power supply formulas

[David, K3TUE]

I recall a conversation I heard, can't recall if it was here or on the air, where the people talking were discussing proper power supply filtering.  One was talking about how you should just slap the biggest capacitors you could find on, while the other was saying that was silly, and the formulas existed in some old Handbook to calculate the proper values for filter elements (choke, capacitors, etc.) for the least ripple.  Is anyone familiar with where I should look for this information?

[KB2WIG]

The handbook  ( pre 72?) has some sections on this subject....   klc

[WV Hoopie]

The old ARRL stuff from the 50's thru the 60's contain lots of gud stuff. Any old cap isn't the correct way to go. One thing mentioned in the old literature is to stay away from a L-C combination which would be resonate at power line frequencies.

Been a while since I've peeked in the library, but when I have a question the old books have the answers........

Craig,

[k4kyv]

There are other factors to  consider besides ripple, mainly dynamic regulation.  If there is a varying load, such as in the case of a slopbucket  linear, a cw transmitter, or a class-B modulator with separate power supply for the rf final, you need as much filter capacitance as your fuses can tolerate before the charging surge begins to blow fuses or kick out circuit breakers.

To check dynamic regulation, you need an oscilloscope.  Analogue meters may read little variation of the voltage when going from minimum to maximum load, but the scope may show that the voltage is actually kicking 30 or 40% above and below the average value during transients, but this variation does not register due to the mechanical inertia of the meter movement.

Using a choke input filter in my transmitter, I have found 25 mfd to be the optimum amount of filter capacitance in a single section filter with swinging choke.  That is the best compromise between dynamic regulation and charging surge current.  The optimum capacitance depends on the voltage/current ratio of the power supply under full load.  25 mfd seems optimum for 2000-2500 volts at maybe 500 ma maximum current, or about 5000 ohm load.

One of my homebrew transmitters runs with 1350 volts on the plate, with common power supply for modulator, final and class-B audio driver stage, maximum total current drain on the power supply 600-700 m.a.  I use a 53 mfd capacitor in that power supply, with the load @ slightly less than 2000 ohms.

This calculates out to dividing 100,000 by the effective  load resistance on the power supply at maximum load current to give the optimum output capacitance in micofarads.  Example:  the above transmitter with 1350 volts, at the maximum load current of 700 m.a. gives 1928.6 ohms.  100,000/1928.6= 51.9 mfd.

Still better dynamic regulation can be achieved by using up to 100 mfd in the capacitor, but some kind of step-start is required to charge up that much capacitance without blowing the fuse.  This can be problematic especially if you kill the hv during receive and have to recharge the capacitors every time you hit the transmit switch.  I don't like to leave the HV on my transmitter while it is in standby.  The idling final amplifier tube with plate voltage still on tends to generate white noise that can interfere with reception, and there is the increased likelihood of crapping out components, and above all, it is a safety hazard because it is too easy to forget that the HV is still on while the transmitter is not operating.

[Steve - WB3HUZ]

Read about dynamic regulation here

http://www.amwindow.org/tech/pdf/geps1.pdf


and here

http://www.amwindow.org/tech/pdf/geps2.pdf

[WA1GFZ]

You get the biggest cap you can find with the lowest resistance choke you can find
then pray.

[k4kyv]

Read about dynamic regulation here

Tnx for posting that.  I studied those G-E articles (acquired a hardcover, professionally  bound dead tree edition covering several years of flyers) back when I first started playing with high power in the early 60's.  At that time 99% of ham transmitters had atrocious dynamic regulation, which generated horrible cw waveform and unacceptable levels of distortion in class-B modulators.

George W2WLR wrote up a similar presentation in one of his AM P/X articles back in the mid 1980's.