keying or modulation. 
Monitoring the cw output from a transmitter using a modulation monitor scope 
may show a horrible keyed waveform.

The solution that has worked for me is to use as much capacitance in the 
output side of the filter as possible, without kicking the overload relay 
when the power supply is first turned on.  I prefer to kill the HV during 
receive.  Don't like having the full HV turned on while the transmitter is 
in standby - a death trap plus rf noise generator, and it hastens breakdown 
of HV components.  You can use step-start if you want to further increase 
the capacitance, but I prefer to avoid that complication.  Of course,  
increasing the inductance of the choke will reduce the inrush current to the 
capacitor by slowing down its charging rate.

I converted my Gates BC1-T to CW capability, and was able to get a 
reasonable looking waveform by increasing the capacitance in the one-section 
choke input filter from 8 mfd to 25 mfd.  It is still not perfect, but is 
satisfactory.  My homebrew transmitter uses about the same capacitance each 
in the rf final plate supply and modulator plate supply.  I have found that 
with a 2-2.5 kv power supply with a load not exceeding 700-800 mills total, 
a 25 mfd 4 kv output capacitor is about the best compromise between dynamic 
regulation and inrush current.

With AM, the best solution is to use a common power supply for the modulator 
and final.  The rf final acts like a heavy bleeder resistor, and maintains 
good dynamic regulation for the class-B modulator plates.  Poor dynamic 
regulation can result in poor sounding audio, even though every component in 
the modulator is of the highest quality and the circuit is of the best 

If you are having an inexplicable audio quality problem, take a look at your 
HV power supply output voltage on an oscilloscope, to see what the dynamic 
regulation looks like.  Connect up a series string of resistors, so that the 
total resistance is several times the normal bleeder resistance, and connect 
the vertical plates of the scope across the bottom resistor.  Make sure the 
wattage and voltage rating of the resistors is not exceeded.  Another way 
would be to temporarily disconnect the bottom end of the normal bleeder 
resistor, and insert a resistor of approximately 5% of the bleeder 
resistance in series with the bottom side.  Place the scope probes between 
the connection between this resistor and the bleeder, and ground, to read 
the voltage across this resistor.  If the resistor is 5% of the total 
resistance, 5% of the total voltage will appear across it.  Thus, with a 2kv 
power supply, 100 volts will appear across the resistor, which is much 
easier to measure with a scope than the full 2000.

Don k4kyv


This message was typed using the DVORAK keyboard layout.  Try it - you'll 
like it.

More information about the AMRadio mailing list

This page last updated 25 Feb 2018.